mirror of
https://github.com/LCTT/TranslateProject.git
synced 2025-01-13 22:30:37 +08:00
Merge remote-tracking branch 'LCTT/master'
This commit is contained in:
wxy
commit
37a8371678
110
published/20171214 IPv6 Auto-Configuration in Linux.md
Normal file
110
published/20171214 IPv6 Auto-Configuration in Linux.md
Normal file
@ -0,0 +1,110 @@
|
||||
在 Linux 中自动配置 IPv6 地址
|
||||
======
|
||||
|
||||
|
||||
|
||||
在 [KVM 中测试 IPv6 网络：第 1 部分][1] 一文中，我们学习了关于<ruby>唯一本地地址<rt>unique local addresses</rt></ruby>（ULA）的相关内容。在本文中，我们将学习如何为 ULA 自动配置 IP 地址。
|
||||
|
||||
### 何时使用唯一本地地址
|
||||
|
||||
<ruby>唯一本地地址<rt>unique local addresses</rt></ruby>（ULA）使用 `fd00::/8` 地址块，它类似于我们常用的 IPv4 的私有地址：`10.0.0.0/8`、`172.16.0.0/12`、以及 `192.168.0.0/16`。但它们并不能直接替换。IPv4 的私有地址分类和网络地址转换（NAT）功能是为了缓解 IPv4 地址短缺的问题，这是个明智的解决方案，它延缓了本该被替换的 IPv4 的生命周期。IPv6 也支持 NAT，但是我想不出使用它的理由。IPv6 的地址数量远远大于 IPv4；它是不一样的，因此需要做不一样的事情。
|
||||
|
||||
那么，ULA 存在的意义是什么呢？尤其是在我们已经有了<ruby>本地链路地址<rt>link-local addresses</rt></ruby>（`fe80::/10`）时，到底需不需要我们去配置它们呢？它们之间（LCTT 译注：指的是唯一本地地址和本地链路地址）有两个重要的区别。一是，本地链路地址是不可路由的，因此，你不能跨子网使用它。二是，ULA 是你自己管理的；你可以自己选择它用于子网的地址范围，并且它们是可路由的。
|
||||
|
||||
使用 ULA 的另一个好处是，如果你只是在局域网中“混日子”的话，你不需要为它们分配全局单播 IPv6 地址。当然了，如果你的 ISP 已经为你分配了 IPv6 的<ruby>全局单播地址<rt>global unicast addresses</rt></ruby>，就不需要使用 ULA 了。你也可以在同一个网络中混合使用全局单播地址和 ULA，但是，我想不出这样使用的一个好理由，并且要一定确保你不使用网络地址转换（NAT）以使 ULA 可公共访问。在我看来，这是很愚蠢的行为。
|
||||
|
||||
ULA 是仅为私有网络使用的，并且应该阻止其流出你的网络，不允许进入因特网。这很简单，在你的边界设备上只要阻止整个 `fd00::/8` 范围的 IPv6 地址即可实现。
|
||||
|
||||
### 地址自动配置
|
||||
|
||||
ULA 不像本地链路地址那样自动配置的，但是使用 radvd 设置自动配置是非常容易的，radva 是路由器公告守护程序。在你开始之前，运行 `ifconfig` 或者 `ip addr show` 去查看你现有的 IP 地址。
|
||||
|
||||
在生产系统上使用时，你应该将 radvd 安装在一台单独的路由器上，如果只是测试使用，你可以将它安装在你的网络中的任意 Linux PC 上。在我的小型 KVM 测试实验室中，我使用 `apt-get install radvd` 命令把它安装在 Ubuntu 上。安装完成之后，我先不启动它，因为它还没有配置文件：
|
||||
|
||||
```
|
||||
$ sudo systemctl status radvd
|
||||
● radvd.service - LSB: Router Advertising Daemon
|
||||
Loaded: loaded (/etc/init.d/radvd; bad; vendor preset: enabled)
|
||||
Active: active (exited) since Mon 2017-12-11 20:08:25 PST; 4min 59s ago
|
||||
Docs: man:systemd-sysv-generator(8)
|
||||
|
||||
Dec 11 20:08:25 ubunut1 systemd[1]: Starting LSB: Router Advertising Daemon...
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: Starting radvd:
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: * /etc/radvd.conf does not exist or is empty.
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: * See /usr/share/doc/radvd/README.Debian
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: * radvd will *not* be started.
|
||||
Dec 11 20:08:25 ubunut1 systemd[1]: Started LSB: Router Advertising Daemon.
|
||||
```
|
||||
|
||||
这些所有的消息有点让人困惑，实际上 radvd 并没有运行，你可以使用经典命令 `ps | grep radvd` 来验证这一点。因此，我们现在需要去创建 `/etc/radvd.conf` 文件。拷贝这个示例，将第一行的网络接口名替换成你自己的接口名字：
|
||||
|
||||
```
|
||||
interface ens7 {
|
||||
AdvSendAdvert on;
|
||||
MinRtrAdvInterval 3;
|
||||
MaxRtrAdvInterval 10;
|
||||
prefix fd7d:844d:3e17:f3ae::/64
|
||||
{
|
||||
AdvOnLink on;
|
||||
AdvAutonomous on;
|
||||
};
|
||||
|
||||
};
|
||||
```
|
||||
|
||||
前缀（`prefix`）定义了你的网络地址，它是地址的前 64 位。前两个字符必须是 `fd`，前缀接下来的剩余部分你自己定义它，最后的 64 位留空，因为 radvd 将去分配最后的 64 位。前缀后面的 16 位用来定义子网，剩余的地址定义为主机地址。你的子网必须总是 `/64`。RFC 4193 要求地址必须随机生成；查看 [在 KVM 中测试 IPv6 Networking：第 1 部分][1] 学习创建和管理 ULAs 的更多知识。
|
||||
|
||||
### IPv6 转发
|
||||
|
||||
IPv6 转发必须要启用。下面的命令去启用它，重启后生效：
|
||||
|
||||
```
|
||||
$ sudo sysctl -w net.ipv6.conf.all.forwarding=1
|
||||
```
|
||||
|
||||
取消注释或者添加如下的行到 `/etc/sysctl.conf` 文件中，以使它永久生效：
|
||||
|
||||
```
|
||||
net.ipv6.conf.all.forwarding = 1
|
||||
```
|
||||
|
||||
启动 radvd 守护程序：
|
||||
|
||||
```
|
||||
$ sudo systemctl stop radvd
|
||||
$ sudo systemctl start radvd
|
||||
```
|
||||
|
||||
这个示例在我的 Ubuntu 测试系统中遇到了一个怪事；radvd 总是停止，我查看它的状态却没有任何问题，做任何改变之后都需要重新启动 radvd。
|
||||
|
||||
启动成功后没有任何输出，并且失败也是如此，因此，需要运行 `sudo systemctl status radvd` 去查看它的运行状态。如果有错误，`systemctl` 会告诉你。一般常见的错误都是 `/etc/radvd.conf` 中的语法错误。
|
||||
|
||||
在 Twitter 上抱怨了上述问题之后，我学到了一件很酷的技巧：当你运行 ` journalctl -xe --no-pager` 去调试 `systemctl` 错误时，你的输出会被换行，然后，你就可以看到错误信息。
|
||||
|
||||
现在检查你的主机，查看它们自动分配的新地址：
|
||||
|
||||
```
|
||||
$ ifconfig
|
||||
ens7 Link encap:Ethernet HWaddr 52:54:00:57:71:50
|
||||
[...]
|
||||
inet6 addr: fd7d:844d:3e17:f3ae:9808:98d5:bea9:14d9/64 Scope:Global
|
||||
[...]
|
||||
```
|
||||
|
||||
本文到此为止，下周继续学习如何为 ULA 管理 DNS，这样你就可以使用一个合适的主机名来代替这些长长的 IPv6 地址。
|
||||
|
||||
通过来自 Linux 基金会和 edX 的 [“Linux 入门”][2] 免费课程学习更多 Linux 的知识。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/learn/intro-to-linux/2017/12/ipv6-auto-configuration-linux
|
||||
|
||||
作者：[Carla Schroder][a]
|
||||
译者：[qhwdw](https://github.com/qhwdw)
|
||||
校对：[wxy](https://github.com/wxy)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/cschroder
|
||||
[1]:https://www.linux.com/learn/intro-to-linux/2017/11/testing-ipv6-networking-kvm-part-1
|
||||
[2]:https://training.linuxfoundation.org/linux-courses/system-administration-training/introduction-to-linux
|
||||
224
published/20180116 Analyzing the Linux boot process.md
Normal file
224
published/20180116 Analyzing the Linux boot process.md
Normal file
@ -0,0 +1,224 @@
|
||||
Linux 启动过程分析
|
||||
======
|
||||
|
||||
> 理解运转良好的系统对于处理不可避免的故障是最好的准备。
|
||||
|
||||
|
||||
|
||||
*图片由企鹅和靴子“赞助”，由 Opensource.com 修改。CC BY-SA 4.0。*
|
||||
|
||||
关于开源软件最古老的笑话是：“代码是<ruby>自具文档化的<rt>self-documenting</rt></ruby>”。经验表明，阅读源代码就像听天气预报一样：明智的人依然出门会看看室外的天气。本文讲述了如何运用调试工具来观察和分析 Linux 系统的启动。分析一个功能正常的系统启动过程，有助于用户和开发人员应对不可避免的故障。
|
||||
|
||||
从某些方面看，启动过程非常简单。内核在单核上以单线程和同步状态启动，似乎可以理解。但内核本身是如何启动的呢？[initrd（initial ramdisk）][1] 和<ruby>引导程序<rt>bootloader</rt></ruby>具有哪些功能？还有，为什么以太网端口上的 LED 灯是常亮的呢？
|
||||
|
||||
请继续阅读寻找答案。在 GitHub 上也提供了 [介绍演示和练习的代码][2]。
|
||||
|
||||
### 启动的开始：OFF 状态
|
||||
|
||||
#### <ruby>局域网唤醒<rt>Wake-on-LAN</rt></ruby>
|
||||
|
||||
OFF 状态表示系统没有上电，没错吧？表面简单，其实不然。例如，如果系统启用了局域网唤醒机制（WOL），以太网指示灯将亮起。通过以下命令来检查是否是这种情况：
|
||||
|
||||
```
|
||||
# sudo ethtool <interface name>
|
||||
```
|
||||
|
||||
其中 `<interface name>` 是网络接口的名字，比如 `eth0`。（`ethtool` 可以在同名的 Linux 软件包中找到。）如果输出中的 `Wake-on` 显示 `g`，则远程主机可以通过发送 [<ruby>魔法数据包<rt>MagicPacket</rt></ruby>][3] 来启动系统。如果您无意远程唤醒系统，也不希望其他人这样做，请在系统 BIOS 菜单中将 WOL 关闭，或者用以下方式：
|
||||
|
||||
```
|
||||
# sudo ethtool -s <interface name> wol d
|
||||
```
|
||||
|
||||
响应魔法数据包的处理器可能是网络接口的一部分，也可能是 [<ruby>底板管理控制器<rt>Baseboard Management Controller</rt></ruby>][4]（BMC）。
|
||||
|
||||
#### 英特尔管理引擎、平台控制器单元和 Minix
|
||||
|
||||
BMC 不是唯一的在系统关闭时仍在监听的微控制器（MCU）。x86_64 系统还包含了用于远程管理系统的英特尔管理引擎（IME）软件套件。从服务器到笔记本电脑，各种各样的设备都包含了这项技术，它开启了如 KVM 远程控制和英特尔功能许可服务等 [功能][5]。根据 [Intel 自己的检测工具][7]，[IME 存在尚未修补的漏洞][6]。坏消息是，要禁用 IME 很难。Trammell Hudson 发起了一个 [me_cleaner 项目][8]，它可以清除一些相对恶劣的 IME 组件，比如嵌入式 Web 服务器，但也可能会影响运行它的系统。
|
||||
|
||||
IME 固件和<ruby>系统管理模式<rt>System Management Mode</rt></ruby>（SMM）软件是 [基于 Minix 操作系统][9] 的，并运行在单独的<ruby>平台控制器单元<rt>Platform Controller Hub</rt></ruby>上（LCTT 译注：即南桥芯片），而不是主 CPU 上。然后，SMM 启动位于主处理器上的<ruby>通用可扩展固件接口<rt>Universal Extensible Firmware Interface</rt></ruby>（UEFI）软件，相关内容 [已被提及多次][10]。Google 的 Coreboot 小组已经启动了一个雄心勃勃的 [<ruby>非扩展性缩减版固件<rt>Non-Extensible Reduced Firmware</rt></ruby>][11]（NERF）项目，其目的不仅是要取代 UEFI，还要取代早期的 Linux 用户空间组件，如 systemd。在我们等待这些新成果的同时，Linux 用户现在就可以从 Purism、System76 或 Dell 等处购买 [禁用了 IME][12] 的笔记本电脑，另外 [带有 ARM 64 位处理器笔记本电脑][13] 还是值得期待的。
|
||||
|
||||
#### 引导程序
|
||||
|
||||
除了启动那些问题不断的间谍软件外，早期引导固件还有什么功能呢？引导程序的作用是为新上电的处理器提供通用操作系统（如 Linux）所需的资源。在开机时，不但没有虚拟内存，在控制器启动之前连 DRAM 也没有。然后，引导程序打开电源，并扫描总线和接口，以定位内核镜像和根文件系统的位置。U-Boot 和 GRUB 等常见的引导程序支持 USB、PCI 和 NFS 等接口，以及更多的嵌入式专用设备，如 NOR 闪存和 NAND 闪存。引导程序还与 [<ruby>可信平台模块<rt>Trusted Platform Module</rt></ruby>][14]（TPM）等硬件安全设备进行交互，在启动最开始建立信任链。
|
||||
|
||||
![Running the U-boot bootloader][16]
|
||||
|
||||
*在构建主机上的沙盒中运行 U-boot 引导程序。*
|
||||
|
||||
包括树莓派、任天堂设备、汽车主板和 Chromebook 在内的系统都支持广泛使用的开源引导程序 [U-Boot][17]。它没有系统日志，当发生问题时，甚至没有任何控制台输出。为了便于调试，U-Boot 团队提供了一个沙盒，可以在构建主机甚至是夜间的持续集成（CI）系统上测试补丁程序。如果系统上安装了 Git 和 GNU Compiler Collection（GCC）等通用的开发工具，使用 U-Boot 沙盒会相对简单：
|
||||
|
||||
```
|
||||
# git clone git://git.denx.de/u-boot; cd u-boot
|
||||
# make ARCH=sandbox defconfig
|
||||
# make; ./u-boot
|
||||
=> printenv
|
||||
=> help
|
||||
```
|
||||
|
||||
在 x86_64 上运行 U-Boot，可以测试一些棘手的功能，如 [模拟存储设备][2] 的重新分区、基于 TPM 的密钥操作以及 USB 设备热插拔等。U-Boot 沙盒甚至可以在 GDB 调试器下单步执行。使用沙盒进行开发的速度比将引导程序刷新到电路板上的测试快 10 倍，并且可以使用 `Ctrl + C` 恢复一个“变砖”的沙盒。
|
||||
|
||||
### 启动内核
|
||||
|
||||
#### 配置引导内核
|
||||
|
||||
引导程序完成任务后将跳转到已加载到主内存中的内核代码，并开始执行，传递用户指定的任何命令行选项。内核是什么样的程序呢？用命令 `file /boot/vmlinuz` 可以看到它是一个 “bzImage”，意思是一个大的压缩的镜像。Linux 源代码树包含了一个可以解压缩这个文件的工具—— [extract-vmlinux][18]：
|
||||
|
||||
```
|
||||
# scripts/extract-vmlinux /boot/vmlinuz-$(uname -r) > vmlinux
|
||||
# file vmlinux
|
||||
vmlinux: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically
|
||||
linked, stripped
|
||||
```
|
||||
|
||||
内核是一个 [<ruby>可执行与可链接格式<rt> Executable and Linking Format</rt></ruby>][19]（ELF）的二进制文件，就像 Linux 的用户空间程序一样。这意味着我们可以使用 `binutils` 包中的命令，如 `readelf` 来检查它。比较一下输出，例如：
|
||||
|
||||
```
|
||||
# readelf -S /bin/date
|
||||
# readelf -S vmlinux
|
||||
```
|
||||
|
||||
这两个二进制文件中的段内容大致相同。
|
||||
|
||||
所以内核必须像其他的 Linux ELF 文件一样启动，但用户空间程序是如何启动的呢？在 `main()` 函数中？并不确切。
|
||||
|
||||
在 `main()` 函数运行之前，程序需要一个执行上下文，包括堆栈内存以及 `stdio`、`stdout` 和 `stderr` 的文件描述符。用户空间程序从标准库（多数 Linux 系统在用 “glibc”）中获取这些资源。参照以下输出：
|
||||
|
||||
```
|
||||
# file /bin/date
|
||||
/bin/date: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically
|
||||
linked, interpreter /lib64/ld-linux-x86-64.so.2, for GNU/Linux 2.6.32,
|
||||
BuildID[sha1]=14e8563676febeb06d701dbee35d225c5a8e565a,
|
||||
stripped
|
||||
```
|
||||
|
||||
ELF 二进制文件有一个解释器，就像 Bash 和 Python 脚本一样，但是解释器不需要像脚本那样用 `#!` 指定，因为 ELF 是 Linux 的原生格式。ELF 解释器通过调用 `_start()` 函数来用所需资源 [配置一个二进制文件][20]，这个函数可以从 glibc 源代码包中找到，可以 [用 GDB 查看][21]。内核显然没有解释器，必须自我配置，这是怎么做到的呢？
|
||||
|
||||
用 GDB 检查内核的启动给出了答案。首先安装内核的调试软件包，内核中包含一个<ruby>未剥离的<rt>unstripped</rt></ruby> vmlinux，例如 `apt-get install linux-image-amd64-dbg`，或者从源代码编译和安装你自己的内核，可以参照 [Debian Kernel Handbook][22] 中的指令。`gdb vmlinux` 后加 `info files` 可显示 ELF 段 `init.text`。在 `init.text` 中用 `l *(address)` 列出程序执行的开头，其中 `address` 是 `init.text` 的十六进制开头。用 GDB 可以看到 x86_64 内核从内核文件 [arch/x86/kernel/head_64.S][23] 开始启动，在这个文件中我们找到了汇编函数 `start_cpu0()`，以及一段明确的代码显示在调用 `x86_64 start_kernel()` 函数之前创建了堆栈并解压了 zImage。ARM 32 位内核也有类似的文件 [arch/arm/kernel/head.S][24]。`start_kernel()` 不针对特定的体系结构，所以这个函数驻留在内核的 [init/main.c][25] 中。`start_kernel()` 可以说是 Linux 真正的 `main()` 函数。
|
||||
|
||||
### 从 start_kernel() 到 PID 1
|
||||
|
||||
#### 内核的硬件清单：设备树和 ACPI 表
|
||||
|
||||
在引导时，内核需要硬件信息，不仅仅是已编译过的处理器类型。代码中的指令通过单独存储的配置数据进行扩充。有两种主要的数据存储方法：[<ruby>设备树<rt>device-tree</rt></ruby>][26] 和 [高级配置和电源接口（ACPI）表][27]。内核通过读取这些文件了解每次启动时需要运行的硬件。
|
||||
|
||||
对于嵌入式设备，设备树是已安装硬件的清单。设备树只是一个与内核源代码同时编译的文件，通常与 `vmlinux` 一样位于 `/boot` 目录中。要查看 ARM 设备上的设备树的内容，只需对名称与 `/boot/*.dtb` 匹配的文件执行 `binutils` 包中的 `strings` 命令即可，这里 `dtb` 是指<ruby>设备树二进制文件<rt>device-tree binary</rt></ruby>。显然，只需编辑构成它的类 JSON 的文件并重新运行随内核源代码提供的特殊 `dtc` 编译器即可修改设备树。虽然设备树是一个静态文件，其文件路径通常由命令行引导程序传递给内核，但近年来增加了一个 [设备树覆盖][28] 的功能，内核在启动后可以动态加载热插拔的附加设备。
|
||||
|
||||
x86 系列和许多企业级的 ARM64 设备使用 [ACPI][27] 机制。与设备树不同的是，ACPI 信息存储在内核在启动时通过访问板载 ROM 而创建的 `/sys/firmware/acpi/tables` 虚拟文件系统中。读取 ACPI 表的简单方法是使用 `acpica-tools` 包中的 `acpidump` 命令。例如：
|
