reimplement website using shell scripts

major change, i know. now i'm using ssg and rssg by roman zolotarev.
okay, well, not exactly. ssg is modified to generate pages for gemini
too. it's hard to maintain two different things simultaneously.

bye-bye jekyll!

7 files changed, 0 insertions, 665 deletions

diff --git a/_posts/2020-12-06-fossil-to-git.md b/_posts/2020-12-06-fossil-to-git.md
deleted file mode 100644
index cacb4c9..0000000
--- a/_posts/2020-12-06-fossil-to-git.md
+++ /dev/null
@@ -1,68 +0,0 @@
----
-title: fossil export to git
-author: la-ninpre
-tags: fossil git tutorial
----
-
-i was trying to export my website repo to fossil using suggested method from
-[fossil website][1]:
-
-```
-git fast-export --all | fossil import --git repo.fossil
-```
-
-[1]:https://www.fossil-scm.org/home/doc/trunk/www/inout.wiki
-
-but i didn't like that fossil recognizes my email as username and so commit
-messages user was `user@example.com` instead of `user`.
-
-<!--more-->
-
-i then read a bit about options of `git fast-export` and found `--anonymize`
-flag. but it's results weren't satisfying either.
-
-when i looked on a raw output of `git fast-export`, i noticed that commit author
-is specified there as 
-
-```
-author user <user@example.com>
-```
-
-and then it's flashed in my head: why not pipe git export through sed and just 
-replace the contents of `<>` with username instead of email.
-
-so the final command looks like this:
-
-```
-git fast-export --all | \
-    sed -E 's/^((author)|(committer))[[:blank:]]+([[:graph:]]+)[[:blank:]]+(<[[:alnum:]]+@[[:alnum:]]+\.[[:alnum:]]+>)/\1 \4 <\4>/' | \
-    fossil import --git repo.fossil
-```
-
-and it converts
-
-```
-author user <user@example.com>
-```
-
-to 
-
-```
-author user <user>
-```
-
-which is odd, but fine for fossil import.
-
----
-
-update: i tested this on a bigger repo with older history and found that this
-regexp was not perfect, i updated it to handle situations like
-`user@example.co.uk` and also names that consist of more than one word.
-
-```
-git fast-export --all | \
-    sed -E 's/^((author)|(committer))[[:blank:]]+([[:graph:]]+([[:blank:]]+[[:graph:]]+)*)[[:blank:]]+(<[[:graph:]]+@[[:graph:]]+(\.[[:graph:]]+)+>)/\1 \4<\4>/' | \
-    fossil import --git repo.fossil
-```
-
-it's veery evil looking horrible thing, but it works.
diff --git a/_posts/2020-12-09-fossil-autoupdate-cronjob.md b/_posts/2020-12-09-fossil-autoupdate-cronjob.md
deleted file mode 100644
index 70f8fc8..0000000
--- a/_posts/2020-12-09-fossil-autoupdate-cronjob.md
+++ /dev/null
@@ -1,67 +0,0 @@
----
-title: auto-update fossil using cron(8)
-tags: openbsd tutorial fossil
-date: 2020-12-09T01:37+03:00
----
-
-i'm running an instance of fossil on my openbsd server (it's the same that is
-powering this website) and for some reason i want it to be up-to-date. more
-presicely, bleeding edge.
-
-<!--more-->
-
-for that i added this part to my `daily.local` script (for those of you who
-don't know, it's script that running every day by cron(8)):
-
-```
-cd /root/fossil && \
-    /usr/local/bin/fossil up | \
-    awk '/changes:/ {
-        if ($2 == "None."){
-            print "No changes, exiting...";
-            exit 1
-        }else{
-            out="";
-            for(i=2; i<=NF; i++){
-                out=out" "$i
-            };
-        }
-        print out;
-        exit 0
-    }' && \
-    /usr/local/bin/fossil revert src/repolist.c >/dev/null && \
-    patch src/repolist.c /var/www/htdocs/fsl.aaoth.xyz/repolist.c.patch \
-        >/dev/null && \
-    ./configure --static >/dev/null && \
-    make >/dev/null && \
-    cp fossil /var/www/bin && \
-    make distclean >/dev/null && \
-    /usr/local/bin/fossil stat
-```
-
-it is very straightforward and simple. firstly, it's changing directory into
-place, where i have fossil checkout (made with `fossil clone` and
-`fossil open`). then it runs `fossil up` and piping it to a small awk script
-that is checking, is there any changes pulled down.
-
-after that there's one interesting part. `fossil revert src/repolist.c` is there
-because i modified it a little bit to make my [repolist][1] page look better.
-after my edits, i exported a patch by executing:
-
-```
-fossil diff > repolist.c.patch
-```
-
-maybe it would be better if i committed those changes, but i don't want to hold
-a full fossil repo among my other fossils, because its history is fairly long.
