My personal website
git clone git://
Log | Files | Refs

rss.xml (26154B)

      1 <?xml version="1.0" encoding="UTF-8"?>
      2 <rss version="2.0"
      3     xmlns:atom=""
      4     xmlns:dc=""
      5     xmlns:sy=""
      6 >
      7   <channel>
      8     <title>Cem's Website</title>
      9     <description>Personal blog/website on Linux/tech/nerdy stuff</description>
     10     <link></link>
     11     <atom:link href="" rel="self" type="application/rss+xml" />
     12     <lastBuildDate>Tue Oct 19 2021 12:00</lastBuildDate>
     13 <item>
     14 <title>My thoughts on execline</title>
     15 <pubDate>Tue, 12 Oct 2021</pubDate>
     16 <dc:creator>Cem Keylan</dc:creator>
     17 <link></link>
     18 <description>&lt;h1&gt;My thoughts on execline&lt;/h1&gt;
     20 &lt;p&gt;With the gaining popularity of the &lt;a href=&quot;;&gt;s6-rc&lt;/a&gt;, I have recently decided to check out
     21 &lt;a href=&quot;;&gt;execline&lt;/a&gt;. It is a shell-like scripting language that is built around chaining
     22 commands together. There is no interpreter, there is no &amp;lsquo;built-in&amp;rsquo; functionality
     23 even though the docs might make you think there are. Execline is best described
     24 as a suite of tools that imitate the functionality of a shell.&lt;/p&gt;
     26 &lt;p&gt;There are a ton of information on the execline&amp;rsquo;s page, especially discussing
     27 why skarnet prefers execline instead of &lt;code&gt;/bin/sh&lt;/code&gt;. Those points are mostly
     28 valid, shells are a nightmare, and they suck at being compliant to the POSIX
     29 specification. What I don&amp;rsquo;t agree with the &lt;a href=&quot;;&gt;why not sh&lt;/a&gt; page, however, is the
     30 part on performance. Even though execline does not have an interactive shell
     31 implementation of its own, it is still much slower than other shells simply by
     32 its design. Since the design of execline is built on process chaining, it
     33 requires spawning new processes for things as basic as variable declaration.
     34 Variable manipulation is the cheapest operation you would expect from a shell,
     35 but in execline, every operation costs the same regardless of how simple it is.&lt;/p&gt;
     37 &lt;p&gt;Throughout my weeks of toying around with execline, I have came to the
     38 conclusion that execline is much better in simple scripts only. Execline is
     39 as usable as any shell out there, but even with its advantages over &lt;code&gt;sh&lt;/code&gt;,
     40 execline is only better if it&amp;rsquo;s simple. Execline is really good for certain
     41 specific situations such as service scripts (as used in s6-rc), or where you
     42 were already meant to chain a couple of commands together. After all, skarnet
     43 already presents these limitations on the website of execline.&lt;/p&gt;
     45 &lt;p&gt;Execline can be leveraged as how s6-rc compiles service databases with other
     46 utilities, but I don&amp;rsquo;t think it can be used as a shell replacement itself. It&amp;rsquo;s
     47 not the next shiny thing to jump on to replace all your shell scripts with
     48 (unless you have really basic shell scripts). It does not have the flexibility
     49 nor the performance of the shell for scripts that can be considered a little
     50 over than the &amp;ldquo;basic&amp;rdquo;.&lt;/p&gt;</description>
     51 </item>
     52 <item>
     53 <title>Reimplementing `sysmgr` in C</title>
     54 <pubDate>Fri, 02 Oct 2020</pubDate>
     55 <dc:creator>Cem Keylan</dc:creator>
     56 <link></link>
     57 <description>&lt;h1&gt;Reimplementing &lt;code&gt;sysmgr&lt;/code&gt; in C&lt;/h1&gt;
     59 &lt;p&gt;For a while, I have been thinking about implementing &lt;a href=&quot;;&gt;sysmgr&lt;/a&gt; in C. I started