||||
|
||||
![ACPI tables on Lenovo laptops][30]
|
||||
|
||||
*联想笔记本电脑的 ACPI 表都是为 Windows 2001 设置的。*
|
||||
|
||||
是的，你的 Linux 系统已经准备好用于 Windows 2001 了，你要考虑安装吗？与设备树不同，ACPI 具有方法和数据，而设备树更多地是一种硬件描述语言。ACPI 方法在启动后仍处于活动状态。例如，运行 `acpi_listen` 命令（在 `apcid` 包中），然后打开和关闭笔记本机盖会发现 ACPI 功能一直在运行。暂时地和动态地 [覆盖 ACPI 表][31] 是可能的，而永久地改变它需要在引导时与 BIOS 菜单交互或刷新 ROM。如果你遇到那么多麻烦，也许你应该 [安装 coreboot][32]，这是开源固件的替代品。
|
||||
|
||||
#### 从 start_kernel() 到用户空间
|
||||
|
||||
[init/main.c][25] 中的代码竟然是可读的，而且有趣的是，它仍然在使用 1991 - 1992 年的 Linus Torvalds 的原始版权。在一个刚启动的系统上运行 `dmesg | head`，其输出主要来源于此文件。第一个 CPU 注册到系统中，全局数据结构被初始化，并且调度程序、中断处理程序（IRQ）、定时器和控制台按照严格的顺序逐一启动。在 `timekeeping_init()` 函数运行之前，所有的时间戳都是零。内核初始化的这部分是同步的，也就是说执行只发生在一个线程中，在最后一个完成并返回之前，没有任何函数会被执行。因此，即使在两个系统之间，`dmesg` 的输出也是完全可重复的，只要它们具有相同的设备树或 ACPI 表。Linux 的行为就像在 MCU 上运行的 RTOS（实时操作系统）一样，如 QNX 或 VxWorks。这种情况持续存在于函数 `rest_init()` 中，该函数在终止时由 `start_kernel()` 调用。
|
||||
|
||||
![Summary of early kernel boot process.][34]
|
||||
|
||||
*早期的内核启动流程。*
|
||||
|
||||
函数 `rest_init()` 产生了一个新进程以运行 `kernel_init()`，并调用了 `do_initcalls()`。用户可以通过将 `initcall_debug` 附加到内核命令行来监控 `initcalls`，这样每运行一次 `initcall` 函数就会产生 一个 `dmesg` 条目。`initcalls` 会历经七个连续的级别：early、core、postcore、arch、subsys、fs、device 和 late。`initcalls` 最为用户可见的部分是所有处理器外围设备的探测和设置：总线、网络、存储和显示器等等，同时加载其内核模块。`rest_init()` 也会在引导处理器上产生第二个线程，它首先运行 `cpu_idle()`，然后等待调度器分配工作。
|
||||
|
||||
`kernel_init()` 也可以 [设置对称多处理（SMP）结构][35]。在较新的内核中，如果 `dmesg` 的输出中出现 “Bringing up secondary CPUs...” 等字样，系统便使用了 SMP。SMP 通过“热插拔” CPU 来进行，这意味着它用状态机来管理其生命周期，这种状态机在概念上类似于热插拔的 U 盘一样。内核的电源管理系统经常会使某个<ruby>核<rt>core</rt></ruby>离线，然后根据需要将其唤醒，以便在不忙的机器上反复调用同一段的 CPU 热插拔代码。观察电源管理系统调用 CPU 热插拔代码的 [BCC 工具][36] 称为 `offcputime.py`。
|
||||
|
||||
请注意，`init/main.c` 中的代码在 `smp_init()` 运行时几乎已执行完毕：引导处理器已经完成了大部分一次性初始化操作，其它核无需重复。尽管如此，跨 CPU 的线程仍然要在每个核上生成，以管理每个核的中断（IRQ）、工作队列、定时器和电源事件。例如，通过 `ps -o psr` 命令可以查看服务每个 CPU 上的线程的 softirqs 和 workqueues。
|
||||
|
||||
```
|
||||
# ps -o pid,psr,comm $(pgrep ksoftirqd)
|
||||
PID PSR COMMAND
|
||||
7 0 ksoftirqd/0
|
||||
16 1 ksoftirqd/1
|
||||
22 2 ksoftirqd/2
|
||||
28 3 ksoftirqd/3
|
||||
|
||||
# ps -o pid,psr,comm $(pgrep kworker)
|
||||
PID PSR COMMAND
|
||||
4 0 kworker/0:0H
|
||||
18 1 kworker/1:0H
|
||||
24 2 kworker/2:0H
|
||||
30 3 kworker/3:0H
|
||||
[ . . . ]
|
||||
```
|
||||
|
||||
其中，PSR 字段代表“<ruby>处理器<rt>processor</rt></ruby>”。每个核还必须拥有自己的定时器和 `cpuhp` 热插拔处理程序。
|
||||
|
||||
那么用户空间是如何启动的呢？在最后，`kernel_init()` 寻找可以代表它执行 `init` 进程的 `initrd`。如果没有找到，内核直接执行 `init` 本身。那么为什么需要 `initrd` 呢？
|
||||
|
||||
#### 早期的用户空间：谁规定要用 initrd？
|
||||
|
||||
除了设备树之外，在启动时可以提供给内核的另一个文件路径是 `initrd` 的路径。`initrd` 通常位于 `/boot` 目录中，与 x86 系统中的 bzImage 文件 vmlinuz 一样，或是与 ARM 系统中的 uImage 和设备树相同。用 `initramfs-tools-core` 软件包中的 `lsinitramfs` 工具可以列出 `initrd` 的内容。发行版的 `initrd` 方案包含了最小化的 `/bin`、`/sbin` 和 `/etc` 目录以及内核模块，还有 `/scripts` 中的一些文件。所有这些看起来都很熟悉，因为 `initrd` 大致上是一个简单的最小化 Linux 根文件系统。看似相似，其实不然，因为位于虚拟内存盘中的 `/bin` 和 `/sbin` 目录下的所有可执行文件几乎都是指向 [BusyBox 二进制文件][38] 的符号链接，由此导致 `/bin` 和 `/sbin` 目录比 glibc 的小 10 倍。
|
||||
|
||||
如果要做的只是加载一些模块，然后在普通的根文件系统上启动 `init`，为什么还要创建一个 `initrd` 呢？想想一个加密的根文件系统，解密可能依赖于加载一个位于根文件系统 `/lib/modules` 的内核模块，当然还有 `initrd` 中的。加密模块可能被静态地编译到内核中，而不是从文件加载，但有多种原因不希望这样做。例如，用模块静态编译内核可能会使其太大而不能适应存储空间，或者静态编译可能会违反软件许可条款。不出所料，存储、网络和人类输入设备（HID）驱动程序也可能存在于 `initrd` 中。`initrd` 基本上包含了任何挂载根文件系统所必需的非内核代码。`initrd` 也是用户存放 [自定义ACPI][38] 表代码的地方。
|
||||
|
||||
![Rescue shell and a custom <code>initrd</code>.][40]
|
||||
|
||||
*救援模式的 shell 和自定义的 `initrd` 还是很有意思的。*
|
||||
|
||||
`initrd` 对测试文件系统和数据存储设备也很有用。将这些测试工具存放在 `initrd` 中，并从内存中运行测试，而不是从被测对象中运行。
|
||||
|
||||
最后，当 `init` 开始运行时，系统就启动啦！由于第二个处理器现在在运行，机器已经成为我们所熟知和喜爱的异步、可抢占、不可预测和高性能的生物。的确，`ps -o pid,psr,comm -p 1` 很容易显示用户空间的 `init` 进程已不在引导处理器上运行了。
|
||||
|
||||
### 总结
|
||||
|
||||
Linux 引导过程听起来或许令人生畏，即使是简单嵌入式设备上的软件数量也是如此。但换个角度来看，启动过程相当简单，因为启动中没有抢占、RCU 和竞争条件等扑朔迷离的复杂功能。只关注内核和 PID 1 会忽略了引导程序和辅助处理器为运行内核执行的大量准备工作。虽然内核在 Linux 程序中是独一无二的，但通过一些检查 ELF 文件的工具也可以了解其结构。学习一个正常的启动过程，可以帮助运维人员处理启动的故障。
|
||||
|
||||
要了解更多信息，请参阅 Alison Chaiken 的演讲——[Linux: The first second][41]，已于 1 月 22 日至 26 日在悉尼举行。参见 [linux.conf.au][42]。
|
||||
|
||||
感谢 [Akkana Peck][43] 的提议和指正。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/1/analyzing-linux-boot-process
|
||||
|
||||
作者：[Alison Chaiken][a]
|
||||
译者：[jessie-pang](https://github.com/jessie-pang)
|
||||
校对：[wxy](https://github.com/wxy)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/don-watkins
|
||||
[1]:https://en.wikipedia.org/wiki/Initial_ramdisk
|
||||
[2]:https://github.com/chaiken/LCA2018-Demo-Code
|
||||
[3]:https://en.wikipedia.org/wiki/Wake-on-LAN
|
||||
[4]:https://lwn.net/Articles/630778/
|
||||
[5]:https://www.youtube.com/watch?v=iffTJ1vPCSo&amp;amp;amp;amp;amp;index=65&amp;amp;amp;amp;amp;list=PLbzoR-pLrL6pISWAq-1cXP4_UZAyRtesk
|
||||
[6]:https://security-center.intel.com/advisory.aspx?intelid=INTEL-SA-00086&amp;amp;amp;amp;amp;languageid=en-fr
|
||||
[7]:https://www.intel.com/content/www/us/en/support/articles/000025619/software.html
|
||||
[8]:https://github.com/corna/me_cleaner
|
||||
[9]:https://lwn.net/Articles/738649/
|
||||
[10]:https://lwn.net/Articles/699551/
|
||||
[11]:https://trmm.net/NERF
|
||||
[12]:https://www.extremetech.com/computing/259879-dell-now-shipping-laptops-intels-management-engine-disabled
|
||||
[13]:https://lwn.net/Articles/733837/
|
||||
[14]:https://linuxplumbersconf.org/2017/ocw/events/LPC2017/tracks/639
|
||||
[15]:/file/383501
|
||||
[16]:https://opensource.com/sites/default/files/u128651/linuxboot_1.png "Running the U-boot bootloader"
|
||||
[17]:http://www.denx.de/wiki/DULG/Manual
|
||||
[18]:https://github.com/torvalds/linux/blob/master/scripts/extract-vmlinux
|
||||
[19]:http://man7.org/linux/man-pages/man5/elf.5.html
|
||||
[20]:https://0xax.gitbooks.io/linux-insides/content/Misc/program_startup.html
|
||||
[21]:https://github.com/chaiken/LCA2018-Demo-Code/commit/e543d9812058f2dd65f6aed45b09dda886c5fd4e
|
||||
[22]:http://kernel-handbook.alioth.debian.org/
|
||||
[23]:https://github.com/torvalds/linux/blob/master/arch/x86/boot/compressed/head_64.S
|
||||
[24]:https://github.com/torvalds/linux/blob/master/arch/arm/boot/compressed/head.S
|
||||
[25]:https://github.com/torvalds/linux/blob/master/init/main.c
|
||||
[26]:https://www.youtube.com/watch?v=m_NyYEBxfn8
|
||||
[27]:http://events.linuxfoundation.org/sites/events/files/slides/x86-platform.pdf
|
||||
[28]:http://lwn.net/Articles/616859/
|
||||
[29]:/file/383506
|
||||
[30]:https://opensource.com/sites/default/files/u128651/linuxboot_2.png "ACPI tables on Lenovo laptops"
|
||||
[31]:https://www.mjmwired.net/kernel/Documentation/acpi/method-customizing.txt
|
||||
[32]:https://www.coreboot.org/Supported_Motherboards
|
||||
[33]:/file/383511
|
||||
[34]:https://opensource.com/sites/default/files/u128651/linuxboot_3.png "Summary of early kernel boot process."
|
||||
[35]:http://free-electrons.com/pub/conferences/2014/elc/clement-smp-bring-up-on-arm-soc
|
||||
[36]:http://www.brendangregg.com/ebpf.html
|
||||
[37]:https://www.busybox.net/
|
||||
[38]:https://www.mjmwired.net/kernel/Documentation/acpi/initrd_table_override.txt
|
||||
[39]:/file/383516
|
||||
[40]:https://opensource.com/sites/default/files/u128651/linuxboot_4.png "Rescue shell and a custom <code>initrd</code>."
|
||||
[41]:https://rego.linux.conf.au/schedule/presentation/16/
|
||||
[42]:https://linux.conf.au/index.html
|
||||
[43]:http://shallowsky.com/
|
||||
@ -1,4 +1,4 @@
|
||||
SPARTA —— 用于网络渗透测试的 GUI 工具套件
|
||||
SPARTA：用于网络渗透测试的 GUI 工具套件
|
||||
======
|
||||
|
||||
|
||||
@ -7,12 +7,11 @@ SPARTA 是使用 Python 开发的 GUI 应用程序，它是 Kali Linux 内置的
|
||||
|
||||
SPARTA GUI 工具套件最擅长的事情是扫描和发现目标端口和运行的服务。
|
||||
|
||||
因此，作为枚举阶段的一部分功能，它提供对开放端口和服务的暴力攻击。
|
||||
|
||||
此外，作为枚举阶段的一部分功能，它提供对开放端口和服务的暴力攻击。
|
||||
|
||||
延伸阅读：[网络渗透检查清单][1]
|
||||
|
||||
## 安装
|
||||
### 安装
|
||||
|
||||
请从 GitHub 上克隆最新版本的 SPARTA：
|
||||
|
||||
@ -21,64 +20,58 @@ git clone https://github.com/secforce/sparta.git
|
||||
```
|
||||
|
||||
或者，从 [这里][2] 下载最新版本的 Zip 文件。
|
||||
|
||||
```
|
||||
cd /usr/share/
|
||||
git clone https://github.com/secforce/sparta.git
|
||||
```
|
||||
将 "sparta" 文件放到 /usr/bin/ 目录下并赋于可运行权限。
|
||||
|
||||
将 `sparta` 文件放到 `/usr/bin/` 目录下并赋于可运行权限。
|
||||
|
||||
在任意终端中输入 'sparta' 来启动应用程序。
|
||||
|
||||
### 网络渗透测试的范围
|
||||
|
||||
## 网络渗透测试的范围：
|
||||
|
||||
* 添加一个目标主机或者目标主机的列表到范围中，来发现一个组织的网络基础设备在安全方面的薄弱环节。
|
||||
* 选择菜单条 - File > Add host(s) to scope
|
||||
|
||||
添加一个目标主机或者目标主机的列表到测试范围中，来发现一个组织的网络基础设备在安全方面的薄弱环节。
|
||||
|
||||
选择菜单条 - “File” -> “Add host(s) to scope”
|
||||
|
||||
[![Network Penetration Testing][3]][4]
|
||||
|
||||
[![Network Penetration Testing][5]][6]
|
||||
|
||||
* 上图展示了在扫描范围中添加 IP 地址。根据你网络的具体情况，你可以添加一个 IP 地址的范围去扫描。
|
||||
* 扫描范围添加之后，Nmap 将开始扫描，并很快得到结果，扫描阶段结束。
|
||||
|
||||
|
||||
|
||||
## 打开 Ports & Services:
|
||||
|
||||
* Nmap 扫描结果提供了目标上开放的端口和服务。
|
||||
上图展示了在扫描范围中添加 IP 地址。根据你网络的具体情况，你可以添加一个 IP 地址的范围去扫描。
|
||||
扫描范围添加之后，Nmap 将开始扫描，并很快得到结果，扫描阶段结束。
|
||||
|
||||
### 打开的端口及服务
|
||||
|
||||
Nmap 扫描结果提供了目标上开放的端口和服务。
|
||||
|
||||
[![Network Penetration Testing][7]][8]
|
||||
|
||||
* 上图展示了扫描发现的目标操作系统、开发的端口和服务。
|
||||
|
||||
|
||||
|
||||
## 在开放端口上实施暴力攻击：
|
||||
|
||||
* 我们来通过 445 端口的服务器消息块（SMB）协议来暴力获取用户列表和它们的有效密码。
|
||||
上图展示了扫描发现的目标操作系统、开发的端口和服务。
|
||||
|
||||
### 在开放端口上实施暴力攻击
|
||||
|
||||
我们来通过 445 端口的服务器消息块（SMB）协议来暴力获取用户列表和它们的有效密码。
|
||||
|
||||
[![Network Penetration Testing][9]][10]
|
||||
|
||||
* 右键并选择 “Send to Brute” 选项。也可以选择发现的目标上的开放端口和服务。
|
||||
* 浏览和在用户名密码框中添加字典文件。
|
||||
|
||||
右键并选择 “Send to Brute” 选项。也可以选择发现的目标上的开放端口和服务。
|
||||
|
||||
浏览和在用户名密码框中添加字典文件。
|
||||
|
||||
[![Network Penetration Testing][11]][12]
|
||||
|
||||
* 点击 “Run” 去启动对目标的暴力攻击。上图展示了对目标 IP 地址进行的暴力攻击取得成功，找到了有效的密码。
|
||||
* 在 Windows 中失败的登陆尝试总是被记录到事件日志中。
|
||||
* 密码每 15 到 30 天改变一次的策略是非常好的一个实践经验。
|
||||
* 强烈建议使用强密码策略。密码锁定策略是阻止这种暴力攻击的最佳方法之一（ 5 次失败的登陆尝试之后将锁定帐户）
|
||||
* 将关键业务资产整合到 SIEM（ 安全冲突 & 事件管理）中将尽可能快地检测到这类攻击行为。
|
||||
点击 “Run” 去启动对目标的暴力攻击。上图展示了对目标 IP 地址进行的暴力攻击取得成功，找到了有效的密码。
|
||||
|
||||
在 Windows 中失败的登陆尝试总是被记录到事件日志中。
|
||||
|
||||
密码每 15 到 30 天改变一次的策略是非常好的一个实践经验。
|
||||
|
||||
强烈建议使用强密码策略。密码锁定策略是阻止这种暴力攻击的最佳方法之一（ 5 次失败的登录尝试之后将锁定帐户）。
|
||||
|
||||
将关键业务资产整合到 SIEM（ 安全冲突 & 事件管理）中将尽可能快地检测到这类攻击行为。
|
||||
|
||||
SPARTA 对渗透测试的扫描和枚举阶段来说是一个非常省时的 GUI 工具套件。SPARTA 可以扫描和暴力破解各种协议。它有许多的功能！祝你测试顺利！
|
||||
|
||||
@ -88,7 +81,7 @@ via: https://gbhackers.com/sparta-network-penetration-testing-gui-toolkit/
|
||||
|
||||
作者：[Balaganesh][a]
|
||||
译者：[qhwdw](https://github.com/qhwdw)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
校对：[wxy](https://github.com/wxy)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
@ -1,28 +1,28 @@
|
||||
为初学者介绍的 Linux tee 命令（6 个例子）
|
||||
======
|
||||
|
||||
有时候，你会想手动跟踪命令的输出内容，同时又想将输出的内容写入文件，确保之后可以用来参考。如果你想寻找这相关的工具，那么恭喜你，Linux 已经有了一个叫做 **tee** 的命令可以帮助你。
|
||||
有时候，你会想手动跟踪命令的输出内容，同时又想将输出的内容写入文件，确保之后可以用来参考。如果你想寻找这相关的工具，那么恭喜你，Linux 已经有了一个叫做 `tee` 的命令可以帮助你。
|
||||
|
||||
本教程中，我们将基于 tee 命令，用一些简单的例子开始讨论。但是在此之前，值得一提的是，本文我们所有的测试实例都基于 Ubuntu 16.04 LTS。
|
||||
本教程中，我们将基于 `tee` 命令，用一些简单的例子开始讨论。但是在此之前，值得一提的是，本文我们所有的测试实例都基于 Ubuntu 16.04 LTS。
|
||||
|
||||
### Linux tee 命令
|
||||
|
||||
tee 命令基于标准输入读取数据，标准输出或文件写入数据。感受下这个命令的语法：
|
||||
`tee` 命令基于标准输入读取数据，标准输出或文件写入数据。感受下这个命令的语法：
|
||||
|
||||
```
|
||||
tee [OPTION]... [FILE]...