-and also i'm not very good at c programming, so i'll keep it as is for now.
-if you're interested this patch is free to use and you can [check it out][2].
-
-after that, there's just a normal configure and make procedure and also final
-cleanup.
-
-i also have a mail server running there, so i get an email of what changes were
-applied and that everything went fine.
-
-[1]:https://fsl.aaoth.xyz
-[2]:https://fsl.aaoth.xyz/repolist.c.patch
diff --git a/_posts/2021-01-06-run-cgit-on-obsd.md b/_posts/2021-01-06-run-cgit-on-obsd.md
deleted file mode 100644
index e1c74bd..0000000
--- a/_posts/2021-01-06-run-cgit-on-obsd.md
+++ /dev/null
@@ -1,272 +0,0 @@
----
-title: run cgit on openbsd
-date: 2021-01-06T13:30+03:00
-author: la-ninpre
-tags: openbsd git tutorial
----
-
-i started using git for my personal pet projects. a little time then i decided
-that it would be a nice idea to make them as open as i can. i use github, but
-to support that idea of self-hosting, i wanted some free and easy web frontend
-to git. cgit is one of the most popular ones, but it was kinda tough to run on
-openbsd.
-
-<!--more-->
-
-i know gitweb exists, but i just like cgit more.
-
-cgit is quite easy to install still, but needs some work done, it's not like
-two commands.
-
-[official cgit page][1] has some installation instructions. it mentions that
-it's distributed in binary form for some linux distros, but of course openbsd's
-not there, so we'll need to build it from source.
-
-first of all, i wanted to use openbsd's native web server`httpd` and native
-fastcgi server -- `slowcgi`.
-the issue is that openbsd's httpd web server lives in chroot-jail and that fact
-is complicating the configuration.
-but before we need to build cgit from source.
-
-## building cgit
-
-to do that, clone cgit source code from [official cgit page][1]
-(or if you like [my cgit page][2], you can clone it instead):
-
-```
-$ git clone https://git.zx2c4.com/cgit
-```
-
-then move to this directory:
-
-```
-$ cd cgit
-```
-
-there are some source files and a Makefile. by default, it'll install cgit in
-`/var/www/htdocs/cgit`. if you want a different path, make corresponding change
-in a Makefile by editing `CGIT_SCRIPT_PATH` variable.
-
-but before compiling cgit itself, we need to init and build git submodule
-(i suppose, this is the source code of git itself that is needed to make some
-git operations on repositories):
-
-```
-$ git submodule init
-$ git submodule update
-```
-
-and then we can compile cgit. note that gnu version of make utility is used,
-install `gmake` from openbsd repositories (`doas pkg add gmake`):
-
-```
-$ gmake && doas gmake install
-```
-
-notice that this command should be executed by user who has write permissions
-to `/var/www` and `/usr/local/lib` (usually root).
-
-when it's done, the rest is to create directories needed for cgit to work
-and also to configure `httpd` and `slowcgi`.
-
-## creating directories and dev/null
-
-cgit uses following files to work:
-- `/etc/cgitrc` -- needed for configuration
-  (see [man page][3] for available options)
-- `/var/cache/cgit` -- cache that is used by cgit to reduce cpu usage
-- `/var/www/htdocs/cgit/`
-  - `cgit.css` -- stylesheet
-  - `cgit.png` -- logo
-  - `favicon.ico` -- favicon
-  - `robots.txt` -- instructions for indexers
-- `/usr/local/lib/cgit/*` -- different filters and stuff
-  (i didn't need it at all, because it's hard to make it work in a chroot)
-- `/dev/null` -- i don't know exactly why it's needed, but it won't work without
-  it
-
-because cgit will run in chroot-jail, all those files and directories except
-`/var/www/htdocs/cgit` should be located in `/var/www`
-(e.g. `/var/www/etc/cgitrc` and so on).
-
-```
-$ doas mkdir -p /var/www/{cache/cgit,dev,etc,usr/lib,usr/libexec}
-$ doas chown -R www:www /var/www/{cache/cgit,htdocs/cgit}
-```
-
-`/dev/null` is not a regular file, it's a device, so it must be created using:
-
-```
-$ doas install -d -g daemon /template/dev
-$ cd /template/dev
-$ doas mknod -m 666 null c 2 2
-$ doas mount_mfs -s 1M -P /template/dev /dev/sd0b /var/www/dev
-```
-
-this instruction is taken from [fossil docs][4].
-
-## copying libraries
-
-since cgit is not linked statically, it also needs some dynamic libraries.