     60 thinking about the inefficiencies of sysmgr. POSIX sh isn&amp;rsquo;t particularly
     61 designed to have ultimate control over child processes. There are basic job
     62 management features that are &lt;em&gt;just enough&lt;/em&gt; for sysmgr to do its job. The
     63 biggest pain is having to use tools like &lt;code&gt;sleep(1)&lt;/code&gt; and &lt;code&gt;kill(1)&lt;/code&gt;. Calling
     64 sleep every second, and using the kill command to check whether a process is
     65 alive or not is extremely inefficient. Some shells &lt;em&gt;do&lt;/em&gt; include these commands
     66 built-in, but it isn&amp;rsquo;t specified by POSIX, but one should never take this as
     67 granted.&lt;/p&gt;
     69 &lt;p&gt;Lately, I have been adding C utilities to sysmgr to make it more efficient. This
     70 defeats the initial purpose of sysmgr, being a service manager in pure POSIX
     71 shell. My main purpose, however, is making sysmgr efficient and simplistic. It
     72 mostly imitates &lt;code&gt;runit&lt;/code&gt; without dealing with all the complexity of the
     73 over-thinked &lt;code&gt;supervise&lt;/code&gt; directory, nor the logging stuff. Most of these can be
     74 handled by the service script itself anyway. That&amp;rsquo;s why instead of this ugly
     75 C/POSIX sh hybrid, I decided to implement it all in C.&lt;/p&gt;
     77 &lt;p&gt;I am not a C expert or anything, I am learning a lot as I am writing the
     78 program. I want it to be C99 and portable (for BSD). It&amp;rsquo;s currently not
     79 functional at all, but, you can see its current state &lt;a href=&quot;;&gt;here&lt;/a&gt;.&lt;/p&gt;
     81 &lt;p&gt;EDIT Oct 10 2020:&lt;/p&gt;
     83 &lt;p&gt;I did the initial release of this C version of sysmgr, which is more stable,
     84 and performant than the POSIX sh version. It still has rough edges, but is
     85 completely usable.&lt;/p&gt;</description>
     86 </item>
     87 <item>
     88 <title>Trust in Distributed Environments</title>
     89 <pubDate>Tue, 08 Sep 2020</pubDate>
     90 <dc:creator>Cem Keylan</dc:creator>
     91 <link></link>
     92 <description>&lt;h1&gt;Trust in Distributed Environments&lt;/h1&gt;
     94 &lt;p&gt;A few days ago, in the &lt;code&gt;#kisslinux&lt;/code&gt; IRC channel, jedahan mentioned an
     95 implementation for trust in the package manager. While I was intrigued by the
     96 idea initially, I decided not to implement this for the upcoming 4.0.0 release.
     97 That is because the package manager and the distribution itself is already
     98 centered on trust. However, this idea made me think a lot about &amp;ldquo;trust&amp;rdquo; in
     99 distributed environments.&lt;/p&gt;
    101 &lt;p&gt;Who and what would you trust? Would you trust Microsoft? Would you trust a
    102 binary? Would you only trust a so called &amp;ldquo;reproducible&amp;rdquo; build?&lt;/p&gt;
    104 &lt;p&gt;Jedahan mentioned the possibility that a repository maintainer could create a
    105 package that would be normally in the distribution so they could mess your
    106 system up. He suggested a &amp;ldquo;source&amp;rdquo; system where you know where each package
    107 comes from. This way the package manager can warn you when the source of a
    108 package is changed. As I have said this idea intrigued me at the beginning, but
    109 here is why it is complex and unnecessary.&lt;/p&gt;
    111 &lt;p&gt;The package manager would warn you every time you fork a package and apply your
    112 changes. Both with kiss and CPT, you already see git logs when the repositories
    113 are fetched. Those logs address each and every file that has been edited, added,
    114 removed, or renamed. CPT also has support for rsync, which is called verbosely.