|
||||
```
|
||||
|
||||
这里是帮助文档的说明：
|
||||
```
|
||||
从标准输入中复制到每一个文件，并输出到标准输出。
|
||||
```
|
||||
|
||||
> 从标准输入中复制到每一个文件，并输出到标准输出。
|
||||
|
||||
|
||||
让 Q&A（问&答）风格的实例给我们带来更多灵感，深入了解这个命令。
|
||||
|
||||
### Q1. 如何在 Linux 上使用这个命令？
|
||||
### Q1、 如何在 Linux 上使用这个命令？
|
||||
|
||||
假设因为某些原因，你正在使用 ping 命令。
|
||||
假设因为某些原因，你正在使用 `ping` 命令。
|
||||
|
||||
```
|
||||
ping google.com
|
||||
@ -30,29 +30,29 @@ ping google.com
|
||||
|
||||
[![如何在 Linux 上使用 tee 命令][1]][2]
|
||||
|
||||
然后同时，你想要输出的信息也同时能写入文件。这个时候，tee 命令就有其用武之地了。
|
||||
然后同时，你想要输出的信息也同时能写入文件。这个时候，`tee` 命令就有其用武之地了。
|
||||
|
||||
```
|
||||
ping google.com | tee output.txt
|
||||
```
|
||||
|
||||
下面的截图展示了这个输出内容不仅被写入 ‘output.txt’ 文件，也被显示在标准输出中。
|
||||
下面的截图展示了这个输出内容不仅被写入 `output.txt` 文件，也被显示在标准输出中。
|
||||
|
||||
[![tee command 输出][3]][4]
|
||||
|
||||
如此应当明确了 tee 的基础用法。
|
||||
如此应当明白了 `tee` 的基础用法。
|
||||
|
||||
### Q2. 如何确保 tee 命令追加信息到文件中？
|
||||
### Q2、 如何确保 tee 命令追加信息到文件中？
|
||||
|
||||
默认情况下，在同一个文件下再次使用 tee 命令会覆盖之前的信息。如果你想的话，可以通过 -a 命令选项改变默认设置。
|
||||
默认情况下，在同一个文件下再次使用 `tee` 命令会覆盖之前的信息。如果你想的话，可以通过 `-a` 命令选项改变默认设置。
|
||||
|
||||
```
|
||||
[command] | tee -a [file]
|
||||
```
|
||||
|
||||
基本上，-a 选项强制 tee 命令追加信息到文件。
|
||||
基本上，`-a` 选项强制 `tee` 命令追加信息到文件。
|
||||
|
||||
### Q3. 如何让 tee 写入多个文件？
|
||||
### Q3、 如何让 tee 写入多个文件？
|
||||
|
||||
这非常之简单。你仅仅只需要写明文件名即可。
|
||||
|
||||
@ -70,7 +70,7 @@ ping google.com | tee output1.txt output2.txt output3.txt
|
||||
|
||||
### Q4. 如何让 tee 命令的输出内容直接作为另一个命令的输入内容？
|
||||
|
||||
使用 tee 命令，你不仅可以将输出内容写入文件，还可以把输出内容作为另一个命令的输入内容。比如说，下面的命令不仅会将文件名存入‘output.txt’文件中，还会通过 wc 命令让你知道输入到 output.txt 中的文件数目。
|
||||
使用 `tee` 命令，你不仅可以将输出内容写入文件，还可以把输出内容作为另一个命令的输入内容。比如说，下面的命令不仅会将文件名存入 `output.txt` 文件中，还会通过 `wc` 命令让你知道输入到 `output.txt` 中的文件数目。
|
||||
|
||||
```
|
||||
ls file* | tee output.txt | wc -l
|
||||
@ -80,11 +80,11 @@ ls file* | tee output.txt | wc -l
|
||||
|
||||
### Q5. 如何使用 tee 命令提升文件写入权限？
|
||||
|
||||
假如你使用 [Vim editor][9] 打开文件，并且做了很多更改，然后当你尝试保存修改时，你得到一个报错，让你意识到那是一个 root 所拥有的文件，这意味着你需要使用 sudo 权限保存修改。
|
||||
假如你使用 [Vim 编辑器][9] 打开文件，并且做了很多更改，然后当你尝试保存修改时，你得到一个报错，让你意识到那是一个 root 所拥有的文件，这意味着你需要使用 `sudo` 权限保存修改。
|
||||
|
||||
[![如何使用 tee 命令提升文件写入权限][10]][11]
|
||||
|
||||
如此情况下，你可以使用 tee 命令来提高权限。
|
||||
如此情况下，你可以（在 Vim 内）使用 `tee` 命令来提高权限。
|
||||
|
||||
```
|
||||
:w !sudo tee %
|
||||
@ -94,17 +94,17 @@ ls file* | tee output.txt | wc -l
|
||||
|
||||
### Q6. 如何让 tee 命令忽视中断？
|
||||
|
||||
-i 命令行选项使 tee 命令忽视通常由 crl+c 组合键发起的中断信号（`SIGINT`）。
|
||||
`-i` 命令行选项使 `tee` 命令忽视通常由 `ctrl+c` 组合键发起的中断信号（`SIGINT`）。
|
||||
|
||||
```
|
||||
[command] | tee -i [file]
|
||||
```
|
||||
|
||||
当你想要使用 crl+c 中断命令的同时，让 tee 命令优雅的退出，这个选项尤为实用。
|
||||
当你想要使用 `ctrl+c` 中断该命令，同时让 `tee` 命令优雅的退出，这个选项尤为实用。
|
||||
|
||||
### 总结
|
||||
|
||||
现在你可能已经认同 tee 是一个非常实用的命令。基于 tee 命令的用法，我们已经介绍了其绝大多数的命令行选项。这个工具并没有什么陡峭的学习曲线，所以，只需跟随这几个例子练习，你就可以运用自如了。更多信息，请查看 [帮助文档][12].
|
||||
现在你可能已经认同 `tee` 是一个非常实用的命令。基于 `tee` 命令的用法，我们已经介绍了其绝大多数的命令行选项。这个工具并没有什么陡峭的学习曲线，所以，只需跟随这几个例子练习，你就可以运用自如了。更多信息，请查看 [帮助文档][12].
|
||||
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
@ -113,7 +113,7 @@ via: https://www.howtoforge.com/linux-tee-command/
|
||||
|
||||
作者：[Himanshu Arora][a]
|
||||
译者：[CYLeft](https://github.com/CYLeft)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
校对：[wxy](https://github.com/wxy)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
@ -1,53 +0,0 @@
|
||||
translating---geekpi
|
||||
|
||||
An old DOS BBS in a Docker container
|
||||
======
|
||||
Awhile back, I wrote about [my Debian Docker base images][1]. I decided to extend this concept a bit further: to running DOS applications in Docker.
|
||||
|
||||
But first, a screenshot:
|
||||
|
||||
![][2]
|
||||
|
||||
It turns out this is possible, but difficult. I went through all three major DOS emulators available (dosbox, qemu, and dosemu). I got them all running inside the Docker container, but had a number of, er, fun issues to resolve.
|
||||
|
||||
The general thing one has to do here is present a fake modem to the DOS environment. This needs to be exposed outside the container as a TCP port. That much is possible in various ways -- I wound up using tcpser. dosbox had a TCP modem interface, but it turned out to be too buggy for this purpose.
|
||||
|
||||
The challenge comes in where you want to be able to accept more than one incoming telnet (or TCP) connection at a time. DOS was not a multitasking operating system, so there were any number of hackish solutions back then. One might have had multiple physical computers, one for each incoming phone line. Or they might have run multiple pseudo-DOS instances under a multitasking layer like [DESQview][3], OS/2, or even Windows 3.1.
|
||||
|
||||
(Side note: I just learned of [DESQview/X][4], which integrated DESQview with X11R5 and [replaced the Windows 3 drivers][5] to allow running Windows as an X application).
|
||||
|
||||
For various reasons, I didn't want to try running one of those systems inside Docker. That left me with emulating the original multiple physical node setup. In theory, pretty easy -- spin up a bunch of DOS boxes, each using at most 1MB of emulated RAM, and go to town. But here came the challenge.
|
||||
|
||||
In a multiple-physical-node setup, you need some sort of file sharing, because your nodes have to access the shared message and file store. There were a myriad of clunky ways to do this in the old DOS days - [Netware][6], [LAN manager][7], even some PC NFS clients. I didn't have access to Netware. I tried the Microsoft LM client in DOS, talking to a Samba server running inside the Docker container. This I got working, but the LM client used so much RAM that, even with various high memory tricks, BBS software wasn't going to run. I couldn't just mount an underlying filesystem in multiple dosbox instances either, because dosbox did caching that wasn't going to be compatible.
|
||||
|
||||
This is why I wound up using dosemu. Besides being a more complete emulator than dosbox, it had a way of sharing the host's filesystems that was going to work.
|
||||
|
||||
So, all of this wound up with this: [jgoerzen/docker-bbs-renegade][8].
|
||||
|
||||
I also prepared building blocks for others that want to do something similar: [docker-dos-bbs][9] and the lower-level [docker-dosemu][10].
|
||||
|
||||
As a side bonus, I also attempted running this under Joyent's Triton (SmartOS, Solaris-based). I was pleasantly impressed that I got it all almost working there. So yes, a Renegade DOS BBS running under a Linux-based DOS emulator in a container on a Solaris machine.
|
||||
|
||||
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: http://changelog.complete.org/archives/9836-an-old-dos-bbs-in-a-docker-container
|
||||
|
||||
作者：[John Goerzen][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:http://changelog.complete.org/archives/author/jgoerzen
|
||||
[1]:https://changelog.complete.org/archives/9794-fixing-the-problems-with-docker-images
|
||||
[2]:https://raw.githubusercontent.com/jgoerzen/docker-bbs-renegade/master/renegade-login.png
|
||||
[3]:https://en.wikipedia.org/wiki/DESQview
|
||||
[4]:http://toastytech.com/guis/dvx.html
|
||||
[5]:http://toastytech.com/guis/dvx3.html
|
||||
[6]:https://en.wikipedia.org/wiki/NetWare
|
||||
[7]:https://en.wikipedia.org/wiki/LAN_Manager
|
||||
[8]:https://github.com/jgoerzen/docker-bbs-renegade
|
||||
[9]:https://github.com/jgoerzen/docker-dos-bbs
|
||||
[10]:https://github.com/jgoerzen/docker-dosemu
|
||||
@ -0,0 +1,82 @@
|
||||
Testing IPv6 Networking in KVM: Part 1
|
||||
======
|
||||
|
||||
|
||||
|
||||
Nothing beats hands-on playing with IPv6 addresses to get the hang of how they work, and setting up a little test lab in KVM is as easy as falling over — and more fun. In this two-part series, we will learn about IPv6 private addressing and configuring test networks in KVM.
|
||||
|
||||
### QEMU/KVM/Virtual Machine Manager
|
||||
|
||||
Let's start with understanding what KVM is. Here I use KVM as a convenient shorthand for the combination of QEMU, KVM, and the Virtual Machine Manager that is typically bundled together in Linux distributions. The simplified explanation is that QEMU emulates hardware, and KVM is a kernel module that creates the guest state on your CPU and manages access to memory and the CPU. Virtual Machine Manager is a lovely graphical overlay to all of this virtualization and hypervisor goodness.
|
||||
|
||||
But you're not stuck with pointy-clicky, no, for there are also fab command-line tools to use — such as virsh and virt-install.
|
||||
|
||||
If you're not experienced with KVM, you might want to start with [Creating Virtual Machines in KVM: Part 1][1] and [Creating Virtual Machines in KVM: Part 2 - Networking][2].
|
||||
|
||||
### IPv6 Unique Local Addresses
|
||||
|
||||
Configuring IPv6 networking in KVM is just like configuring IPv4 networks. The main difference is those weird long addresses. [Last time][3], we talked about the different types of IPv6 addresses. There is one more IPv6 unicast address class, and that is unique local addresses, fc00::/7 (see [RFC 4193][4]). This is analogous to the private address classes in IPv4, 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16.
|
||||
|
||||
This diagram illustrates the structure of the unique local address space. 48 bits define the prefix and global ID, 16 bits are for subnets, and the remaining 64 bits are the interface ID:
|
||||
```
|
||||
| 7 bits |1| 40 bits | 16 bits | 64 bits |
|
||||
+--------|-+------------|-----------|----------------------------+
|
||||
| Prefix |L| Global ID | Subnet ID | Interface ID |
|
||||
+--------|-+------------|-----------|----------------------------+
|
||||
|
||||
```
|
||||
|
||||
Here is another way to look at it, which is might be more helpful for understanding how to manipulate these addresses:
|
||||
```
|
||||
| Prefix | Global ID | Subnet ID | Interface ID |
|
||||
+--------|--------------|-------------|----------------------+
|
||||
| fd | 00:0000:0000 | 0000 | 0000:0000:0000:0000 |
|
||||
+--------|--------------|-------------|----------------------+
|
||||
|
||||
```
|
||||
|
||||
fc00::/7 is divided into two /8 blocks, fc00::/8 and fd00::/8. fc00::/8 is reserved for future use. So, unique local addresses always start with fd, and the rest is up to you. The L bit, which is the eighth bit, is always set to 1, which makes fd00::/8. Setting it to zero makes fc00::/8. You can see this with subnetcalc:
|
||||
```
|
||||
$ subnetcalc fd00::/8 -n
|
||||
Address = fd00::
|
||||
fd00 = 11111101 00000000
|
||||
|
||||
$ subnetcalc fc00::/8 -n
|
||||
Address = fc00::
|
||||
fc00 = 11111100 00000000
|
||||
|
||||
```
|
||||
|
||||
RFC 4193 requires that addresses be randomly generated. You can invent addresses any way you choose, as long as they start with fd, because the IPv6 cops aren't going to invade your home and give you a hard time. Still, it is a best practice to follow what RFCs say. The addresses must not be assigned sequentially or with well-known numbers. RFC 4193 includes an algorithm for building a pseudo-random address generator, or you can find any number of generators online.
|
||||
|
||||
Unique local addresses are not centrally managed like global unicast addresses (assigned to you by your Internet service provider), but even so the probability of address collisions is very low. This is a nice benefit when you need to merge some local networks or want to route between discrete private networks.
|
||||
|
||||
You can mix unique local addresses and global unicast addresses on the same subnets. Unique local addresses are routable and require no extra router tweaks. However, you should configure your border routers and firewalls to not allow them to leave your network except between private networks at different locations.
|
||||
|
||||
RFC 4193 advises against mingling AAAA and PTR records with your global unicast address records, because there is no guarantee that they will be unique, even though the odds of duplicates are low. Just like we do with IPv4 addresses, keep your private local name services and public name services separate. The tried-and-true combination of Dnsmasq for local name services and BIND for public name services works just as well for IPv6 as it does for IPv4.
|
||||
|
||||
### Pseudo-Random Address Generator
|
||||
|
||||
One example of an online address generator is [Local IPv6 Address Generator][5]. You can find many cool online tools like this. You can use it to create a new address for you, or use it with your existing global ID and play with creating subnets.
|
||||
|
||||
Come back next week to learn how to plug all of this IPv6 goodness into KVM and do live testing.
|
||||
|
||||
Learn more about Linux through the free ["Introduction to Linux" ][6]course from The Linux Foundation and edX.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/learn/intro-to-linux/2017/11/testing-ipv6-networking-kvm-part-1
|
||||
|
||||
作者：[Carla Schroder][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/cschroder
|
||||
[1]:https://www.linux.com/learn/intro-to-linux/2017/5/creating-virtual-machines-kvm-part-1
|
||||
[2]:https://www.linux.com/learn/intro-to-linux/2017/5/creating-virtual-machines-kvm-part-2-networking
|
||||
[3]:https://www.linux.com/learn/intro-to-linux/2017/10/calculating-ipv6-subnets-linux
|
||||
[4]:https://tools.ietf.org/html/rfc4193
|
||||
[5]:https://www.ultratools.com/tools/rangeGenerator
|
||||
[6]:https://training.linuxfoundation.org/linux-courses/system-administration-training/introduction-to-linux
|
||||
@ -1,45 +0,0 @@
|
||||
translating---geekpi
|
||||
|
||||
How to create better documentation with a kanban board
|
||||
======
|
||||
|
||||
If you're working on documentation, a website, or other user-facing content, it's helpful to know what users expect to find--both the information they want and how the information is organized and structured. After all, great content isn't very useful if people can't find what they're looking for.
|
||||
|
||||
Card sorting is a simple and effective way to gather input from users about what they expect from menu interfaces and pages. The simplest implementation is to label a stack of index cards with the sections you plan to include in your website or documentation and ask users to sort the cards in the way they would look for the information. Variations include letting people write their own menu headers or content elements.
|
||||
|
||||
The goal is to learn what your users expect and where they expect to find it, rather than having to figure out your menu and layout on your own. This is relatively straightforward when you have users in the same physical location, but it's more challenging when you are trying to get feedback from people in many locations.
|
||||
|
||||
I've found [kanban][1] boards are a great tool for these situations. They allow people to easily drag virtual cards around to categorize and rank them, and they are multi-purpose, unlike dedicated card-sorting software.
|
||||
|
||||
I often use Trello for card sorting, but there are several [open source alternatives][2] that you might want to try.
|
||||
|
||||
### How it works
|
||||
|
||||
My most successful kanban experiment was when I was working on documentation for [Gluster][3], a free and open source scalable network-attached storage filesystem. I needed to take a large pile of documentation that had grown over time and break it into categories to create a navigation system. BEcause I didn't have the technical knowledge necessary to sort it, I turned to the Gluster team and developer community for guidance.