-they all need to be accessible from chroot, so we need to copy them to
-`/var/www/usr/lib`. to check, what should be copied, run:
-
-```
-$ ldd /var/www/htdocs/cgit/cgit.cgi
-
-/var/www/htdocs/cgit/cgit.cgi:
-	Start            End              Type  Open Ref GrpRef Name
-	00000b068a590000 00000b068a7b6000 exe   2    0   0      /var/www/htdocs/cgit/cgit.cgi
-	00000b0927dcb000 00000b0927de7000 rlib  0    1   0      /usr/lib/libz.so.5.0
-	00000b0937409000 00000b093750b000 rlib  0    2   0      /usr/local/lib/libiconv.so.7.0
-	00000b0978c28000 00000b0978c37000 rlib  0    1   0      /usr/local/lib/libintl.so.7.0
-	00000b091fdc0000 00000b091fdcc000 rlib  0    2   0      /usr/lib/libpthread.so.26.1
-	00000b0920331000 00000b09203be000 rlib  0    1   0      /usr/local/lib/libluajit-5.1.so.1.0
-	00000b091cc5f000 00000b091cd54000 rlib  0    1   0      /usr/lib/libc.so.96.0
-	00000b089fffb000 00000b08a002b000 rlib  0    1   0      /usr/lib/libm.so.10.1
-	00000b08b2542000 00000b08b2585000 rlib  0    1   0      /usr/lib/libc++abi.so.3.0
-	00000b08cebc7000 00000b08cebc7000 ld.so 0    1   0      /usr/libexec/ld.so
-```
-
-and it'll return a list of all dependencies. copy them to `/var/www/lib`:
-
-```
-$ doas cp /usr/lib/{libz.so.5.0,libpthread.so.26.1,libc.so.96.0,libm.so.10.1,libc++abi.so.3.0} /var/www/lib
-$ doas cp /usl/local/lib/{libiconv.so.7.0,libintl.so.7.0,libluajit-5.1.so.1.0} /var/www/lib
-$ doas cp /usr/libexec/ld.so /var/www/usr/libexec
-```
-
-you should be able now to test cgit using this command:
-
-```
-$ doas chroot -u www /var/www /htdocs/cgit/cgit.cgi
-```
-
-it should return no errors but a webpage.
-
-## configuring cgit
-
-as already mentioned, cgit is configured using `/var/www/etc/cgitrc`. i suggest
-reading [manpage][3] for detailed overview of all available options, but here's
-an example cgitrc to start with:
-
-```
-#cache
-cache-size=1000
-cache-dynamic-ttl=60
-cache-static-ttl=44640
-cache-root-ttl=6
-cache-repo=5
-
-#index page
-enable-index-links=1
-enable-index-owner=0
-max-repodesc-length=60
-root-title=aaoth's git repos
-root-desc=some personal projects
-
-#repo global
-enable-git-config=1
-enable-commit-graph=1
-enable-follow-links=1
-enable-blame=1
-enable-http-clone=1
-enable-log-filecount=1
-enable-log-linecount=1
-enable-html-serving=1
-branch-sort=age
-snapshots=tar.gz zip
-side-by-side-diffs=0
-max-stats=week
-
-#root
-readme=:README.md
-readme=:readme.md
-readme=:README
-readme=:readme
-
-#mimetypes
-mimetype.html=text/html
-mimetype.gif=image/gif
-mimetype.jpg=image/jpeg
-mimetype.jpeg=image/jpeg
-mimetype.png=image/png
-mimetype.svg=image/svg+xml
-mimetype.pdf=application/pdf
-
-scan-path=/git
-```
-
-some of the settings are omitted, but you can tweak it further as you wish.
-
-note that i use autoscan feature of cgit. i have all my repos located in
-`var/www/git` as described by `scan-path` option.
-all of them are chowned by www user and have `cgitrc` text file inside.
-
-each repo-specific `cgitrc` looks like this:
-
-```
-name=test_repo
-desc=test repository to test cgit
-owner=username
-max-stats=month
-```
-
-## configuring httpd and slowcgi
-
-and now the last part is to actually serve cgit using httpd and slowcgi
-
-first of all, enable and start slowcgi:
-
-```
-$ doas rcctl enable slowcgi
-$ doas rcctl start slowcgi
-```
-
-then edit your `/etc/httpd.conf`, you need to create a simple server statement
-
-```
-server "example.com" {
-
-    listen on egress port 80
-    root "/htdocs/cgit"
-
-    location "/cgit.css" {
-        root "/htdocs/cgit"
-    }
-
-    location "/cgit.png" {
-        root "/htdocs/cgit"
-    }
-
-    location "/robots.txt" {
-        root "/htdocs/cgit"
-    }
-
-    location "/favicon.ico" {
-        root "/htdocs/cgit"
-    }
-
-    location "/*" {
-        fastcgi {
-            socket "/run/slowcgi.sock"
-            param SCRIPT_FILENAME "/htdocs/cgit/cgit.cgi"
-        }
-    }
-}
-```
-
-i know it can seem *very* odd, but it's the only way it works for me. as always,
-all improvement suggestions are welcome.