    115 While not as descriptive, rsync also shows what&amp;rsquo;s changed/added and what&amp;rsquo;s
    116 deleted.&lt;/p&gt;
    118 &lt;p&gt;Also, back on April, I have added submodule support to my fork of kiss, which
    119 Dylan adapted on May 19. I have added this feature because it solves a similar
    120 issue. I want to have only some packages from a repository and leave the rest
    121 of them. This way I am the one in control of what goes inside my repositories.&lt;/p&gt;
    123 &lt;p&gt;Minor annoyances aside, would this solve the issue of trust? Maybe this evil
    124 repository maintainer decides to botch a package that was already in their
    125 repository not provided by your distribution. Should we then track the source
    126 files, build files as well? But those change all the time.&lt;/p&gt;
    128 &lt;p&gt;I believe that this environment is as trustworthy as it can get, a repository
    129 system with package build instructions that easy to read and understand, easy to
    130 history check, easy to limit, and easy to allow. KISS and Carbs Linux have a
    131 single maintainer. I maintain Carbs and Dylan maintains KISS. You are not
    132 trusting an organization, you are trusting individuals that you can easily
    133 converse on the internet. The same goes for most community repository
    134 maintainers out there. Trying to implement more would be a &amp;ldquo;security theater&amp;rdquo;
    135 that would be a hassle for the maintainers, the users and the package manager
    136 without a noticeable benefit to any.&lt;/p&gt;</description>
    137 </item>
    138 <item>
    139 <title>wpa_add script</title>
    140 <pubDate>Fri, 28 Aug 2020</pubDate>
    141 <dc:creator>Cem Keylan</dc:creator>
    142 <link></link>
    143 <description>&lt;h1&gt;wpa_add script&lt;/h1&gt;
    145 &lt;p&gt;I have this script named &lt;code&gt;wpa_add&lt;/code&gt;, which I use to easily add new WiFi when I
    146 am outside, possibly in a cafe. I have written this script because I don&amp;rsquo;t like
    147 the way my girlfriend looks at me while thinking that I am an absolute moron for
    148 not using Windows 10, and the entirety of Linux is a circlejerk. It is only
    149 natural that she thinks this way. I use my own distribution that doesn&amp;rsquo;t have
    150 things like &lt;code&gt;dbus&lt;/code&gt;, or &lt;code&gt;NetworkManager&lt;/code&gt;, or one of those common desktop
    151 environments. You could install it by creating a simple package, but I am happy
    152 to not have any of those in my system.&lt;/p&gt;
    154 &lt;p&gt;This script uses wpa-supplicant to add a new network and reconfigure. It uses
    155 dmenu for input, however you could replace dmenu calls with some command line
    156 prompts. I am doing the following assumptions:
    157 - You can manipulate &lt;code&gt;wpa_supplicant&lt;/code&gt; without root access.
    158 - The configuration is on &lt;code&gt;/etc/wpa_supplicant.conf&lt;/code&gt;.
    159 - You can edit &lt;code&gt;/etc/wpa/supplicant.conf&lt;/code&gt;.&lt;/p&gt;
    161 &lt;p&gt;If you want to ensure the above just do the following (as root):&lt;/p&gt;
    163 &lt;pre&gt;&lt;code class=&quot;sh&quot;&gt;# Add yourself to the wheel group if you aren&apos;t already.
    164 adduser user wheel
    166 # Change the ownership of /etc/wpa_supplicant.conf
    167 chown root:wheel /etc/wpa_supplicant.conf
    169 # Make sure the configuration can be edited by the wheel group.
    170 chmod 664 /etc/wpa_supplicant.conf
    171 &lt;/code&gt;&lt;/pre&gt;
    173 &lt;p&gt;Your &lt;code&gt;wpa_supplicant&lt;/code&gt; configuration must include the following line (or something similar):&lt;/p&gt;
    175 &lt;pre&gt;&lt;code class=&quot;plaintext&quot;&gt;ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wheel
    176 &lt;/code&gt;&lt;/pre&gt;
    178 &lt;p&gt;Here is the script&lt;/p&gt;
    180 &lt;pre&gt;&lt;code class=&quot;sh&quot;&gt;#!/bin/sh
    181 # Script to add wpa_supplicant networks through dmenu
    183 if [ &quot;$1&quot; ]; then
    184     name=$1
    185 else
    186     name=$(dmenu -p &quot;Please enter network name, leave empty if you want to search&quot; &amp;lt;&amp;amp;-)