|
||||
|
||||
First, I created a shared Kanban board. I gave the columns general names that would enable sorting and created cards for all the topics I planned to cover in the documentation. I flagged some cards with different colors to indicate either a topic was missing and needed to be created, or it was present and needed to be removed. Then I put all the cards into an "unsorted" column and asked people to drag them where they thought the cards should be organized and send me a screen capture of what they thought was the ideal state.
|
||||
|
||||
Dealing with all the screen captures was the trickiest part. I wish there was a merge or consensus feature that would've helped me aggregate everyone's data, rather than having to examine a bunch of screen captures. Fortunately, after the first person sorted the cards, people more or less agreed on the structure and made only minor modifications. When opinions differed on a topic's placement, I set up flash meetings where people could explain their thinking and we could hash out the disagreements.
|
||||
|
||||
### Using the data
|
||||
|
||||
From here, it was easy to convert the information I captured into menus and refine it. If users thought items should become submenus, they usually told me in comments or when we talked on the phone. Perceptions of menu organization vary depending upon people's job tasks, so you never have complete agreement, but testing with users means you won't have as many blind spots about what people use and where they will look for it.
|
||||
|
||||
Pairing card sorting with analytics gives you even more insight on what people are looking for. Once, when I ran analytics on some training documentation I was working on, I was surprised that to learn that the most searched page was about title capitalization. So I surfaced that page at the top-menu level, even though my "logical" setting put it far down in a sub-menu.
|
||||
|
||||
I've found kanban card-sorting a great way to help me create content that users want to see and put it where they expect to find it. Have you found another great way to organize your content for users' benefit? Or another interesting use for kanban boards? If so, please share your thoughts in the comments.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/17/11/kanban-boards-card-sorting
|
||||
|
||||
作者：[Heidi Waterhouse][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/hwaterhouse
|
||||
[1]:https://en.wikipedia.org/wiki/Kanban
|
||||
[2]:https://opensource.com/alternatives/trello
|
||||
[3]:https://www.gluster.org/
|
||||
@ -1,214 +0,0 @@
|
||||
translating by Flowsnow
|
||||
|
||||
Parsing HTML with Python
|
||||
======
|
||||
|
||||
|
||||
|
||||
Image by : Jason Baker for Opensource.com.
|
||||
|
||||
As a long-time member of the documentation team at Scribus, I keep up-to-date with the latest updates of the source so I can help make updates and additions to the documentation. When I recently did a "checkout" using Subversion on a computer I had just upgraded to Fedora 27, I was amazed at how long it took to download the documentation, which consists of HTML pages and associated images. I became concerned that the project's documentation seemed much larger than it should be and suspected that some of the content was "zombie" documentation--HTML files that aren't used anymore and images that have lost all references in the currently used HTML.
|
||||
|
||||
I decided to create a project for myself to figure this out. One way to do this is to search for existing image files that aren't used. If I could scan through all the HTML files for image references, then compare that list to the actual image files, chances are I would see a mismatch.
|
||||
|
||||
Here is a typical image tag:
|
||||
```
|
||||
<img src="images/edit_shapes.png" ALT="Edit examples" ALIGN=left>
|
||||
```
|
||||
|
||||
I'm interested in the part between the first set of quotation marks, after `src=`. After some searching for a solution, I found a Python module called [BeautifulSoup][1]. The tasty part of the script I wrote looks like this:
|
||||
```
|
||||
soup = BeautifulSoup(all_text, 'html.parser')
|
||||
match = soup.findAll("img")
|
||||
if len(match) > 0:
|
||||
for m in match:
|
||||
imagelist.append(str(m))
|
||||
```
|
||||
|
||||
We can use this `findAll` method to pluck out the image tags. Here is a tiny piece of the output:
|
||||
```
|
||||
|
||||
|
||||
<img src="images/pdf-form-ht3.png"/><img src="images/pdf-form-ht4.png"/><img src="images/pdf-form-ht5.png"/><img src="images/pdf-form-ht6.png"/><img align="middle" alt="GSview - Advanced Options Panel" src="images/gsadv1.png" title="GSview - Advanced Options Panel"/><img align="middle" alt="Scribus External Tools Preferences" src="images/gsadv2.png" title="Scribus External Tools Preferences"/>
|
||||
```
|
||||
|
||||
So far, so good. I thought that the next step might be to just carve this down, but when I tried some string methods in the script, it returned errors about this being tags and not strings. I saved the output to a file and went through the process of editing in [KWrite][2]. One nice thing about KWrite is that you can do a "find & replace" using regular expressions (regex), so I could replace `<img` with `\n<img`, which made it easier to see how to carve this down from there. Another nice thing with KWrite is that, if you make an injudicious choice with regex, you can undo it.
|
||||
|
||||
But I thought, surely there is something better than this, so I turned to regex, or more specifically the `re` module for Python. The relevant part of this new script looks like this:
|
||||
```
|
||||
match = re.findall(r'src="(.*)/>', all_text)
|
||||
if len(match)>0:
|
||||
for m in match:
|
||||
imagelist.append(m)
|
||||
```
|
||||
|
||||
And a tiny piece of its output looks like this:
|
||||
```
|
||||
images/cmcanvas.png" title="Context Menu for the document canvas" alt="Context Menu for the document canvas" /></td></tr></table><br images/eps-imp1.png" title="EPS preview in a file dialog" alt="EPS preview in a file dialog" images/eps-imp5.png" title="Colors imported from an EPS file" alt="Colors imported from an EPS file" images/eps-imp4.png" title="EPS font substitution" alt="EPS font substitution" images/eps-imp2.png" title="EPS import progress" alt="EPS import progress" images/eps-imp3.png" title="Bitmap conversion failure" alt="Bitmap conversion failure"
|
||||
```
|
||||
|
||||
At first glance, it looks similar to the output above, and has the nice feature of trimming out parts of the image tag, but there are puzzling inclusions of table tags and other content. I think this relates to this regex expression `src="(.*)/>`, which is termed greedy, meaning it doesn't necessarily stop at the first instance of `/>` it encounters. I should add that I also tried `src="(.*)"` which was really no better. Not being a regexpert (just made this up), my searching around for various ideas to improve this didn't help.
|
||||
|
||||
After a series of other things, even trying out `HTML::Parser` with Perl, I finally tried to compare this to the situation of some scripts that I wrote for Scribus that analyze the contents of a text frame, character by character, then take some action. For my purposes, what I finally came up with improves on all these methods and requires no regex or HTML parser at all. Let's go back to that example `img` tag I showed.
|
||||
```
|
||||
<img src="images/edit_shapes.png" ALT="Edit examples" ALIGN=left>
|
||||
```
|
||||
|
||||
I decided to home in on the `src=` piece. One way would be to wait for an occurrence of `s`, then see if the next character is `r`, the next `c`, and the next `=`. If so, bingo! Then what follows between two sets of double quotation marks is what I need. The problem with this is the structure it takes to hang onto these. One way of looking at a string of characters representing a line of HTML text would be:
|
||||
```
|
||||
for c in all_text:
|
||||
```
|
||||
|
||||
But the logic was just too messy to hang onto the previous `c`, and the one before that, the one before that, and the one before that.
|
||||
|
||||
In the end, I decided to focus on the `=` and to use an indexing method whereby I could easily reference any prior or future character in the string. Here is the searching part:
|
||||
```
|
||||
index = 3
|
||||
while index < linelength:
|
||||
if (all_text[index] == '='):
|
||||
if (all_text[index-3] == 's') and (all_text[index-2] == 'r') and (all_text[index-1] == 'c'):
|
||||
imagefound(all_text, imagelist, index)
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
```
|
||||
|
||||
I start the search with the fourth character (indexing starts at 0), so I don't get an indexing error down below, and realistically, there will not be an equal sign before the fourth character of a line. The first test is to see if we find `=` as we're marching through the string, and if not, we march on. If we do see one, then we ask if the three previous characters were `s`, `r`, and `c`, in that order. If that happens, we call the function `imagefound`:
|
||||
```
|
||||
def imagefound(all_text, imagelist, index):
|
||||
end = 0
|
||||
index += 2
|
||||
newimage = ''
|
||||
while end == 0:
|
||||
if (all_text[index] != '"'):
|
||||
newimage = newimage + all_text[index]
|
||||
index += 1
|
||||
else:
|
||||
newimage = newimage + '\n'
|
||||
imagelist.append(newimage)
|
||||
end = 1
|
||||
return
|
||||
```
|
||||
|
||||
We're sending the function the current index, which represents the `=`. We know the next character will be `"`, so we jump two characters and begin adding characters to a holding string named `newimage`, until we reach the following `"`, at which point we're done. We add the string plus a `newline` character to our list `imagelist` and `return`, keeping in mind there may be more image tags in this remaining string of HTML, so we're right back in the middle of our searching loop.
|
||||
|
||||
Here's what our output looks like now:
|
||||
```
|
||||
images/text-frame-link.png
|
||||
images/text-frame-unlink.png
|
||||
images/gimpoptions1.png
|
||||
images/gimpoptions3.png
|
||||
images/gimpoptions2.png
|
||||
images/fontpref3.png
|
||||
images/font-subst.png
|
||||
images/fontpref2.png
|
||||
images/fontpref1.png
|
||||
images/dtp-studio.png
|
||||
```
|
||||
|
||||
Ahhh, much cleaner, and this only took a few seconds to run. I could have jumped seven more index spots to cut out the `images/` part, but I like having it there to make sure I haven't chopped off the first letter of the image filename, and this is so easy to edit out with KWrite--you don't even need regex. After doing that and saving the file, the next step was to run another script I wrote called `sortlist.py`:
|
||||
```
|
||||
#!/usr/bin/env python
|
||||
# -*- coding: utf-8 -*-
|
||||
# sortlist.py
|
||||
|
||||
import os
|
||||
|
||||
imagelist = []
|
||||
for line in open('/tmp/imagelist_parse4.txt').xreadlines():
|
||||
imagelist.append(line)
|
||||
|
||||
imagelist.sort()
|
||||
|
||||
outfile = open('/tmp/imagelist_parse4_sorted.txt', 'w')
|
||||
outfile.writelines(imagelist)
|
||||
outfile.close()
|
||||
```
|
||||
|
||||
This pulls in the file contents as a list, sorts it, then saves it as another file. After that I could just do the following:
|
||||
```
|
||||
ls /home/gregp/development/Scribus15x/doc/en/images/*.png > '/tmp/actual_images.txt'
|
||||
```
|
||||
|
||||
Then I need to run `sortlist.py` on that file too, since the method `ls` uses to sort is different from Python. I could have run a comparison script on these files, but I preferred to do this visually. In the end, I ended up with 42 images that had no HTML reference from the documentation.
|
||||
|
||||
Here is my parsing script in its entirety:
|
||||
```
|
||||
#!/usr/bin/env python
|
||||
# -*- coding: utf-8 -*-
|
||||
# parseimg4.py
|
||||
|
||||
import os
|
||||
|
||||
def imagefound(all_text, imagelist, index):
|
||||
end = 0
|
||||
index += 2
|
||||
newimage = ''
|
||||
while end == 0:
|
||||
if (all_text[index] != '"'):
|
||||
newimage = newimage + all_text[index]
|
||||
index += 1
|
||||
else:
|
||||
newimage = newimage + '\n'
|
||||
imagelist.append(newimage)
|
||||
end = 1
|
||||
return
|
||||
|
||||
htmlnames = []
|
||||
imagelist = []
|
||||
tempstring = ''
|
||||
filenames = os.listdir('/home/gregp/development/Scribus15x/doc/en/')
|
||||
for name in filenames:
|
||||
if name.endswith('.html'):
|
||||
htmlnames.append(name)
|
||||
#print htmlnames
|
||||
for htmlfile in htmlnames:
|
||||
all_text = open('/home/gregp/development/Scribus15x/doc/en/' + htmlfile).read()
|
||||
linelength = len(all_text)
|
||||
index = 3
|
||||
while index < linelength:
|
||||
if (all_text[index] == '='):
|
||||
if (all_text[index-3] == 's') and (all_text[index-2] == 'r') and
|
||||
(all_text[index-1] == 'c'):
|
||||
imagefound(all_text, imagelist, index)
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
|
||||
outfile = open('/tmp/imagelist_parse4.txt', 'w')
|
||||
outfile.writelines(imagelist)
|
||||
outfile.close()
|
||||
imageno = len(imagelist)
|
||||
print str(imageno) + " images were found and saved"
|
||||
```
|
||||
|
||||
Its name, `parseimg4.py`, doesn't really reflect the number of scripts I wrote along the way, with both minor and major rewrites, plus discards and starting over. Notice that I've hardcoded these directory and filenames, but it would be easy enough to generalize, asking for user input for these pieces of information. Also as they were working scripts, I sent the output to `/tmp`, so they disappear once I reboot my system.
|
||||
|
||||
This wasn't the end of the story, since the next question was: What about zombie HTML files? Any of these files that are not used might reference images not picked up by the previous method. We have a `menu.xml` file that serves as the table of contents for the online manual, but I also needed to consider that some files listed in the TOC might reference files not in the TOC, and yes, I did find some.
|
||||
|
||||
I'll conclude by saying that this was a simpler task than this image search, and it was greatly helped by the processes I had already developed.
|
||||
|
||||
|
||||
### About the author
|
||||
|
||||
[][7] Greg Pittman - Greg is a retired neurologist in Louisville, Kentucky, with a long-standing interest in computers and programming, beginning with Fortran IV in the 1960s. When Linux and open source software came along, it kindled a commitment to learning more, and eventually contributing. He is a member of the Scribus Team.[More about me][8]
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/1/parsing-html-python
|
||||
|
||||
作者：[Greg Pittman][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/greg-p
|
||||
[1]:https://www.crummy.com/software/BeautifulSoup/
|
||||
[2]:https://www.kde.org/applications/utilities/kwrite/
|
||||
[7]:https://opensource.com/users/greg-p
|
||||
[8]:https://opensource.com/users/greg-p
|
||||
@ -1,3 +1,5 @@
|
||||
translating----geekpi
|
||||
|
||||
Install AWFFull web server log analysis application on ubuntu 17.10
|
||||
======
|
||||
|
||||
|
||||
@ -0,0 +1,78 @@
|
||||
3 open source tools for scientific publishing
|
||||
======
|
||||
|
||||
|
||||
One industry that lags behind others in the adoption of digital or open source tools is the competitive and lucrative world of scientific publishing. Worth over £19B ($26B) annually, according to figures published by Stephen Buranyi in [The Guardian][1] last year, the system for selecting, publishing, and sharing even the most important scientific research today still bears many of the constraints of print media. New digital-era technologies present a huge opportunity to accelerate discovery, make science collaborative instead of competitive, and redirect investments from infrastructure development into research that benefits society.
|
||||
|
||||
The non-profit [eLife initiative][2] was established by the funders of research, in part to encourage the use of these technologies to this end. In addition to publishing an open-access journal for important advances in life science and biomedical research, eLife has made itself into a platform for experimentation and showcasing innovation in research communication—with most of this experimentation based around the open source ethos.
|
||||
|
||||
Working on open publishing infrastructure projects gives us the opportunity to accelerate the reach and adoption of the types of technology and user experience (UX) best practices that we consider important to the advancement of the academic publishing industry. Speaking very generally, the UX of open source products is often left undeveloped, which can in some cases dissuade people from using it. As part of our investment in OSS development, we place a strong emphasis on UX in order to encourage users to adopt these products.
|
||||
|
||||
All of our code is open source, and we actively encourage community involvement in our projects, which to us means faster iteration, more experimentation, greater transparency, and increased reach for our work.
|
||||
|
||||
The projects that we are involved in, such as the development of Libero (formerly known as [eLife Continuum][3]) and the [Reproducible Document Stack][4], along with our recent collaboration with [Hypothesis][5], show how OSS can be used to bring about positive changes in the assessment, publication, and communication of new discoveries.
|
||||
|
||||
### Libero
|
||||
|
||||
Libero is a suite of services and applications available to publishers that includes a post-production publishing system, a full front-end user interface pattern suite, Libero's Lens Reader, an open API, and search and recommendation engines.
|
||||
|
||||
Last year, we took a user-driven approach to redesigning the front end of Libero, resulting in less distracting site “furniture” and a greater focus on research articles. We tested and iterated all the key functional areas of the site with members of the eLife community to ensure the best possible reading experience for everyone. The site’s new API also provides simpler access to content for machine readability, including text mining, machine learning, and online application development.
|
||||
|
||||
The content on our website and the patterns that drive the new design are all open source to encourage future product development for both eLife and other publishers that wish to use it.
|
||||
|
||||
### The Reproducible Document Stack
|
||||
|
||||
In collaboration with [Substance][6] and [Stencila][7], eLife is also engaged in a project to create a Reproducible Document Stack (RDS)—an open stack of tools for authoring, compiling, and publishing computationally reproducible manuscripts online.
|
||||
|
||||
Today, an increasing number of researchers are able to document their computational experiments through languages such as [R Markdown][8] and [Python][9]. These can serve as important parts of the experimental record, and while they can be shared independently from or alongside the resulting research article, traditional publishing workflows tend to relegate these assets as a secondary class of content. To publish papers, researchers using these languages often have little option but to submit their computational results as “flattened” outputs in the form of figures, losing much of the value and reusability of the code and data references used in the computation. And while electronic notebook solutions such as [Jupyter][10] can enable researchers to publish their code in an easily reusable and executable form, that’s still in addition to, rather than as an integral part of, the published manuscript.
|
||||
|
||||
The [Reproducible Document Stack][11] project aims to address these challenges through development and publication of a working prototype of a reproducible manuscript that treats code and data as integral parts of the document, demonstrating a complete end-to-end technology stack from authoring through to publication. It will ultimately allow authors to submit their manuscripts in a format that includes embedded code blocks and computed outputs (statistical results, tables, or graphs), and have those assets remain both visible and executable throughout the publication process. Publishers will then be able to preserve these assets directly as integral parts of the published online article.
|
||||
|
||||
### Open annotation with Hypothesis
|
||||
|
||||
Most recently, we introduced open annotation in collaboration with [Hypothesis][12] to enable users of our website to make comments, highlight important sections of articles, and engage with the reading public online.
|
||||
|
||||
Through this collaboration, the open source Hypothesis software was customized with new moderation features, single sign-on authentication, and user-interface customization options, giving publishers more control over its implementation on their sites. These enhancements are already driving higher-quality discussions around published scholarly content.
|
||||
|
||||
The tool can be integrated seamlessly into publishers’ websites, with the scholarly publishing platform [PubFactory][13] and content solutions provider [Ingenta][14] already taking advantage of its improved feature set. [HighWire][15] and [Silverchair][16] are also offering their publishers the opportunity to implement the service.
|
||||
|
||||
### Other industries and open source
|
||||
|
||||
Over time, we hope to see more publishers adopt Hypothesis, Libero, and other projects to help them foster the discovery and reuse of important scientific research. But the opportunities for innovation eLife has been able to leverage because of these and other OSS technologies are also prevalent in other industries.
|
||||
|
||||
The world of data science would be nowhere without the high-quality, well-supported open source software and the communities built around it; [TensorFlow][17] is a leading example of this. Thanks to OSS and its communities, all areas of AI and machine learning have seen rapid acceleration and advancement compared to other areas of computing. Similar is the explosion in usage of Linux as a cloud web host, followed by containerization with Docker, and now the growth of Kubernetes, one of the most popular open source projects on GitHub.
|
||||
|
||||
All of these technologies enable organizations to do more with less and focus on innovation instead of reinventing the wheel. And in the end, that’s the real benefit of OSS: It lets us all learn from each other’s failures while building on each other's successes.
|
||||
|
||||
We are always on the lookout for opportunities to engage with the best emerging talent and ideas at the interface of research and technology. Find out more about some of these engagements on [eLife Labs][18], or contact [innovation@elifesciences.org][19] for more information.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/3/scientific-publishing-software
|
||||
|
||||
作者：[Paul Shanno][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/pshannon
|
||||
[1]:https://www.theguardian.com/science/2017/jun/27/profitable-business-scientific-publishing-bad-for-science
|
||||
[2]:https://elifesciences.org/about
|
||||
[3]:https://elifesciences.org/inside-elife/33e4127f/elife-introduces-continuum-a-new-open-source-tool-for-publishing
|
||||
[4]:https://elifesciences.org/for-the-press/e6038800/elife-supports-development-of-open-technology-stack-for-publishing-reproducible-manuscripts-online
|
||||
[5]:https://elifesciences.org/for-the-press/81d42f7d/elife-enhances-open-annotation-with-hypothesis-to-promote-scientific-discussion-online
|
||||
[6]:https://github.com/substance
|
||||
[7]:https://github.com/stencila/stencila
|
||||
[8]:https://rmarkdown.rstudio.com/
|
||||
[9]:https://www.python.org/
|
||||
[10]:http://jupyter.org/
|
||||
[11]:https://elifesciences.org/labs/7dbeb390/reproducible-document-stack-supporting-the-next-generation-research-article
|
||||
[12]:https://github.com/hypothesis
|
||||
[13]:http://www.pubfactory.com/
|
||||
[14]:http://www.ingenta.com/
|
||||
[15]:https://github.com/highwire
|
||||
[16]:https://www.silverchair.com/community/silverchair-universe/hypothesis/
|
||||
[17]:https://www.tensorflow.org/
|
||||
[18]:https://elifesciences.org/labs
|
||||
[19]:mailto:innovation@elifesciences.org
|
||||
@ -0,0 +1,92 @@
|
||||
translating---geekpi
|
||||
|
||||
An Open Source Desktop YouTube Player For Privacy-minded People
|
||||
======
|
||||
|
||||
|
||||
|
||||
You already know that we need a Google account to subscribe channels and download videos from YouTube. If you don’t want Google track what you’re doing on YouTube, well, there is an open source YouTube player named **“FreeTube”**. It allows you to watch, search and download Youtube videos and subscribe your favorite channels without an account, which prevents Google from having your information. It gives you complete ad-free experience. Another notable advantage is it has a built-in basic HTML5 player to watch videos. Since we’re not using the built-in YouTube player, Google can’t track the “views” and the video analytics either. FreeTube only sends your IP details, but this also can be overcome by using a VPN. It is completely free, open source and available for GNU/Linux, Mac OS X, and Windows.
|
||||
|
||||
### Features
|
||||
|
||||
* Watch videos without ads.
|
||||
* Prevent Google from tracking what you watch using cookies or JavaScript.
|
||||
* Subscribe to channels without an account.
|
||||
* Store subscriptions, history, and saved videos locally.
|
||||
* Import / Backup subscriptions.
|
||||
* Mini Player.
|
||||
* Light / Dark Theme.
|
||||
* Free, Open Source.
|
||||
* Cross-platform.
|
||||
|
||||
|
||||
|
||||
### Installing FreeTube
|
||||
|
||||
Go to the [**releases page**][1] and grab the version depending upon the OS you use. For the purpose of this guide, I will be using **.tar.gz** file.
|
||||
```
|
||||
$ wget https://github.com/FreeTubeApp/FreeTube/releases/download/v0.1.3-beta/FreeTube-linux-x64.tar.xz
|
||||
|
||||
```
|
||||
|
||||
Extract the downloaded archive:
|
||||
```
|
||||
$ tar xf FreeTube-linux-x64.tar.xz
|
||||
|
||||
```
|
||||
|
||||
Go to the Freetube folder:
|
||||
```
|
||||
$ cd FreeTube-linux-x64/
|
||||
|
||||
```
|
||||
|
||||
Launch Freeube using command:
|
||||
```
|
||||
$ ./FreeTub
|
||||
|
||||
```
|
||||
|
||||
This is how FreeTube default interface looks like.
|
||||
|
||||
![][3]
|
||||
|
||||
### Usage
|
||||
|
||||
FreeTube currently uses **YouTube API** to search for videos. And then, It uses **Youtube-dl HTTP API** to grab the raw video files and play them in a basic HTML5 video player. Since subscriptions, history, and saved videos are stored locally on your system, your details will not be sent to Google or anyone else.
|
||||
|
||||
Enter the video name in the search box and hit ENTER key. FreeTube will list out the results based on your search query.
|
||||
|
||||
![][4]
|
||||
|
||||
You can click on any video to play it.
|
||||
|
||||
![][5]
|
||||
|
||||
If you want to change the theme or default API, import/export subscriptions, go to the **Settings** section.
|
||||
|
||||
![][6]
|
||||
|
||||
Please note that FreeTube is still in **beta** stage, so there will be bugs. If there are any bugs, please report them in the GitHub page given at the end of this guide.
|
||||
|
||||
Cheers!
|
||||
|
||||
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.ostechnix.com/freetube-an-open-source-desktop-youtube-player-for-privacy-minded-people/
|
||||
|
||||
作者：[SK][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.ostechnix.com/author/sk/
|
||||
[1]:https://github.com/FreeTubeApp/FreeTube/releases
|
||||
[2]:data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7
|
||||
[3]:http://www.ostechnix.com/wp-content/uploads/2018/03/FreeTube-1.png
|
||||
[4]:http://www.ostechnix.com/wp-content/uploads/2018/03/FreeTube-3.png
|
||||
[5]:http://www.ostechnix.com/wp-content/uploads/2018/03/FreeTube-5-1.png
|
||||
[6]:http://www.ostechnix.com/wp-content/uploads/2018/03/FreeTube-2.png
|
||||
@ -0,0 +1,87 @@
|
||||
How to set up a print server on a Raspberry Pi
|
||||
======
|
||||
|
||||
|
||||
|
||||
I like to work on small projects at home, so this year I picked up a [Raspberry Pi 3 Model B][1], a great model for home hobbyists like me. With built-in wireless on the Raspberry Pi 3 Model B, I can connect the Pi to my home network without a cable. This makes it really easy to put the Raspberry Pi to use right where it is needed.
|
||||
|
||||
At our house, my wife and I both have laptops, but we have just one printer: a slightly used HP Color LaserJet. Because our printer doesn't have a wireless card and can't connect to wireless networks, we usually leave the LaserJet connected to my laptop, since I do most of the printing. While that arrangement works most of the time, sometimes my wife would like to print something without having to go through me.
|
||||
|
||||
### Basic setup
|
||||
|
||||
I realized we really needed a solution to connect the printer to the wireless network so both of us could print to it whenever we wanted. I could buy a wireless print server to connect the USB printer to the wireless network, but I decided instead to use my Raspberry Pi to build a print server to make the LaserJet available to anyone in our house.
|
||||
|
||||
Setting up the Raspberry Pi is fairly straightforward. I downloaded the [Raspbian][2] image and wrote that to my microSD card. Then, I booted the Raspberry Pi with an HDMI display, a USB keyboard, and a USB mouse. With that, I was ready to go!
|
||||
|
||||
The Raspbian system automatically boots into a graphical desktop environment where I performed most of the basic setup: setting the keyboard language, connecting to my wireless network, setting the password for the regular user account (`pi`), and setting the password for the system administrator account (`root`).
|
||||
|
||||
I don't plan to use the Raspberry Pi as a desktop system. I only want to use it remotely from my regular Linux computer. So, I also used Raspbian's graphical administration tool to set the Raspberry Pi to boot into console mode, but not to automatically login as the `pi` user.
|
||||
|
||||
Once I rebooted the Raspberry Pi, I needed to make a few other system tweaks so I could use the Pi as a "server" on my network. I set the Dynamic Host Configuration Protocol (DHCP) client to use a static IP address; by default, the DHCP client might pick any available network address, which would make it tricky to know how to connect to the Raspberry Pi over the network. My home network uses a private class A network, so my router's IP address is `10.0.0.1` and all my IP addresses are `10.0.0.x`. In my case, IP addresses in the lower range are safe, so I set up a static IP address on the wireless network at `10.0.0.11` by adding these lines to the `/etc/dhcpcd.conf` file:
|
||||
```
|
||||
interface wlan0
|
||||
|
||||
static ip_address=10.0.0.11/24
|
||||
|
||||
static routers=10.0.0.1
|
||||
|
||||
static domain_name_servers=8.8.8.8 8.8.4.4
|
||||
|
||||
```
|
||||
|
||||
Before I rebooted again, I made sure that the secure shell daemon (SSHD) was running (you can set what services start at boot-up in Preferences). This allowed me to use a secure shell (SSH) client from my regular Linux system to connect to the Raspberry Pi over the network.
|
||||
|
||||
### Print setup
|
||||
|
||||
Now that my Raspberry Pi was on the network, I did the rest of the setup remotely, using SSH, from my regular Linux desktop machine. Make sure your printer is connected to the Raspberry Pi before taking the following steps.
|
||||
|
||||
Setting up printing is fairly easy. The modern print server is called CUPS, which stands for the Common Unix Printing System. Any recent Unix system should be able to print through a CUPS print server. To set up CUPS on Raspberry Pi, you just need to enter a few commands to install the CUPS software, allow printing by other systems, and restart the print server with the new configuration:
|
||||
```
|
||||
$ sudo apt-get install cups
|
||||
|
||||
$ sudo cupsctl --remote-any
|
||||
|
||||
$ sudo /etc/init.d/cups restart
|
||||
|
||||
```
|
||||
|
||||
Setting up a printer in CUPS is also straightforward and can be done through a web browser. CUPS listens on port 631, so just use your favorite web browser and surf to:
|
||||
```
|
||||
https://10.0.0.11:631/
|
||||
|
||||
```
|
||||
|
||||
Your web browser may complain that it doesn't recognize the web server's https certificate; just accept it, and login as the system administrator. You should see the standard CUPS panel:
|
||||
|
||||
|
||||
From there, navigate to the Administration tab, and select Add Printer.
|
||||
|
||||
|
||||
|
||||
If your printer is already connected via USB, you should be able to easily select the printer's make and model. Don't forget to tick the Share This Printer box so others can use it, too. And now your printer should be set up in CUPS:
|
||||
|
||||
|
||||
|
||||
### Client setup
|
||||
|
||||
Setting up a network printer from the Linux desktop should be quite simple. My desktop is GNOME, and you can add the network printer right from the GNOME Settings application. Just navigate to Devices and Printers and unlock the panel. Click on the Add button to add the printer.
|
||||
|
||||
On my system, GNOME Settings automatically discovered the network printer and added it. If that doesn't happen for you, you may need to add the IP address for your Raspberry Pi to manually add the printer.
|
||||
|
||||
|
||||
|
||||
And that's it! We are now able to print to the Color LaserJet over the wireless network from wherever we are in the house. I no longer need to be physically connected to the printer, and everyone in the family can print on their own.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/3/print-server-raspberry-pi
|
||||
|
||||
作者：[Jim Hall][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/jim-hall
|
||||
[1]:https://www.raspberrypi.org/products/raspberry-pi-3-model-b/
|
||||
[2]:https://www.raspberrypi.org/downloads/
|
||||
128
sources/tech/20180309 A Comparison of Three Linux -App Stores.md
Normal file
128
sources/tech/20180309 A Comparison of Three Linux -App Stores.md
Normal file
@ -0,0 +1,128 @@
|
||||
A Comparison of Three Linux 'App Stores'
|
||||
======
|
||||
|
||||
|
||||
I remember, long, long ago, when installing apps in Linux required downloading and compiling source packages. If you were really lucky, some developer might have packaged the source code into a form that was more easily installable. Without those developers, installing packages could become a dependency nightmare.
|
||||
|
||||
But then, package managers like rpm and dpkg began to rise in popularity, followed quickly by the likes of yum and apt. This was an absolute boon to anyone looking to make Linux their operating system of choice. Although dependencies could still be an issue, they weren’t nearly as bad as they once were. In fact, many of these package managers made short shrift of picking up all the dependencies required for installation.
|
||||
|
||||
And the Linux world rejoiced! Hooray!
|
||||
|
||||
But, with those package managers came a continued requirement of the command line. That, of course, is all fine and good for old hat Linux users. However, there’s a new breed of Linux users who don’t necessarily want to work with the command line. For that user-base, the Linux “app store” was created.
|
||||
|
||||
This all started with the [Synaptic Package Manager][1]. This graphical front end for apt was first released in 2001 and was a breath of fresh air. Synaptic enabled user to easily search for a piece of software and install it with a few quick clicks. Dependencies would be picked up and everything worked. Even when something didn’t work, Synaptic included the means to fix broken packages—all from a drop-down menu.
|
||||
|
||||
Since then, a number of similar tools have arrived on the market, all of which improve on the usability of Synaptic. Although Synaptic is still around (and works quite well), new users demand more modern tools that are even easier to use. And Linux delivered.
|
||||
|
||||
I want to highlight three of the more popular “app stores” to be found on various Linux distributions. In the end, you’ll see that installing applications on Linux, regardless of your distribution, doesn’t have to be a nightmare.
|
||||
|
||||
### GNOME Software
|
||||
|
||||
GNOME’s take on the graphical package manager, [Software][2], hit the scene just in time for the Ubuntu Software Center to finally fade into the sunset (which was fortuitous, considering Canonical’s shift from Unity to GNOME). Any distribution that uses GNOME will include GNOME Software. Unlike the now-defunct Ubuntu Software Center, GNOME Software allows users to both install and update apps from within the same interface (Figure 1).
|
||||
|
||||
![GNOME Software][4]
|
||||
|
||||
Figure 1: The GNOME Software main window.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
To find a piece of software to install, click the Search button (top left, looking glass icon), type the name of the software you want to install, and wait for the results. When you find a title you want to install, click the Install button (Figure 2) and, when prompted, type your user (sudo) password.
|
||||
|
||||
![GNOME Software][7]
|
||||
|
||||
Figure 2: Installing Slack from GNOME Software.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
GNOME Software also includes easy to navigate categories, Editor’s Picks, and GNOME add-ons. As a bonus feature, GNOME Software also supports both snaps and flatpak software. Out of the box, GNOME Software on Ubuntu (and derivatives) support snaps. If you’re adventurous, you can add support for flatpak by opening a terminal window and issuing the command sudo apt install gnome-software-plugin-flatpak.
|
||||
|
||||
GNOME Software makes it so easy to install software on Linux, any user (regardless of experience level) can install and update apps with zero learning curve.
|
||||
|
||||
### KDE Discover
|
||||
|
||||
[Discover][8] is KDE’s answer to GNOME Software. Although the layout (Figure 3) is slightly different, Discover should feel immediately familiar.
|
||||
|
||||
![KDE Discover][10]
|
||||
|
||||
Figure 3: The KDE Discover main window is equally user friendly.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
One of the primary differences between Discover and Software is that Discover differentiates between Plasma (the KDE desktop) and application add-ons. Say, for example, you want to find an “extension” for the Kate text editor; click on Application Addons and search “kate” to see all available addons for the application.
|
||||
|
||||
The Plasma Addons feature makes it easy for users to search through the available desktop widgets and easily install them.
|
||||
|
||||
The one downfall of KDE Discover is that applications are listed in a reverse alphabetical order. Click on one of the given categories, from the main page, and you’ll be given a listing of available apps to scroll through, from Z to A (Figure 4).
|
||||
|
||||
![KDE Discover][12]
|
||||
|
||||
Figure 4: The KDE Discover app listing.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
You will also notice no apparent app rating system. With GNOME Software, it’s not only easy to rate a software title, it’s easy to decide if you want to pass on an app or not (based on a given rating). With KDE Discover, there is no rating system to be found.
|
||||
|
||||
One bonus that Discover adds, is the ability to quickly configure repositories. From the main window, click on Settings, and you can enable/disable any of the included sources (Figure 5). Click the drop-down in the upper right corner, and you can even add new sources.
|
||||
|
||||
![KDE Discover][14]
|
||||
|
||||
Figure 5: Enabling, disable, and add sources, all from within Discover.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
### Pamac
|
||||
|
||||
If you’re hoping to soon count yourself among the growing list of Arch Linux users, you’ll be glad to know that the Linux distribution often considered for the more “elite”, also includes a graphical package manager. [Pamac][15] does an outstanding job of making installing applications on Arch easy. Although Pamac isn’t quite on the design level of either GNOME Software or KDE Discover, it still does a great job of simplifying the installing and updating of applications. From the Pamac main window (Figure 6), you can either click on the search button, or click a Category or Group to find the software you’re looking to install.
|
||||
|
||||
![Pamac][17]
|
||||
|
||||
Figure 6: The Pamac main window.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
If you can’t find the software you’re looking for, you might need to enable one of the many repositories. Click on the Repository button and then search through the categories (Figure 7) to locate the repository to be added.
|
||||
|
||||
![Pamac][19]
|
||||
|
||||
Figure 7: Adding new repositories in Pamac.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
Updates are smoothly handled with Pamac. Click on the Updates button (in the left navigation) and then, in the resulting window (Figure 8), click Apply. All of your Arch updates will be installed.
|
||||
|
||||
![Pamac][21]
|
||||
|
||||
Figure 8: Updating Arch via Pamac.
|
||||
|
||||
[Used with permission][5]
|
||||
|
||||
### More where that came from
|
||||
|
||||
I’ve only listed three graphical package managers. That is not to say these three are the only options to be found. Other distributions have their own takes on the package manager GUI. However, these three do an outstanding job of representing just how far installing software on Linux has come, since those early days of only being able to install via source.
|
||||
|
||||
Learn more about Linux through the free ["Introduction to Linux" ][22]course from The Linux Foundation and edX.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/learn/intro-to-linux/2018/3/comparison-three-linux-app-stores
|
||||
|
||||
作者：[JACK WALLEN][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/jlwallen
|
||||
[1]:https://code.launchpad.net/synaptic
|
||||
[2]:https://wiki.gnome.org/Apps/Software
|
||||
[4]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/gnome_software.jpg?itok=MvRQRX3- (GNOME Software)
|
||||
[7]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/gnome_software_2.jpg?itok=5nzpUQa7 (GNOME Software)
|
||||
[8]:https://userbase.kde.org/Discover
|
||||
[10]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/kde_discover.jpg?itok=LDTmkkMV (KDE Discover)
|
||||
[12]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/kde_discover_2.jpg?itok=f5P7elG_ (KDE Discover)
|
||||
[14]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/kde_discover_3.jpg?itok=JvS3s6FB (KDE Discover)
|
||||
[15]:https://github.com/manjaro/pamac
|
||||
[17]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/pamac.jpg?itok=gZ9X-Z05 (Pamac)
|
||||
[19]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/pamac_1.jpg?itok=Ygt5_U8A (Pamac)
|
||||
[21]:https://www.linux.com/sites/lcom/files/styles/rendered_file/public/pamac_2.jpg?itok=cIjKM51m (Pamac)
|
||||
[22]:https://training.linuxfoundation.org/linux-courses/system-administration-training/introduction-to-linux
|
||||
@ -0,0 +1,100 @@
|
||||
How to check your network connections on Linux
|
||||
======
|
||||
|
||||
|
||||
The **ip** command has a lot to tell you about the configuration and state of your network connections, but what do all those words and numbers mean? Let’s take a deep dive in and see what all the displayed values are trying to tell you.