    187 fi
    189 [ &quot;$name&quot; ] || {
    190     wpa_cli scan
    191     name=$(
    192     wpa_cli scan_results | sed 1,2d | while read -r _ _ _ _ ssid _; do
    193         # Hidden wifi are not to be returned
    194         [ &quot;$ssid&quot; ] || continue
    195         echo &quot;$ssid&quot;
    196     done | sort -u | dmenu -l 10 -p &quot;Please choose WiFi&quot;)
    197     [ &quot;$name&quot; ] || exit 1
    198 }
    200 pass=$(dmenu -P -p &quot;Please enter your password, leave empty if the network has open access.&quot;)
    202 if [ &quot;$pass&quot; ]; then
    203     wpa_passphrase &quot;$name&quot; &amp;lt;&amp;lt;EOF&amp;gt;&amp;gt; /etc/wpa_supplicant.conf
    204 $pass
    205 EOF
    206 else
    207     printf &apos;network={\n\tssid=&quot;%s&quot;\n\tkey_mgmt=NONE\n\tpriority=-999\n}\n&apos; &quot;$name&quot; &amp;gt;&amp;gt; /etc/wpa_supplicant.conf
    208 fi
    210 wpa_cli reconfigure
    211 &lt;/code&gt;&lt;/pre&gt;
    213 &lt;p&gt;As I have said, you could do something similar in a command-line-only tool as
    214 well. This one uses &lt;code&gt;fzf&lt;/code&gt; on WiFi selection.&lt;/p&gt;
    216 &lt;pre&gt;&lt;code class=&quot;sh&quot;&gt;#!/bin/sh -e
    218 stty=&quot;$(stty -g)&quot;
    219 trap &quot;stty $stty&quot; EXIT INT TERM HUP
    221 if [ &quot;$1&quot; ]; then
    222     name=$1
    223 else
    224     printf &apos;Network Name, leave empty if you want to search: &apos;
    225     read -r name
    226 fi
    228 [ &quot;$name&quot; ] || {
    229     wpa_cli scan &amp;gt;/dev/null
    230     name=$(
    231     wpa_cli scan_results | sed 1,2d | while read -r _ _ _ _ ssid _; do
    232         # Hidden wifi are not to be returned
    233         [ &quot;$ssid&quot; ] || continue
    234         echo &quot;$ssid&quot;
    235     done | sort -u | fzf --prompt &quot;Please choose WiFi: &quot;)
    236 }
    238 [ &quot;$name&quot; ] || exit 1
    240 stty -echo
    241 printf &apos;Please enter your password, leave empty if the network has open access.\nPassword: &apos;
    242 read -r pass
    244 if [ &quot;$pass&quot; ]; then
    245     wpa_passphrase &quot;$name&quot; &amp;lt;&amp;lt;EOF&amp;gt;&amp;gt; /etc/wpa_supplicant.conf
    246 $pass
    247 EOF
    248 else
    249     printf &apos;network={\n\tssid=&quot;%s&quot;\n\tkey_mgmt=NONE\n\tpriority=-999\n}\n&apos; &quot;$name&quot; &amp;gt;&amp;gt; /etc/wpa_supplicant.conf
    250 fi
    252 wpa_cli reconfigure
    253 &lt;/code&gt;&lt;/pre&gt;
    255 &lt;p&gt;These scripts can be found as a gist &lt;a href=&quot;;&gt;here&lt;/a&gt;&lt;/p&gt;</description>
    256 </item>
    257 <item>
    258 <title>Static linking</title>
    259 <pubDate>Fri, 28 Aug 2020</pubDate>
    260 <dc:creator>Cem Keylan</dc:creator>
    261 <link></link>
    262 <description>&lt;h1&gt;Static linking&lt;/h1&gt;
    264 &lt;p&gt;While I was working on a new initramfs generator for Carbs, I was once again
    265 reminded of the advantages of static linking software. Previously, I was using
    266 some really dumb script that was just basically using the package manager as a
    267 library for building the whole initramfs system from scratch. This system
    268 structure was completely statically linked, and the whole thing weighed around
    269 1.3MiB.&lt;/p&gt;
    271 &lt;p&gt;Now, while &lt;code&gt;rd&lt;/code&gt; (the small script that I had written) was good enough for me, it
    272 wouldn&amp;rsquo;t be a fit to distribute with the system. It doesn&amp;rsquo;t deal with dynamic
    273 binaries, kernel modules or library installation. So I have written this script
    274 that deals with those (kernel modules aren&amp;rsquo;t done yet, though).&lt;/p&gt;
    276 &lt;p&gt;The issue with build systems today are that the binaries are built dynamically
    277 unless you build the whole thing static. As long as there are shared libraries,
    278 the binaries will be dynamic as well. That&amp;rsquo;s why the core repository of Carbs
    279 still contains dynamic binaries for gcc, binutils, util-linux and some other
    280 packages.&lt;/p&gt;
    282 &lt;p&gt;The size of the new image with exactly the same binaries is a whopping 1.9MiB.