|
||||
|
||||
When you use the **ip a** (or **ip addr** ) command to get information on all the network interfaces on your system, you're going to see something like this:
|
||||
```
|
||||
$ ip a
|
||||
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
|
||||
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
|
||||
inet 127.0.0.1/8 scope host lo
|
||||
valid_lft forever preferred_lft forever
|
||||
inet6 ::1/128 scope host
|
||||
valid_lft forever preferred_lft forever
|
||||
2: enp0s25: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
|
||||
link/ether 00:1e:4f:c8:43:fc brd ff:ff:ff:ff:ff:ff
|
||||
inet 192.168.0.24/24 brd 192.168.0.255 scope global dynamic enp0s25
|
||||
valid_lft 57295sec preferred_lft 57295sec
|
||||
inet6 fe80::2c8e:1de0:a862:14fd/64 scope link
|
||||
valid_lft forever preferred_lft forever
|
||||
|
||||
```
|
||||
|
||||
The two interfaces on this system — the loopback (lo) and network (enp0s25) — are displayed along with a lot of stats. The "lo" interface is clearly the loopback. We can see the loopback IPv4 address (127.0.0.1) and the loopback IPv6 ( **::1** ) in the listing. The normal network interface is more interesting.
|
||||
|
||||
### Why enp0s25 and not eth0
|
||||
|
||||
If you're wondering why it's called **enp0s25** on this system instead of the likely more familiar **eth0** , a little explanation is in order.
|
||||
|
||||
The new naming scheme is referred to as the “Predictable Network Interface” naming. It’s been used on systemd-based Linux systems for some time. The interface name depends on the physical location of the hardware. The " **en** " simply means "ethernet" just like "eth" does for eth0. The " **p** " is the bus number of the ethernet card and the " **s** " is the slot number. So "enp0s25" tells us a lot about the hardware we're working with.
|
||||
|
||||
The <BROADCAST,MULTICAST,UP,LOWER_UP> string of settings tell us that ...
|
||||
```
|
||||
BROADCAST the interface supports broadcasting
|
||||
MULTICAST the interface supports multicasting
|
||||
UP the network interface is enabled
|
||||
LOWER_UP the network cable is plugged in and device connected to network
|
||||
mtu 1500 the maximum transfer unit (packet size) is 1,500 bytes
|
||||
|
||||
```
|
||||
|
||||
The other values listed also tell us a lot about the interface, but we need to know what words like "brd" and "qlen" represent. So, here's a translation of the rest of the **ip a** shown above.
|
||||
```
|
||||
mtu 1500 maximum transfer unit (packet size)
|
||||
qdisc pfifo_fast used for packet queueing
|
||||
state UP network interface is up
|
||||
group default interface group
|
||||
qlen 1000 transmission queue length
|
||||
link/ether 00:1e:4f:c8:43:fc MAC(hardware) address of the interface
|
||||
brd ff:ff:ff:ff:ff:ff broadcast address
|
||||
inet 192.168.0.24/24 IPv4 address
|
||||
brd 192.168.0.255 broadcast address
|
||||
scope global valid everywhere
|
||||
dynamic enp0s25 address is dynamically assigned
|
||||
valid_lft 80866sec valid lifetime for IPv4 address
|
||||
preferred_lft 80866sec preferred lifetime for IPv4 address
|
||||
inet6 fe80::2c8e:1de0:a862:14fd/64 IPv6 address
|
||||
scope link valid only on this device
|
||||
valid_lft forever valid lifetime for IPv6 address
|
||||
preferred_lft forever preferred lifetime for IPv6 address
|
||||
|
||||
```
|
||||
|
||||
You might have noticed that some of the information that the ifconfig command provides is not included in the **ip a** output — such as the stats on transmitted packets. If you want to see a list of the number of packets transmitted and received along with collisions, you can use this ip command:
|
||||
```
|
||||
$ ip -s link show enp0s25
|
||||
2: enp0s25: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP mode DEFAULT group default qlen 1000
|
||||
link/ether 00:1e:4f:c8:43:fc brd ff:ff:ff:ff:ff:ff
|
||||
RX: bytes packets errors dropped overrun mcast
|
||||
224258568 418718 0 0 0 84376
|
||||
TX: bytes packets errors dropped carrier collsns
|
||||
6131373 78152 0 0 0 0
|
||||
|
||||
```
|
||||
|
||||
Another **ip** command provides information on a system's routing table.
|
||||
```
|
||||
$ ip route show
|
||||
default via 192.168.0.1 dev enp0s25 proto static metric 100
|
||||
169.254.0.0/16 dev enp0s25 scope link metric 1000
|
||||
192.168.0.0/24 dev enp0s25 proto kernel scope link src 192.168.0.24 metric 100
|
||||
|
||||
```
|
||||
|
||||
The **ip** command is extremely versatile. You can get a helpful cheat sheet on the **ip** command and its options from [Red Hat][1].
|
||||
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.networkworld.com/article/3262045/linux/checking-your-network-connections-on-linux.html
|
||||
|
||||
作者：[Sandra Henry-Stocker][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.networkworld.com/author/Sandra-Henry_Stocker/
|
||||
[1]:https://access.redhat.com/sites/default/files/attachments/rh_ip_command_cheatsheet_1214_jcs_print.pdf
|
||||
@ -0,0 +1,44 @@
|
||||
如何使用看板创建更好的文档
|
||||
======
|
||||
|
||||
如果你正在处理文档、网站或其他面向用户的内容，那么了解用户希望找到的内容（包括他们想要的信息以及信息的组织和结构）很有帮助。毕竟，如果人们无法找到他们想要的东西，那么出色的内容就没有用。
|
||||
|
||||
|
||||
卡片分类是一种简单而有效的方式，可以从用户那里收集有关菜单界面和页面的内容。最简单的实现方式是在你计划在网站或文章分类标注一些索引卡，并要求用户按照查找信息的方式对卡进行分类。变体包括让人们编写自己的菜单标题或内容元素。
|
||||
|
||||
我们的目标是了解用户的期望以及他们希望在哪里找到它，而不是自己弄清楚菜单和布局。当与用户处于相同的物理位置时，这是相对简单的，但当莫尝试从多个位置的人员获得反馈时，这会更具挑战性。
|
||||
|
||||
我发现[看板][1]对于这些情况是一个很好的工具。它允许人们轻松拖动虚拟卡进行分类和排名，而且与专门卡片分类软件不同，它们是多用途的。
|
||||
|
||||
我经常使用 Trello 进行卡片分类，但有几种你可能想尝试的[开源替代品][2]。
|
||||
|
||||
### 怎么运行的
|
||||
|
||||
我最成功的 kanban 实验是在写 [Gluster][3] 文档的时候- 一个免费开源的可扩展网络存储文件系统。我需要携带大量随时间增长的文档，并将其分成若干类别以创建引导系统。由于我没有必要的技术知识来分类，我向 Gluster 团队和开发人员社区寻求指导。
|
||||
|
||||
首先，我创建了一个共享看板。我列出了一些通用名称，这些名称可以为我计划在文档中涵盖的所有主题排序和创建卡片。我标记了一些不同颜色的卡片，以表明某个主题缺失并需要创建，或者它存在并需要删除。然后，我把所有卡片放入“未排序”一列，并要求人们将它们拖到他们认为应该组织卡片的地方，然后给我一个他们认为是理想状态的截图。
|
||||
|
||||
处理所有截图是最棘手的部分。我希望有一个合并或共识功能可以帮助我汇总每个人的数据，而不必检查一堆截图。幸运的是，在第一个人对卡片进行分类之后，人们或多或少地对该结构达成一致，而只做了很小的修改。当对某个主题的位置有不同意见时，我发起一个快速会议，让人们可以解释他们的想法，并且可以排除分歧。
|
||||
|
||||
### 使用数据
|
||||
|
||||
在这里，很容易将捕捉到的信息转换为菜单并对其进行优化。如果用户认为项目应该成为子菜单，他们通常会在评论中或在电话聊天时告诉我。对菜单组织的看法因人们的工作任务而异，所以从来没有完全达成一致意见，但用户进行测试意味着你不会对人们使用什么以及在哪里查找有很多盲点。
|
||||
|
||||
将卡片分类与分析功能配对，可以让你更深入地了解人们在寻找什么。有一次，当我对一些我正在写培训文档进行分析时，我惊讶地发现搜索量最大的页面是关于资本的。所以我在顶层菜单层面上显示了该页面，即使我的“逻辑”设置将它放在了子菜单中。
|
||||
|
||||
我发现看板卡片分类是一种很好的方式，可以帮助我创建用户想要查看的内容，并将其放在希望被找到的位置。你是否发现了另一种对用户友好的组织内容的方法？或者看板的另一种有趣用途是什么？如果有的话，请在评论中分享你的想法。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/17/11/kanban-boards-card-sorting
|
||||
|
||||
作者：[Heidi Waterhouse][a]
|
||||
译者：[geekpi](https://github.com/geekpi)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/hwaterhouse
|
||||
[1]:https://en.wikipedia.org/wiki/Kanban
|
||||
[2]:https://opensource.com/alternatives/trello
|
||||
[3]:https://www.gluster.org/
|
||||
@ -1,109 +0,0 @@
|
||||
在 Linux 中自动配置 IPv6 地址
|
||||
======
|
||||
|
||||
|
||||
|
||||
在 [ KVM 中测试 IPv6 网络：第 1 部分][1] 一文中，我们学习了关于唯一本地地址（ULAs）的相关内容。在本文中，我们将学习如何为 ULAs 自动配置 IP 地址。
|
||||
|
||||
### 何时使用唯一本地地址
|
||||
|
||||
唯一本地地址使用 fd00::/8 地址块，它类似于我们常用的 IPv4 的私有地址：10.0.0.0/8、172.16.0.0/12、以及 192.168.0.0/16。但它们并不能直接替换。IPv4 的私有地址分类和网络地址转换（NAT）功能是为了缓解 IPv4 地址短缺的问题，这是个明智的解决方案，它延缓了本该被替换的 IPv4 的生命周期。IPv6 也支持 NAT，但是我想不出使用它的理由。IPv6 的地址数量远远大于 IPv4；它是不一样的，因此需要做不一样的事情。
|
||||
|
||||
那么，ULAs 存在的意义是什么呢？尤其是在我们已经有了本地链路地址（fe80::/10）时，到底需不需要我们去配置它们呢？它们之间（译者注：指的是唯一本地地址和本地链路地址）有两个重要的区别。一是，本地链路地址是不可路由的，因此，你不能跨子网使用它。二是，ULAs 是你自己管理的；你可以自己选择它用于子网的地址范围，并且它们是可路由的。
|
||||
|
||||
使用 ULAs 的另一个好处是，如果你只是在局域网中“混日子”的话，你不需要为它们分配全局单播 IPv6 地址。当然了，如果你的 ISP 已经为你分配了 IPv6 的全局单播地址，就不需要使用 ULAs 了。你也可以在同一个网络中混合使用全局单播地址和 ULAs，但是，我想不出这样使用的一个好理由，并且要一定确保你不使用网络地址转换以使 ULAs 可公共访问。在我看来，这是很愚蠢的行为。
|
||||
|
||||
ULAs 是仅为私有网络使用的，并且它会阻塞所有流出你的网络的数据包，不允许进入因特网。这很简单，在你的边界设备上只要阻止整个 fd00::/8 范围的 IPv6 地址即可实现。
|
||||
|
||||
### 地址自动配置
|
||||
|
||||
ULAs 不像本地链路地址那样自动配置的，但是使用 radvd 设置自动配置是非常容易的，radva 是路由器公告守护程序。在你开始之前，运行 `ifconfig` 或者 `ip addr show` 去查看你现有的 IP 地址。
|
||||
|
||||
在生产系统上使用时，你应该将 radvd 安装在一台单独的路由器上，如果只是测试使用，你可以将它安装在你的网络中的任意 Linux PC 上。在我的小型 KVM 测试实验室中，我使用 `apt-get install radvd` 命令把它安装在 Ubuntu 上。安装完成之后，我先不启动它，因为它还没有配置文件：
|
||||
```
|
||||
$ sudo systemctl status radvd
|
||||
● radvd.service - LSB: Router Advertising Daemon
|
||||
Loaded: loaded (/etc/init.d/radvd; bad; vendor preset: enabled)
|
||||
Active: active (exited) since Mon 2017-12-11 20:08:25 PST; 4min 59s ago
|
||||
Docs: man:systemd-sysv-generator(8)
|
||||
|
||||
Dec 11 20:08:25 ubunut1 systemd[1]: Starting LSB: Router Advertising Daemon...
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: Starting radvd:
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: * /etc/radvd.conf does not exist or is empty.
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: * See /usr/share/doc/radvd/README.Debian
|
||||
Dec 11 20:08:25 ubunut1 radvd[3541]: * radvd will *not* be started.
|
||||
Dec 11 20:08:25 ubunut1 systemd[1]: Started LSB: Router Advertising Daemon.
|
||||
|
||||
```
|
||||
|
||||
这些所有的消息有点让人困惑，实际上 radvd 并没有运行，你可以使用经典命令 `ps|grep radvd` 来验证这一点。因此，我们现在需要去创建 `/etc/radvd.conf` 文件。拷贝这个示例，将第一行的网络接口名替换成你自己的接口名字：
|
||||
```
|
||||
interface ens7 {
|
||||
AdvSendAdvert on;
|
||||
MinRtrAdvInterval 3;
|
||||
MaxRtrAdvInterval 10;
|
||||
prefix fd7d:844d:3e17:f3ae::/64
|
||||
{
|
||||
AdvOnLink on;
|
||||
AdvAutonomous on;
|
||||
};
|
||||
|
||||
};
|
||||
|
||||
```
|
||||
|
||||
前缀定义了你的网络地址，它是地址的前 64 位。前两个字符必须是 `fd`，前缀接下来的剩余部分你自己定义它，最后的 64 位留空，因为 radvd 将去分配最后的 64 位。前缀后面的 16 位用来定义子网，剩余的地址定义为主机地址。你的子网必须总是 /64。RFC 4193 要求地址必须随机生成；查看 [在 KVM 中测试 IPv6 Networking：第 1 部分][1] 学习创建和管理 ULAs 的更多知识。
|
||||
|
||||
### IPv6 转发
|
||||
|
||||
IPv6 转发必须要启用。下面的命令去启用它，重启后生效：
|
||||
```
|
||||
$ sudo sysctl -w net.ipv6.conf.all.forwarding=1
|
||||
|
||||
```
|
||||
|
||||
取消注释或者添加如下的行到 `/etc/sysctl.conf` 文件中，以使它永久生效：
|
||||
```
|
||||
net.ipv6.conf.all.forwarding = 1
|
||||
```
|
||||
|
||||
启动 radvd 守护程序：
|
||||
```
|
||||
$ sudo systemctl stop radvd
|
||||
$ sudo systemctl start radvd
|
||||
|
||||
```
|
||||
|
||||
这个示例在我的 Ubuntu 测试系统中遇到了一个怪事；radvd 总是停止，我查看它的状态却没有任何问题，做任何改变之后都需要重新启动 radvd。
|
||||
|
||||
启动成功后没有任何输出，并且失败也是如此，因此，需要运行 `sudo systemctl radvd status` 去查看它的运行状态。如果有错误，systemctl 会告诉你。一般常见的错误都是 `/etc/radvd.conf` 中的语法错误。
|
||||
|
||||
在 Twitter 上抱怨了上述问题之后，我学到了一件很酷的技巧：当你运行 ` journalctl -xe --no-pager` 去调试 systemctl 错误时，你的输出将被封装打包，然后，你就可以看到错误信息。
|
||||
|
||||
现在检查你的主机，查看它们自动分配的新地址：
|
||||
```
|
||||
$ ifconfig
|
||||
ens7 Link encap:Ethernet HWaddr 52:54:00:57:71:50
|
||||
[...]
|
||||
inet6 addr: fd7d:844d:3e17:f3ae:9808:98d5:bea9:14d9/64 Scope:Global
|
||||
[...]
|
||||
|
||||
```
|
||||
|
||||
本文到此为止，下周继续学习如何为 ULAs 管理 DNS，这样你就可以使用一个合适的主机名来代替这些长长的 IPv6 地址。
|
||||
|
||||
通过来自 Linux 基金会和 edX 的 ["Linux 入门" ][2] 免费课程学习更多 Linux 的知识。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/learn/intro-to-linux/2017/12/ipv6-auto-configuration-linux
|
||||
|
||||
作者：[Carla Schroder][a]
|
||||
译者：[qhwdw](https://github.com/qhwdw)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/cschroder
|
||||
[1]:https://www.linux.com/learn/intro-to-linux/2017/11/testing-ipv6-networking-kvm-part-1
|
||||
[2]:https://training.linuxfoundation.org/linux-courses/system-administration-training/introduction-to-linux
|
||||
@ -1,260 +0,0 @@
|
||||
Linux 启动过程分析
|
||||
======
|
||||
|
||||
|
||||
|
||||
图片由企鹅和靴子“赞助”，由 Opensource.com 修改。CC BY-SA 4.0。
|
||||
|
||||
关于开源软件最古老的笑话是：“代码是自文档化的（self-documenting）”。经验表明，阅读源代码就像听天气预报一样：明智的人依然出门会看看室外的天气。本文讲述了如何运用调试工具来观察和分析 Linux 系统的启动。分析一个正常的系统启动过程，有助于用户和开发人员应对不可避免的故障。
|
||||
|
||||
从某些方面看，启动过程非常简单。内核在单核上启动单线程和同步，似乎可以理解。但内核本身是如何启动的呢？[initrd（initial ramdisk）][1]和引导程序（bootloaders）具有哪些功能？还有，为什么以太网端口上的 LED 灯是常亮的呢？
|
||||
|
||||
请继续阅读寻找答案。GitHub 也提供了 [介绍演示和练习的代码][2]。
|
||||
|
||||
### 启动的开始：OFF 状态
|
||||
|
||||
#### 局域网唤醒（Wake-on-LAN）
|
||||
|
||||
OFF 状态表示系统没有上电，没错吧？表面简单，其实不然。例如，如果系统启用连局域网唤醒机制（WOL），以太网指示灯将亮起。通过以下命令来检查是否是这种情况：
|
||||
|
||||
```
|
||||
$# sudo ethtool <interface name>
|
||||
```
|
||||
|
||||
其中 `<interface name>` 是网络接口的名字，比如 `eth0`。（`ethtool` 可以在同名的 Linux 软件包中找到。）如果输出中的 “Wake-on” 显示 “g”，则远程主机可以通过发送 [魔法数据包（MagicPacket）][3] 来启动系统。如果您无意远程唤醒系统，也不希望其他人这样做，请在系统 BIOS 菜单中将 WOL 关闭，或者用以下方式：
|
||||
|
||||
```
|
||||
$# sudo ethtool -s <interface name> wol d
|
||||
```
|
||||
|
||||
响应魔法数据包的处理器可能是网络接口的一部分，也可能是 [底板管理控制器（Baseboard Management Controller，BMC）][4]。
|
||||
|
||||
#### 英特尔管理引擎、平台路径控制器和 Minix
|
||||
|
||||
BMC 不是唯一的在系统关闭时仍在监听的微控制器（MCU）。x86_64 系统还包含了用于远程管理系统的英特尔管理引擎（IME）软件套件。从服务器到笔记本电脑，各种各样的设备都包含了这项技术，开启了如 KVM 远程控制和英特尔功能许可服务等 [功能][5]。根据 [Intel 自己的检测工具][7]，[IME 存在尚未修补的漏洞][6]。坏消息是，要禁用 IME 很难。Trammell Hudson 发起了一个 [me_cleaner 项目][8]，它可以清除一些相对恶劣的 IME 组件，比如嵌入式 Web 服务器，但也可能会影响运行它的系统。
|
||||
|
||||
IME 固件和系统管理模式（SMM）软件是 [基于 Minix 操作系统][9] 的，并运行在单独的平台路径控制器上，而不是主 CPU 上。然后，SMM 启动位于主处理器上的通用可扩展固件接口（UEFI）软件，相关内容 [已被提及很多][10]。Google 的 Coreboot 小组已经启动了一个雄心勃勃的 [非扩展性缩减版固件][11]（NERF）项目，其目的不仅是要取代 UEFI，还要取代早期的 Linux 用户空间组件，如 systemd。在我们等待这些新成果的同时，Linux 用户现在就可以从 Purism、System76 或 Dell 等处购买 [禁用了 IME][12] 的笔记本电脑，另外 [带有 ARM 64 位处理器笔记本电脑][13] 还是值得期待的。
|
||||
|
||||
####
|
||||
#### 引导程序
|
||||
|
||||
除了启动问题不断的间谍软件外，早期的引导固件还有什么功能呢？引导程序的作用是为新上电的处理器提供运行像 Linux 之类的通用操作系统所需的资源。在开机时，不但没有虚拟内存，在控制器启动之前连 DRAM 也没有。然后，引导程序打开电源，并扫描总线和接口，以定位到内核镜像和根文件系统的位置。U-Boot 和 GRUB 等常见的引导程序支持 USB、PCI 和 NFS 等接口，以及更多的嵌入式专用设备，如 NOR 和 NAND 闪存。引导程序还与 [可信平台模块][14]（TPMs）等硬件安全设备进行交互，在启动最开始建立信任链。
|
||||
|
||||
![Running the U-boot bootloader][16]
|
||||
|
||||
在构建主机上的沙盒中运行 U-boot 引导程序。
|
||||
|
||||
包括树莓派、任天堂设备、汽车板和 Chromebook 在内的系统都支持广泛使用的开源引导程序 [U-Boot][17]。它没有系统日志，当发生问题时，甚至没有任何控制台输出。为了便于调试，U-Boot 团队提供了一个沙盒，可以在构建主机甚至是夜间的持续整合（Continuous Integration）系统上测试补丁程序。如果系统上安装了 Git 和 GNU Compiler Collection（GCC）等通用的开发工具，使用 U-Boot 沙盒会相对简单：
|
||||
|
||||
```
|
||||
|
||||
|
||||
$# git clone git://git.denx.de/u-boot; cd u-boot
|
||||
|
||||
$# make ARCH=sandbox defconfig
|
||||
|
||||
$# make; ./u-boot
|
||||
|
||||
=> printenv
|
||||
|
||||
=> help
|
||||
```
|
||||
|
||||
在 x86_64 上运行 U-Boot，可以测试一些棘手的功能，如 [模拟存储设备][2] 重新分区、基于 TPM 的密钥操作以及 USB 设备热插拔等。U-Boot 沙盒甚至可以在 GDB 调试器下单步执行。使用沙盒进行开发的速度比将引导程序刷新到电路板上的测试快 10 倍，并且可以使用 Ctrl + C 恢复一个“变砖”的沙盒。
|
||||
|
||||
### 启动内核
|
||||
|
||||
#### 配置引导内核
|
||||
|
||||
完成任务后，引导程序将跳转到已加载到主内存中的内核代码，并开始执行，传递用户指定的任何命令行选项。内核是什么样的程序呢？用命令 `file /boot/vmlinuz` 可以看到它是一个“bzImage”，意思是一个大的压缩的镜像。Linux 源代码树包含了一个可以解压缩这个文件的工具—— [extract-vmlinux][18]：
|
||||
|
||||
```
|
||||
|
||||
|
||||
$# scripts/extract-vmlinux /boot/vmlinuz-$(uname -r) > vmlinux
|
||||
|
||||
$# file vmlinux
|
||||
|
||||
vmlinux: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically
|
||||
|
||||
linked, stripped
|
||||
```
|
||||
|
||||
内核是一个 [可执行与可链接格式][19]（ELF）的二进制文件，就像 Linux 的用户空间程序一样。这意味着我们可以使用 `binutils` 包中的命令，如 `readelf` 来检查它。比较一下输出，例如：
|
||||
|
||||
```
|
||||
|
||||
|
||||
$# readelf -S /bin/date
|
||||
|
||||
$# readelf -S vmlinux
|
||||
```
|
||||
|
||||
这两个文件中的段内容大致相同。
|
||||
|
||||
所以内核必须像其他的 Linux ELF 文件一样启动，但用户空间程序是如何启动的呢？在 `main()` 函数中？并不确切。
|
||||
|
||||
在 `main()` 函数运行之前，程序需要一个执行上下文，包括堆栈内存以及 `stdio`、`stdout` 和 `stderr` 的文件描述符。用户空间程序从标准库（多数 Linux 系统在用“glibc”）中获取这些资源。参照以下输出：
|
||||
|
||||
```
|
||||
|
||||
|
||||
$# file /bin/date
|
||||
|
||||
/bin/date: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically
|
||||
|
||||
linked, interpreter /lib64/ld-linux-x86-64.so.2, for GNU/Linux 2.6.32,
|
||||
|
||||
BuildID[sha1]=14e8563676febeb06d701dbee35d225c5a8e565a,
|
||||
|
||||
stripped
|
||||
```
|
||||
|
||||
ELF 二进制文件有一个解释器，就像 Bash 和 Python 脚本一样，但是解释器不需要像脚本那样用 `#!` 指定，因为 ELF 是 Linux 的原生格式。ELF 解释器通过调用 `_start()` 函数来用所需资源 [配置一个二进制文件][20]，这个函数可以从 glibc 源代码包中找到，可以 [用 GDB 查看][21]。内核显然没有解释器，必须自我配置，这是怎么做到的呢？
|
||||
|
||||
用 GDB 检查内核的启动给出了答案。首先安装内核的调试软件包，内核中包含一个未剥离的（unstripped）vmlinux，例如 `apt-get install linux-image-amd64-dbg`，或者从源代码编译和安装你自己的内核，可以参照 [Debian Kernel Handbook][22] 中的指令。`gdb vmlinux` 后加 `info files` 可显示 ELF 段 `init.text`。在 `init.text` 中用 `l *(address)` 列出程序执行的开头，其中 `address` 是 `init.text` 的十六进制开头。用 GDB 可以看到 x86_64 内核从内核文件 [arch/x86/kernel/head_64.S][23] 开始启动，在这个文件中我们找到了汇编函数 `start_cpu0()`，以及一段明确的代码显示在调用 `x86_64 start_kernel()` 函数之前创建了堆栈并解压了 zImage。ARM 32 位内核也有类似的文件 [arch/arm/kernel/head.S][24]。`start_kernel()` 不针对特定的体系结构，所以这个函数驻留在内核的 [init/main.c][25] 中。`start_kernel()` 可以说是 Linux 真正的 `main()` 函数。
|
||||
|
||||
### 从 start_kernel() 到 PID 1
|
||||
|
||||
#### 内核的硬件清单：设备树和 ACPI 表
|
||||
|
||||
在引导时，内核需要硬件信息，不仅仅是已编译过的处理器类型。代码中的指令通过单独存储的配置数据进行扩充。有两种主要的数据存储方法：[设备树][26] 和 [高级配置和电源接口（ACPI）表][27]。内核通过读取这些文件了解每次启动时需要运行的硬件。
|
||||
|
||||
对于嵌入式设备，设备树是已安装硬件的清单。设备树只是一个与内核源代码同时编译的文件，通常与 `vmlinux` 一样位于 `/boot` 目录中。要查看 ARM 设备上的设备树的内容，只需对名称与 `/boot/*.dtb` 匹配的文件执行 `binutils` 包中的 `strings` 命令即可，`dtb` 是指一个设备树二进制文件。显然，只需编辑构成它的类 JSON 文件并重新运行随内核源代码提供的特殊 `dtc` 编译器即可修改设备树。虽然设备树是一个静态文件，其文件路径通常由命令行引导程序传递给内核，但近年来增加了一个 [设备树覆盖][28] 的功能，内核在启动后可以动态加载热插拔的附加设备。
|
||||
|
||||
x86 系列和许多企业级的 ARM64 设备使用 [ACPI][27] 机制。与设备树不同的是，ACPI 信息存储在内核在启动时通过访问板载 ROM 而创建的 `/sys/firmware/acpi/tables` 虚拟文件系统中。读取 ACPI 表的简单方法是使用 `acpica-tools` 包中的 `acpidump` 命令。例如：
|
||||
|
||||
![ACPI tables on Lenovo laptops][30]
|
||||
|
||||
|
||||
联想笔记本电脑的 ACPI 表都是为 Windows 2001 设置的。
|
||||
|
||||
是的，你的 Linux 系统已经准备好用于 Windows 2001 了，你要考虑安装吗？与设备树不同，ACPI 具有方法和数据，而设备树更多地是一种硬件描述语言。ACPI 方法在启动后仍处于活动状态。例如，运行 `acpi_listen` 命令（在 `apcid` 包中），然后打开和关闭笔记本机盖会发现 ACPI 功能一直在运行。暂时地和动态地 [覆盖 ACPI 表][31] 是可能的，而永久地改变它需要在引导时与 BIOS 菜单交互或刷新 ROM。如果你遇到那么多麻烦，也许你应该 [安装 coreboot][32]，这是开源固件的替代品。
|
||||
|
||||
#### 从 start_kernel() 到用户空间
|
||||
|
||||
[init/main.c][25] 中的代码竟然是可读的，而且有趣的是，它仍然在使用 1991 - 1992 年的 Linus Torvalds 的原始版权。在一个刚启动的系统上运行 `dmesg | head`，其输出主要来源于此文件。第一个 CPU 注册到系统中，全局数据结构被初始化，并且调度程序、中断处理程序（IRQ）、定时器和控制台按照严格的顺序逐一启动。在 `timekeeping_init()` 函数运行之前，所有的时间戳都是零。内核初始化的这部分是同步的，也就是说执行只发生在一个线程中，在最后一个完成并返回之前，没有任何函数会被执行。因此，即使在两个系统之间，`dmesg` 的输出也是完全可重复的，只要它们具有相同的设备树或 ACPI 表。Linux 的行为就像在 MCU 上运行的 RTOS（实时操作系统）一样，如 QNX 或 VxWorks。这种情况持续存在于函数 `rest_init()` 中，该函数在终止时由 `start_kernel()` 调用。
|
||||
|
||||
![Summary of early kernel boot process.][34]
|
||||
|
||||
早期的内核启动流程
|
||||
|
||||
函数 `rest_init()` 产生了一个新进程以运行 `kernel_init()`，并调用了 `do_initcalls()`。用户可以通过将 `initcall_debug` 附加到内核命令行来监控 `initcalls`，这样每运行一次 `initcall` 函数就会产生 `dmesg` 条目。`initcalls` 会历经七个连续的级别：early、core、postcore、arch、subsys、fs、device 和 late。`initcalls` 最为用户可见的部分是所有处理器外围设备的探测和设置：总线、网络、存储和显示器等等，同时加载其内核模块。`rest_init()` 也会在引导处理器上产生第二个线程，它首先运行 `cpu_idle()`，然后等待调度器分配工作。
|
||||
|
||||
`kernel_init()` 也可以 [设置对称多处理（SMP）结构][35]。在较新的内核中，如果 `dmesg` 的输出中出现“启动第二个 CPU...”等字样，系统便使用了 SMP。SMP 通过“热插拔”CPU 来进行，这意味着它用状态机来管理其生命周期，这种状态机在概念上类似于热插拔的 U 盘一样。内核的电源管理系统经常会使某个核（core）离线，然后根据需要将其唤醒，以便在不忙的机器上反复调用同一段的 CPU 热插拔代码。观察电源管理系统调用 CPU 热插拔代码的 [BCC 工具][36] 称为 `offcputime.py`。
|
||||
|
||||
请注意，`init/main.c` 中的代码在 `smp_init()` 运行时几乎已执行完毕：引导处理器已经完成了大部分其他核无需重复的一次性初始化操作。尽管如此，跨 CPU 的线程仍然要在每个核上生成，以管理每个核的中断（IRQ）、工作队列、定时器和电源事件。例如，通过 `ps -o psr` 命令可以查看服务 softirqs 和 workqueues 在每个 CPU 上的线程。
|
||||
|
||||
```
|
||||
|
||||
|
||||
$\# ps -o pid,psr,comm $(pgrep ksoftirqd)
|
||||
|
||||
PID PSR COMMAND
|
||||
|
||||
7 0 ksoftirqd/0
|
||||
|
||||
16 1 ksoftirqd/1
|
||||
|
||||
22 2 ksoftirqd/2
|
||||
|
||||
28 3 ksoftirqd/3
|
||||
|
||||
|
||||
|
||||
$\# ps -o pid,psr,comm $(pgrep kworker)
|
||||
|
||||
PID PSR COMMAND
|
||||
|
||||
4 0 kworker/0:0H
|
||||
|
||||
18 1 kworker/1:0H
|
||||
|
||||
24 2 kworker/2:0H
|
||||
|
||||
30 3 kworker/3:0H
|
||||
|
||||
[ . . . ]
|
||||
```
|
||||
|
||||
其中，PSR 字段代表“处理器”。每个核还必须拥有自己的定时器和 `cpuhp` 热插拔处理程序。
|
||||
|
||||
那么用户空间是如何启动的呢？在最后，`kernel_init()` 寻找可以代表它执行 `init` 进程的 `initrd`。如果没有找到，内核直接执行 `init` 本身。那么为什么需要 `initrd` 呢？
|
||||
|
||||
#### 早期的用户空间：谁规定要用 initrd？
|
||||
|
||||
除了设备树之外，在启动时可以提供给内核的另一个文件路径是 `initrd` 的路径。`initrd` 通常位于 `/boot` 目录中，与 x86 系统中的 bzImage 文件 vmlinuz 一样，或是与 ARM 系统中的 uImage 和设备树相同。用 `initramfs-tools-core` 软件包中的 `lsinitramfs` 工具可以列出 `initrd` 的内容。发行版的 `initrd` 方案包含了最小化的 `/bin`、`/sbin` 和 `/etc` 目录以及内核模块，还有 `/scripts` 中的一些文件。所有这些看起来都很熟悉，因为 `initrd` 大致上是一个简单的最小化 Linux 根文件系统。看似相似，其实不然，因为位于虚拟内存盘中的 `/bin` 和 `/sbin` 目录下的所有可执行文件几乎都是指向 [BusyBox binary][38] 的符号链接，由此导致 `/bin` 和 `/sbin` 目录比 glibc 的小 10 倍。
|
||||
|
||||
如果要做的只是加载一些模块，然后在普通的根文件系统上启动 `init`，为什么还要创建一个 `initrd` 呢？想想一个加密的根文件系统，解密可能依赖于加载一个位于根文件系统 `/lib/modules` 的内核模块，当然还有 `initrd` 中的。加密模块可能被静态地编译到内核中，而不是从文件加载，但有多种原因不希望这样做。例如，用模块静态编译内核可能会使其太大而不能适应存储空间，或者静态编译可能会违反软件许可条款。不出所料，存储、网络和人类输入设备（HID）驱动程序也可能存在于 `initrd` 中。`initrd` 基本上包含了任何挂载根文件系统所必需的非内核代码。`initrd` 也是用户存放 [自定义ACPI][38] 表代码的地方。
|
||||
|
||||
![Rescue shell and a custom <code>initrd</code>.][40]
|
||||
|
||||
救援模式的 shell 和自定义的 `initrd` 还是很有意思的。
|
||||
|
||||
`initrd` 对测试文件系统和数据存储设备也很有用。将这些测试工具存放在 `initrd` 中，并从内存中运行测试，而不是从被测对象中运行。
|
||||
|
||||
最后，当 `init` 开始运行时，系统就启动啦！由于辅助处理器正在运行，机器已经成为我们所熟知和喜爱的异步、可抢占、不可预测和高性能的生物。的确，`ps -o pid,psr,comm -p 1` 很容易显示已不在引导处理器上运行的用户空间的 `init` 进程。
|
||||
|
||||
### Summary
|
||||
### 总结
|
||||
|
||||
Linux 引导过程听起来或许令人生畏，即使考虑到简单嵌入式设备上的软件数量。换个角度来看，启动过程相当简单，因为启动中没有抢占、RCU 和竞争条件等扑朔迷离的复杂功能。只关注内核和 PID 1 会忽略了引导程序和辅助处理器为运行内核执行的大量准备工作。虽然内核在 Linux 程序中是独一无二的，但通过一些检查 ELF 文件的工具也可以了解其结构。学习一个正常的启动过程，可以帮助运维人员处理启动的故障。
|
||||
|
||||
要了解更多信息，请参阅 Alison Chaiken 的演讲——[Linux: The first second][41]，将在 1 月 22 日至 26 日在悉尼举行。参见 [linux.conf.au][42]。
|
||||
|
||||
感谢 [Akkana Peck][43] 的提议和指正。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/1/analyzing-linux-boot-process
|
||||
|
||||
作者：[Alison Chaiken][a]
|
||||
译者：[jessie-pang](https://github.com/jessie-pang)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/don-watkins
|
||||
[1]:https://en.wikipedia.org/wiki/Initial_ramdisk
|
||||
[2]:https://github.com/chaiken/LCA2018-Demo-Code
|
||||
[3]:https://en.wikipedia.org/wiki/Wake-on-LAN
|
||||
[4]:https://lwn.net/Articles/630778/
|
||||
[5]:https://www.youtube.com/watch?v=iffTJ1vPCSo&amp;amp;amp;amp;amp;index=65&amp;amp;amp;amp;amp;list=PLbzoR-pLrL6pISWAq-1cXP4_UZAyRtesk
|
||||
[6]:https://security-center.intel.com/advisory.aspx?intelid=INTEL-SA-00086&amp;amp;amp;amp;amp;languageid=en-fr
|
||||
[7]:https://www.intel.com/content/www/us/en/support/articles/000025619/software.html
|
||||
[8]:https://github.com/corna/me_cleaner
|
||||
[9]:https://lwn.net/Articles/738649/
|
||||
[10]:https://lwn.net/Articles/699551/
|
||||
[11]:https://trmm.net/NERF
|
||||
[12]:https://www.extremetech.com/computing/259879-dell-now-shipping-laptops-intels-management-engine-disabled
|
||||
[13]:https://lwn.net/Articles/733837/
|
||||
[14]:https://linuxplumbersconf.org/2017/ocw/events/LPC2017/tracks/639
|
||||
[15]:/file/383501
|
||||
[16]:https://opensource.com/sites/default/files/u128651/linuxboot_1.png "Running the U-boot bootloader"
|
||||
[17]:http://www.denx.de/wiki/DULG/Manual
|
||||
[18]:https://github.com/torvalds/linux/blob/master/scripts/extract-vmlinux
|
||||
[19]:http://man7.org/linux/man-pages/man5/elf.5.html
|
||||
[20]:https://0xax.gitbooks.io/linux-insides/content/Misc/program_startup.html
|
||||
[21]:https://github.com/chaiken/LCA2018-Demo-Code/commit/e543d9812058f2dd65f6aed45b09dda886c5fd4e
|
||||
[22]:http://kernel-handbook.alioth.debian.org/
|
||||
[23]:https://github.com/torvalds/linux/blob/master/arch/x86/boot/compressed/head_64.S
|
||||
[24]:https://github.com/torvalds/linux/blob/master/arch/arm/boot/compressed/head.S
|
||||
[25]:https://github.com/torvalds/linux/blob/master/init/main.c
|
||||
[26]:https://www.youtube.com/watch?v=m_NyYEBxfn8
|
||||
[27]:http://events.linuxfoundation.org/sites/events/files/slides/x86-platform.pdf
|
||||
[28]:http://lwn.net/Articles/616859/
|
||||
[29]:/file/383506
|
||||
[30]:https://opensource.com/sites/default/files/u128651/linuxboot_2.png "ACPI tables on Lenovo laptops"
|
||||
[31]:https://www.mjmwired.net/kernel/Documentation/acpi/method-customizing.txt
|
||||
[32]:https://www.coreboot.org/Supported_Motherboards
|
||||
[33]:/file/383511
|
||||
[34]:https://opensource.com/sites/default/files/u128651/linuxboot_3.png "Summary of early kernel boot process."