    283 While a size increase of 600KiB might not seem like a huge deal, I just want to
    284 tell you that busybox is static in both images, leaving ONLY TWO binaries that
    285 I install to my image; fsck and e2fsck. By switching from a static binary to
    286 dynamic + lib for only two programs, you require 600 KiB more space, and I have
    287 been talking about a gzip compressed cpio archive throughout this whole post.&lt;/p&gt;</description>
    288 </item>
    289 <item>
    290 <title>Starting X without Xinit</title>
    291 <pubDate>Wed, 12 Aug 2020</pubDate>
    292 <dc:creator>Cem Keylan</dc:creator>
    293 <link></link>
    294 <description>&lt;h1&gt;Starting X without Xinit&lt;/h1&gt;
    296 &lt;p&gt;Most people who don&amp;rsquo;t use a desktop environment use the &lt;code&gt;startx&lt;/code&gt; command to
    297 initialize their X windowing system. Now, &lt;code&gt;startx&lt;/code&gt; is a shell script that runs
    298 the C program &lt;code&gt;xinit&lt;/code&gt; which basically runs &lt;code&gt;xorg-server&lt;/code&gt;. Using xinit obviously
    299 has some nice perks. It makes some checks and runs your .xinitrc file. We don&amp;rsquo;t
    300 need any of that though. Here is my X launcher:&lt;/p&gt;
    302 &lt;pre&gt;&lt;code class=&quot;sh&quot;&gt;#!/bin/sh
    304 export DISPLAY=${DISPLAY:-:0}
    305 trap &quot;$HOME/.xinitrc&quot; USR1
    307 (
    308     trap &apos;&apos; USR1
    310     exec X -keeptty :0 vt1
    311 ) &amp;amp;
    313 wait
    314 &lt;/code&gt;&lt;/pre&gt;
    316 &lt;p&gt;You need to keep in mind that your .xinitrc should be an executable.&lt;/p&gt;</description>
    317 </item>
    318 <item>
    319 <title>Why I dislike Arch and Gentoo</title>
    320 <pubDate>Fri, 08 May 2020</pubDate>
    321 <dc:creator>Cem Keylan</dc:creator>
    322 <link></link>
    323 <description>&lt;h1&gt;Why I dislike Arch and Gentoo&lt;/h1&gt;
    325 &lt;p&gt;Over the years, I have used many many Linux distributions. The reason I am
    326 now using a distribution maintained by me, is because I am never truly satisfied
    327 about other people&amp;rsquo;s work. Not that they are bad at they do, it&amp;rsquo;s just that
    328 they don&amp;rsquo;t share the same vision as me. And I have felt this way with almost
    329 every distribution I have used.&lt;/p&gt;
    331 &lt;h2&gt;Arch Linux&lt;/h2&gt;
    333 &lt;p&gt;Arch Linux itself feels like it became a &amp;lsquo;meme distribution&amp;rsquo;. Their user-base
    334 is a cult-like community that think they are superior for using Arch Linux.
    335 Now, you might be an Arch user, and might not be like this. I used Arch for
    336 a long time, and didn&amp;rsquo;t feel this way, ever. I only see this level of cultism
    337 for Arch and systemd.&lt;/p&gt;
    339 &lt;p&gt;If you ever call Arch bloated on an online community website, you will get
    340 lots of crap for it. But in fact, Arch Linux is bloated. Now this isn&amp;rsquo;t due
    341 to having too many packages in the base installation. This is because of their
    342 packaging of software.&lt;/p&gt;
    344 &lt;p&gt;Arch enables almost every single option in the package configuration, meaning
    345 lots of unnecessary dependencies, and packages with humongous sizes.&lt;/p&gt;
    347 &lt;p&gt;Pacman is a rather slow package manager, and missing alternatives. For me,
    348 an alternatives system is a must.&lt;/p&gt;
    350 &lt;p&gt;If you want to use a better binary distribution, use Void Linux. They have
    351 saner package configurations, and the environment just feels more UNIXy. xbps
    352 is really fast, and has an alternatives system.&lt;/p&gt;
    354 &lt;h2&gt;Gentoo Linux&lt;/h2&gt;
    356 &lt;p&gt;This will be the longer part, because my dislike for Gentoo is bigger than
    357 my dislike towards Arch. If you want to see how NOT to maintain a distribution,
    358 check out Gentoo.&lt;/p&gt;
    360 &lt;p&gt;I&amp;rsquo;ve used Gentoo for a few months, and I&amp;rsquo;m saying this right out of the
    361 gate. Portage is the slowest piece of software that I have ever used on
    362 Linux. Maybe that&amp;rsquo;s because I deliberately avoid software using Python,
    363 but Portage is most probably the slowest package manager that is being
    364 used.&lt;/p&gt;
    366 &lt;p&gt;Portage depends on Bash, Python, and GNU wget. I have got a line count from
    367 &lt;code&gt;cloc&lt;/code&gt;, doing a &lt;code&gt;find . \( -name &apos;*.sh -o -name &apos;*.py&apos; \) -exec cloc {} +&lt;/code&gt;.
    368 The source code of just &lt;code&gt;*.sh&lt;/code&gt; and &lt;code&gt;*.py&lt;/code&gt; files are over 100k lines of code.
    369 Then I got curious and runned cloc against the whole repository. Here is
    370 the output.&lt;/p&gt;
    372 &lt;pre&gt;&lt;code&gt;--------------------------------------------------------------------------------
    373 Language                      files          blank        comment           code
    374 --------------------------------------------------------------------------------
    375 Python                          703          20009          21411         102180
    376 Bourne Shell                     13            643            678           3911
    377 Bourne Again Shell               44            583            434           3172
    378 diff                             17             31            298            574
    379 YAML                              6             32             80            573
    380 XSD                               1             27             27            494
    381 C                                 2             56            128            291
    382 make                              1              7              6             19
    383 INI                               1              1              0             15
    384 reStructuredText                  1              5              4              9
    385 XSLT                              1              0              0              5
    386 --------------------------------------------------------------------------------
    387 SUM:                            790          21394          23066         111243
    388 --------------------------------------------------------------------------------
    389 &lt;/code&gt;&lt;/pre&gt;
    391 &lt;p&gt;That&amp;rsquo;s quite a lot.&lt;/p&gt;
    393 &lt;p&gt;Portage is a package manager that tries to ease the configuration process of
    394 packages, but at the process makes it terribly complex to compose packages,
    395 and adds billions of portage configuration options. Configuring your first
    396 kernel is literally easier than configuring portage in a way you want. Users
    397 just do not know that they would be better off doing an LFS build for a much
    398 stabler system. My system was broken countless times while using Gentoo.
    399 Maintaining a Gentoo system is honestly harder than maintaining my own
    400 distribution.&lt;/p&gt;
    402 &lt;p&gt;&lt;strong&gt;EAPI&lt;/strong&gt;, probably the worst thing about the Portage ecosystem. It is the most
    403 complex, hard to read, hard to learn packaging system ever made. Creating a
    404 USE flag system shouldn&amp;rsquo;t have been this hard.&lt;/p&gt;</description>
    405 </item>
    406 <item>
    407 <title>Editor Wizardry</title>
    408 <pubDate>Mon, 13 Apr 2020</pubDate>
    409 <dc:creator>Cem Keylan</dc:creator>
    410 <link></link>
    411 <description>&lt;h1&gt;Editor Wizardry&lt;/h1&gt;
    413 &lt;p&gt;To this day, I have tried lots of IDEs and text editors. Visual Studio, PyCharm,
    414 Sublime, Notepad++, Vim, Emacs, Pico, Atom, etc. The list goes on. I have even
    415 unironically used ed, and ironically used cat for a while.&lt;/p&gt;
    417 &lt;p&gt;I have settled down after years and years of &amp;ldquo;editor-hopping&amp;rdquo;. I now have 3
    418 main editors that I use on a daily basis! Yeah, you have read it correct. I use
    419 3 editors on a daily basis. Those are,&lt;/p&gt;
    421 &lt;ul&gt;
    422 &lt;li&gt;sed&lt;/li&gt;
    423 &lt;li&gt;vi (not vim)&lt;/li&gt;
    424 &lt;li&gt;emacs&lt;/li&gt;
    425 &lt;/ul&gt;
    428 &lt;h2&gt;Emacs&lt;/h2&gt;
    430 &lt;p&gt;Emacs is a beast. Defining Emacs as a text-editor is wrong. It is a lisp
    431 interpreter, with text manipulation abilities.&lt;/p&gt;
    433 &lt;p&gt;Now, I do like the concept of Integrated Development Environments. It&amp;rsquo;s a shame