|
||||
[35]:http://free-electrons.com/pub/conferences/2014/elc/clement-smp-bring-up-on-arm-soc
|
||||
[36]:http://www.brendangregg.com/ebpf.html
|
||||
[37]:https://www.busybox.net/
|
||||
[38]:https://www.mjmwired.net/kernel/Documentation/acpi/initrd_table_override.txt
|
||||
[39]:/file/383516
|
||||
[40]:https://opensource.com/sites/default/files/u128651/linuxboot_4.png "Rescue shell and a custom <code>initrd</code>."
|
||||
[41]:https://rego.linux.conf.au/schedule/presentation/16/
|
||||
[42]:https://linux.conf.au/index.html
|
||||
[43]:http://shallowsky.com/
|
||||
219
translated/tech/20180129 Parsing HTML with Python.md
Normal file
219
translated/tech/20180129 Parsing HTML with Python.md
Normal file
@ -0,0 +1,219 @@
|
||||
用Python解析HTML
|
||||
======
|
||||
|
||||
|
||||
|
||||
图片由Jason Baker为Opensource.com所作。
|
||||
|
||||
作为Scribus文档团队的长期成员，我随时了解最新的源代码更新，以便对文档进行更新和补充。 我最近在刚升级到Fedora 27系统的计算机上使用Subversion进行“checkout”操作时，对于文档下载所需要的时间我感到很惊讶，文档由HTML页面和相关图像组成。 我担心该项目的文档看起来比项目本身大得多，并且怀疑其中的一些内容是“僵尸”文档——不会再使用的HTML文件以及HTML中无法访问到的图像。
|
||||
|
||||
我决定为自己创建一个项目来解决这个问题。 一种方法是搜索未使用的现有图像文件。 如果我可以扫描所有HTML文件中的图像引用，然后将该列表与实际图像文件进行比较，那么我可能会看到不匹配的文件。
|
||||
|
||||
这是一个典型的图像标签：
|
||||
```
|
||||
<img src="images/edit_shapes.png" ALT="Edit examples" ALIGN=left>
|
||||
```
|
||||
|
||||
我对第一组引号之间的部分很感兴趣，在src =之后。 寻找解决方案后，我找到一个名为[BeautifulSoup][1]的Python模块。 脚本的核心部分如下所示：
|
||||
|
||||
```
|
||||
soup = BeautifulSoup(all_text, 'html.parser')
|
||||
match = soup.findAll("img")
|
||||
if len(match) > 0:
|
||||
for m in match:
|
||||
imagelist.append(str(m))
|
||||
```
|
||||
|
||||
我们可以使用这个`findAll` 方法来挖出图片标签。 这是一小部分输出：
|
||||
|
||||
```
|
||||
<img src="images/pdf-form-ht3.png"/><img src="images/pdf-form-ht4.png"/><img src="images/pdf-form-ht5.png"/><img src="images/pdf-form-ht6.png"/><img align="middle" alt="GSview - Advanced Options Panel" src="images/gsadv1.png" title="GSview - Advanced Options Panel"/><img align="middle" alt="Scribus External Tools Preferences" src="images/gsadv2.png" title="Scribus External Tools Preferences"/>
|
||||
```
|
||||
|
||||
到现在为止还挺好。我原以为下一步就可以搞定了，但是当我在脚本中尝试了一些字符串方法时，它返回了有关标记的错误而不是字符串的错误。 我将输出保存到一个文件中，并在[KWrite][2]中进行编辑。 KWrite的一个好处是你可以使用正则表达式（regex）来做“查找和替换”操作，所以我可以用`\n<img` 替换 `<img`，这样可以看得更清楚。 KWrite的另一个好处是，如果你用正则表达式做了一个不明智的选择，你还可以撤消。
|
||||
|
||||
但我认为，肯定有比这更好的东西，所以我转而使用正则表达式，或者更具体地说Python的 `re` 模块。 这个新脚本的相关部分如下所示：
|
||||
|
||||
```
|
||||
match = re.findall(r'src="(.*)/>', all_text)
|
||||
if len(match)>0:
|
||||
for m in match:
|
||||
imagelist.append(m)
|
||||
```
|
||||
|
||||
它的一小部分输出如下所示：
|
||||
```
|
||||
images/cmcanvas.png" title="Context Menu for the document canvas" alt="Context Menu for the document canvas" /></td></tr></table><br images/eps-imp1.png" title="EPS preview in a file dialog" alt="EPS preview in a file dialog" images/eps-imp5.png" title="Colors imported from an EPS file" alt="Colors imported from an EPS file" images/eps-imp4.png" title="EPS font substitution" alt="EPS font substitution" images/eps-imp2.png" title="EPS import progress" alt="EPS import progress" images/eps-imp3.png" title="Bitmap conversion failure" alt="Bitmap conversion failure"
|
||||
```
|
||||
|
||||
乍一看，它看起来与上面的输出类似，并且附带有修剪部分图像标签的好处，但是有令人费解的是还夹杂着表格标签和其他内容。 我认为这涉及到这个正则表达式`src="(.*)/>`，这被称为贪婪，意味着它不一定停止在遇到`/>`的第一个实例。我应该补充一点，我也尝试过`src="(.*)"`，这真的没有什么更好的效果，不是一个正则表达式专家（只是做了这个），我找了各种方法来改进这一点但是并没什么用。
|
||||
|
||||
做了一系列的事情之后，甚至尝试了Perl的`HTML::Parser`模块，最终我试图将这与我为Scribus编写的一些脚本进行比较，这些脚本逐个字符的分析文本内容，然后采取一些行动。 为了最终目的，我终于想出了所有这些方法，并且完全不需要正则表达式或HTML解析器。 让我们回到展示的那个`img`标签的例子。
|
||||
|
||||
```
|
||||
<img src="images/edit_shapes.png" ALT="Edit examples" ALIGN=left>
|
||||
```
|
||||
|
||||
我决定回到`src=`这一块。 一种方法是等待`s`出现，然后看下一个字符是否是`r`，下一个是`c`，下一个是否`=`。 如果是这样，那就匹配上了！ 那么两个双引号之间的内容就是我所需要的。 这种方法的问题在于需要连续识别上面这样的结构。 一种查看代表一行HTML文本的字符串的方法是：
|
||||
|
||||
```
|
||||
for c in all_text:
|
||||
```
|
||||
|
||||
但是这个逻辑太乱了，以至于不能持续匹配到前面的`c`，还有之前的字符，更之前的字符，更更之前的字符。
|
||||
|
||||
最后，我决定专注于`=`并使用索引方法，以便我可以轻松地引用字符串中的任何先前或将来的字符。 这里是搜索部分：
|
||||
|
||||
```
|
||||
index = 3
|
||||
while index < linelength:
|
||||
if (all_text[index] == '='):
|
||||
if (all_text[index-3] == 's') and (all_text[index-2] == 'r') and (all_text[index-1] == 'c'):
|
||||
imagefound(all_text, imagelist, index)
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
```
|
||||
|
||||
我用第四个字符开始搜索（索引从0开始），所以我在下面没有出现索引错误，并且实际上，在每一行的第四个字符之前不会有等号。 第一个测试是看字符串中是否出现了`=`，如果没有，我们就会前进。 如果我们确实看到一个等号，那么我们会看前三个字符是否是`s`，`r`和`c`。 如果全都匹配了，就调用函数`imagefound`：
|
||||
|
||||
```
|
||||
def imagefound(all_text, imagelist, index):
|
||||
end = 0
|
||||
index += 2
|
||||
newimage = ''
|
||||
while end == 0:
|
||||
if (all_text[index] != '"'):
|
||||
newimage = newimage + all_text[index]
|
||||
index += 1
|
||||
else:
|
||||
newimage = newimage + '\n'
|
||||
imagelist.append(newimage)
|
||||
end = 1
|
||||
return
|
||||
```
|
||||
|
||||
我们正在给函数发送当前索引，它代表着`=`。 我们知道下一个字符将会是`"`，所以我们跳过两个字符，并开始向名为`newimage`的控制字符串添加字符，直到我们发现下一个`"`，此时我们完成了一次匹配。 我们将字符串加一个`换行`符添加到列表`imagelist`中并`返回`，请记住，在剩余的这个HTML字符串中可能会有更多图片标签，所以我们马上回到搜索循环的中间。
|
||||
|
||||
以下是我们的输出现在的样子：
|
||||
```
|
||||
images/text-frame-link.png
|
||||
images/text-frame-unlink.png
|
||||
images/gimpoptions1.png
|
||||
images/gimpoptions3.png
|
||||
images/gimpoptions2.png
|
||||
images/fontpref3.png
|
||||
images/font-subst.png
|
||||
images/fontpref2.png
|
||||
images/fontpref1.png
|
||||
images/dtp-studio.png
|
||||
```
|
||||
|
||||
啊，干净多了，而这只花费几秒钟的时间。 我本可以将索引前移7步来剪切`images/`部分，但我更愿意把这个部分保存下来，确保我没有切片掉图像文件名的第一个字母，这很容易用KWrite编辑成功- - 你甚至不需要正则表达式。 做完这些并保存文件后，下一步就是运行我编写的另一个脚本`sortlist.py`：
|
||||
|
||||
```
|
||||
#!/usr/bin/env python
|
||||
# -*- coding: utf-8 -*-
|
||||
# sortlist.py
|
||||
|
||||
import os
|
||||
|
||||
imagelist = []
|
||||
for line in open('/tmp/imagelist_parse4.txt').xreadlines():
|
||||
imagelist.append(line)
|
||||
|
||||
imagelist.sort()
|
||||
|
||||
outfile = open('/tmp/imagelist_parse4_sorted.txt', 'w')
|
||||
outfile.writelines(imagelist)
|
||||
outfile.close()
|
||||
```
|
||||
|
||||
这会读取文件内容，并存储为列表，对其排序，然后另存为另一个文件。 之后，我可以做到以下几点：
|
||||
|
||||
```
|
||||
ls /home/gregp/development/Scribus15x/doc/en/images/*.png > '/tmp/actual_images.txt'
|
||||
```
|
||||
|
||||
然后我需要在该文件上运行`sortlist.py`，因为`ls`方法的排序与Python不同。 我原本可以在这些文件上运行比较脚本，但我更愿意以可视方式进行操作。 最后，我成功找到了42个图像，这些图像没有来自文档的HTML引用。
|
||||
|
||||
这是我的完整解析脚本：
|
||||
```
|
||||
#!/usr/bin/env python
|
||||
# -*- coding: utf-8 -*-
|
||||
# parseimg4.py
|
||||
|
||||
import os
|
||||
|
||||
def imagefound(all_text, imagelist, index):
|
||||
end = 0
|
||||
index += 2
|
||||
newimage = ''
|
||||
while end == 0:
|
||||
if (all_text[index] != '"'):
|
||||
newimage = newimage + all_text[index]
|
||||
index += 1
|
||||
else:
|
||||
newimage = newimage + '\n'
|
||||
imagelist.append(newimage)
|
||||
end = 1
|
||||
return
|
||||
|
||||
htmlnames = []
|
||||
imagelist = []
|
||||
tempstring = ''
|
||||
filenames = os.listdir('/home/gregp/development/Scribus15x/doc/en/')
|
||||
for name in filenames:
|
||||
if name.endswith('.html'):
|
||||
htmlnames.append(name)
|
||||
#print htmlnames
|
||||
for htmlfile in htmlnames:
|
||||
all_text = open('/home/gregp/development/Scribus15x/doc/en/' + htmlfile).read()
|
||||
linelength = len(all_text)
|
||||
index = 3
|
||||
while index < linelength:
|
||||
if (all_text[index] == '='):
|
||||
if (all_text[index-3] == 's') and (all_text[index-2] == 'r') and
|
||||
(all_text[index-1] == 'c'):
|
||||
imagefound(all_text, imagelist, index)
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
else:
|
||||
index += 1
|
||||
|
||||
outfile = open('/tmp/imagelist_parse4.txt', 'w')
|
||||
outfile.writelines(imagelist)
|
||||
outfile.close()
|
||||
imageno = len(imagelist)
|
||||
print str(imageno) + " images were found and saved"
|
||||
```
|
||||
|
||||
脚本名称为`parseimg4.py`，这并不能真实反映我陆续编写的脚本数量，包括微调的和大改的以及丢弃并重新开始写的。 请注意，我已经对这些目录和文件名进行了硬编码，但是总结起来很容易，要求用户输入这些信息。 同样，因为它们是工作脚本，所以我将输出发送到 `/tmp`目录，所以一旦重新启动系统，它们就会消失。
|
||||
|
||||
这不是故事的结尾，因为下一个问题是：僵尸HTML文件怎么办？ 任何未使用的文件都可能会引用到前面比对方法没有提取到的图像。 我们有一个`menu.xml`文件作为联机手册的目录，但我还需要考虑TOC（译者注：TOC是table of contents的缩写）中列出的某些文件可能引用了不在TOC中的文件，是的，我确实找到了一些这样的文件。
|
||||
|
||||
最后我可以说，这是一个比图像搜索更简单的任务，而且开发的过程对我有很大的帮助。
|
||||
|
||||
|
||||
### About the author
|
||||
|
||||
[][7] Greg Pittman - Greg是Kentucky州Louisville市的一名退休的神经学家，从二十世纪六十年代的Fortran IV语言开始长期以来对计算机和编程有着浓厚的兴趣。 当Linux和开源软件出现的时候，Greg深受启发，去学习更多只是，并实现最终贡献的承诺。 他是Scribus团队的成员。[更多关于我][8]
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://opensource.com/article/18/1/parsing-html-python
|
||||
|
||||
作者：[Greg Pittman][a]
|
||||
译者：[Flowsnow](https://github.com/Flowsnow)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://opensource.com/users/greg-p
|
||||
[1]:https://www.crummy.com/software/BeautifulSoup/
|
||||
[2]:https://www.kde.org/applications/utilities/kwrite/
|
||||
[7]:https://opensource.com/users/greg-p
|
||||
[8]:https://opensource.com/users/greg-p
|
||||
@ -0,0 +1,51 @@
|
||||
Docker 容器中的老式 DOS BBS
|
||||
======
|
||||
不久前，我写了一篇[我的 Debian Docker 基本映像][1]。我决定进一步扩展这个概念：在 Docker 中运行 DOS 程序。
|
||||
|
||||
但首先，一张截图：
|
||||
|
||||
![][2]
|
||||
|
||||
事实证明这是可能的，但很难。我使用了所有三种主要的 DOS 模拟器（dosbox、qemu 和 dosemu）。我让它们都能在 Docker 容器中运行，但有很多有趣的问题需要解决。
|
||||
|
||||
普遍要做的事是在 DOS 环境下提供一个伪造的调制解调器。这需要作为 TCP 端口暴露在容器外部。这有很多方法可以做到 - 我使用的是 tcpser结。dosbox 有一个 TCP 调制解调器接口，但事实证明，这样做太问题太多了。
|
||||
|
||||
挑战来自你希望能够一次接受多个传入 telnet（或 TCP）连接。DOS 不是一个多任务操作系统，所以当时有很多 hack 的方法。一种是有多台物理机，每个有一根传入电话线。或者它们可能会在 [DESQview][3]、OS/2 甚至 Windows 3.1 等多任务层下运行多个伪 DOS 实例。
|
||||
|
||||
（注意：我刚刚了解到 [DESQview/X][4]，它将 DESQview 与 X11R5 集成在一起，并[取代了 Windows 3 驱动程序][5]来使 Windows 作为 X 应用程序运行。）
|
||||
|
||||
出于各种原因，我不想尝试在 Docker 中运行其中一个系统。这让我模拟了原来的多物理节点设置。从理论上讲，非常简单 - 运行一组 DOS 实例，每个实例最多使用 1MB 的模拟 RAM，这就行了。但是这里面临挑战。
|
||||
|
||||
在多物理节点设置中，你需要某种文件共享，因为你的节点需要访问共享消息和文件存储。在老式的 DOS 时代，有很多笨重的方法可以做到这一点 - [Netware][6]、[LAN manager][7]，甚至一些 PC NFS 客户端。我没有访问 Netware。我尝试了 DOS 中的 Microsoft LM 客户端，与在 Docker 容器内运行的 Samba 服务器交互。这样可以使用，但 LM 客户端即使有各种高内存技巧还是占用了很多内存，BBS 软件也无法运行。我无法在多个 dosbox 实例中挂载底层文件系统，因为 dosbox 缓存不兼容。
|
||||
|
||||
这就是为什么我使用 dosemu 的原因。除了有比 dosbox 更完整的模拟器之外，它还有一种共享主机文件系统的方式。
|
||||
|
||||
|
||||
所以, 所有这一切都在此：[jgoerzen/docker-bbs-renegade][8]。
|
||||
|
||||
我还为其他想做类似事情的人准备了构建块：[docker-dos-bbs][9] 和底层 [docker-dosemu][10]。
|
||||
|
||||
意外的收获是，我也试图在 Joyent 的 Triton（基于 Solaris 的 SmartOS）下运行它。让我感到高兴的印象是，这几乎可以在这下工作。是的，在 Solaris 机器上在一个基于 Linux 的 DOS 模拟器的容器中运行 Renegade DOS BBS。
|
||||
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: http://changelog.complete.org/archives/9836-an-old-dos-bbs-in-a-docker-container
|
||||
|
||||
作者：[John Goerzen][a]
|
||||
译者：[geekpi](https://github.com/geekpi)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:http://changelog.complete.org/archives/author/jgoerzen
|
||||
[1]:https://changelog.complete.org/archives/9794-fixing-the-problems-with-docker-images
|
||||
[2]:https://raw.githubusercontent.com/jgoerzen/docker-bbs-renegade/master/renegade-login.png
|
||||
[3]:https://en.wikipedia.org/wiki/DESQview
|
||||
[4]:http://toastytech.com/guis/dvx.html
|
||||
[5]:http://toastytech.com/guis/dvx3.html
|
||||
[6]:https://en.wikipedia.org/wiki/NetWare
|
||||
[7]:https://en.wikipedia.org/wiki/LAN_Manager
|
||||
[8]:https://github.com/jgoerzen/docker-bbs-renegade
|
||||
[9]:https://github.com/jgoerzen/docker-dos-bbs
|
||||
[10]:https://github.com/jgoerzen/docker-dosemu
|
||||
Loading…
Reference in New Issue
Block a user