    434 that all of them suck. With Emacs I can fine-tune everything according to my
    435 taste, install the packages I need, configure them the way I like. With IDEs you
    436 get some nice plugins, and a tiny bit of customization, but that&amp;rsquo;s it. You get
    437 an environment limited by the vision of someone else. Not to mention that most
    438 IDEs are proprietary software.&lt;/p&gt;
    440 &lt;p&gt;I have stopped using Vim, because it is only meant to be a text editor. You can
    441 extend its features with plugins, but you can really see the impact with just a
    442 few of them. Vimscript is also really primitive, that&amp;rsquo;s why people write plugins
    443 with Python, JS, and such. This further affects the speed of Vim. Most Emacs
    444 packages I have encountered are written in pure lisp. I have over 70 packages,
    445 yet my load time and overall speed is better than when I had Vim with 8 plugins.&lt;/p&gt;
    447 &lt;h3&gt;Cons&lt;/h3&gt;
    449 &lt;ul&gt;
    450 &lt;li&gt;&lt;strong&gt;It comes with too many unnecessary features&lt;/strong&gt; Emacs comes with multiple IRC
    451 clients, a mail reader, rss reader etc. I don&amp;rsquo;t think they are badly
    452 implemented or anything, I would just prefer building up as I want to instead.&lt;/li&gt;
    453 &lt;li&gt;&lt;strong&gt;The defaults are not intuitive&lt;/strong&gt; Now, I haven&amp;rsquo;t actually tried any of them,
    454 but there is a reason &amp;ldquo;Emacs distributions&amp;rdquo;, such as &amp;ldquo;Spacemacs&amp;rdquo;, &amp;ldquo;DOOM
    455 Emacs&amp;rdquo;, &amp;ldquo;Centaur&amp;rdquo; exist. The base of Emacs, even with its unnecessary
    456 features, is unintuitive and complicated.&lt;/li&gt;
    457 &lt;/ul&gt;
    460 &lt;p&gt;Also, let&amp;rsquo;s not forget that Emacs uses an ancient Lisp dialect.&lt;/p&gt;
    462 &lt;h2&gt;Vi&lt;/h2&gt;
    464 &lt;p&gt;I mostly use Emacs when I am dealing with projects. If my aim is to just make
    465 simple changes when I am on the terminal, I just pop up vi provided by busybox.
    466 I just like that it is fast and featureless. It barely gets the job done, and
    467 that&amp;rsquo;s why I like it.&lt;/p&gt;
    469 &lt;h3&gt;Cons&lt;/h3&gt;
    471 &lt;ul&gt;
    472 &lt;li&gt;&lt;strong&gt;No syntax highlighting&lt;/strong&gt; Syntax highlighting is an important feature for me
    473 but I have learned to live without it. Since I don&amp;rsquo;t edit large files with it,
    474 this is not the biggest con.&lt;/li&gt;
    475 &lt;li&gt;&lt;strong&gt;Hard to configure&lt;/strong&gt; Busybox vi only has a limited featureset, which makes
    476 it hard to configure. It doesn&amp;rsquo;t read an &lt;code&gt;rc&lt;/code&gt; file, it uses the &lt;code&gt;$EXINIT&lt;/code&gt;
    477 variable instead. Available options are limited. For example, you cannot
    478 convert the &amp;ldquo;tab&amp;rdquo; action to use space instead of tabs.&lt;/li&gt;
    479 &lt;li&gt;&lt;strong&gt;No visual selection support&lt;/strong&gt; Sadly, &lt;code&gt;v/V&lt;/code&gt; isn&amp;rsquo;t implemented in busybox vi.&lt;/li&gt;
    480 &lt;/ul&gt;
    483 &lt;h2&gt;Sed&lt;/h2&gt;
    485 &lt;p&gt;I use sed for when I am making small changes to small files, because it is
    486 faster than opening a file, making a change, saving, and exiting. Using regular
    487 expressions are much faster and efficient at such things.&lt;/p&gt;
    489 &lt;h3&gt;Cons&lt;/h3&gt;
    491 &lt;ul&gt;
    492 &lt;li&gt;&lt;strong&gt;Risky unless you know what you are doing&lt;/strong&gt; Since sed is operated on regex,
    493 you need to be extra careful. You are running that regex on the entire file
    494 without an option to &amp;lsquo;undo&amp;rsquo; (unless you have a sed that has a backup
    495 implementation).&lt;/li&gt;
    496 &lt;/ul&gt;</description>
    497 </item>
    498   </channel>
    499 </rss>