-
-and finally, (re-)start httpd:
-
-```
-$ doas rcctl enable httpd
-$ doas rcctl start httpd
-```
-
-[1]:https://git.zx2c4.com/cgit
-[2]:https://git.aaoth.xyz/cgit/cgit.git
-[3]:https://git.zx2c4.com/cgit/tree/cgitrc.5.txt
-[4]:https://www.fossil-scm.org/home/doc/trunk/www/server/openbsd/fastcgi.md#chroot
diff --git a/_posts/2021-05-23-join-the-test-of-my-matrix-server.md b/_posts/2021-05-23-join-the-test-of-my-matrix-server.md
deleted file mode 100644
index 929e509..0000000
--- a/_posts/2021-05-23-join-the-test-of-my-matrix-server.md
+++ /dev/null
@@ -1,58 +0,0 @@
----
-title: join the test of my matrix server
-date: 2021-05-23T19:57:34+03:00
-author: la-ninpre
-tags: openbsd testing matrix
----
-
-i launched my instance of [matrix][0] server recently. it runs on my openbsd vps
-and the server software i'm using is [synapse][1]. homeserver address is
-(unsurprizingly) `matrix.aaoth.xyz`.
-
-i also launched an instance of [element][2] matrix web-client on
-[element.aaoth.xyz][3], so you can try it.
-
-[0]:https://matrix.org
-[1]:https://matrix.org/docs/projects/server/synapse
-[2]:https://element.io
-[3]:https://element.aaoth.xyz
-
-<!--more-->
-
-matrix is relatively new standard for instant messaging. the main reason i am
-interested in it is that it's open-source. it means that anyone could launch
-their instance of synapse and be happy with it.
-it also means that the whole system is decentralized, providing protection
-against global surveillance.
-
-## about my server
-
-after you create an account, you will be connected to the broadcast room.
-it is unencrypted and read-only.
-there are some links to other rooms.
-
-note that everything going on with my server should be considered temporary
-and i could be able to stop, disable or wipe everything completely,
-so don't rely on this as production-ready tool.
-
-also note that pretty much everything is in russian, because i launched matrix
-primarily for my friends.
-
-## about openbsd
-
-here is a tutorial by the great man **robert d herb** who addressed a lot of
-quirks installing synapse on openbsd, which helped me a lot:
-[running a matrix homeserver with synapse and element][4]
-
-because i haven't figured out completely how openbsd's relayd is working,
-i broke my [fossils][5]. i hope i'll fix them later, but now they look messy.
-there are some issues, i think, with internal structure of fossil's ui.
-it needs to be served directly by httpd. but for synapse to work it is mandatory
-to run relayd as reverse proxy.
-
-if you know how to shift some portion of traffic to relayd and some to httpd,
-please, [drop me a line][6].
-
-[4]:https://robertdherb.com/things/matrix.html
-[5]:https://fsl.aaoth.xyz
-[6]:mailto:aaoth@aaoth.xyz
diff --git a/_posts/2021-06-02-i-launched-a-gemini-capsule-recently.md b/_posts/2021-06-02-i-launched-a-gemini-capsule-recently.md
deleted file mode 100644
index 89ff602..0000000
--- a/_posts/2021-06-02-i-launched-a-gemini-capsule-recently.md
+++ /dev/null
@@ -1,98 +0,0 @@
----
-title: i launched a gemini capsule recently
-date: 2021-06-02T22:36:43+03:00
-author: la-ninpre
-tags: gemini openbsd
----
-
-i launched my own gemini capsule, yay!
-
-it is available on the same domain, just type `gemini://aaoth.xyz` in your
-favourite gemini client.
-
-<!--more-->
-
-for those of you that don't know, gemini is a relatively new internet protocol.
-it is already gained popularity among some enthusiasts out there.
-
-it is intended to be simple and lightweight, it's just plaintext. and the whole
-specification is so simple that usable server and/or client for it could be
-written within about 100 lines of code.
-
-learn more on:
-
-[gemini website](https://gemini.circumlunar.space)
-
-## about quirks and vger
-
-firstly i looked through the list of gemini servers on gemini website.
-and one particular server there attracted me. it was `vger`.
-it is saying that it is secure and openbsd-centric.
-
-so i tried installing it. it is even packaged for openbsd, which was pretty
-convenient, even though i don't mind building stuff from source.
-especially if it is not a big bloated thing.
-
-and surprisingly the configuration was **so** simple, that i even hadn't
-realized it for a first couple of minutes.
-but then i wanted to launch some fancy cgi things, such as, for example,
-gemlikes. it is providing simple like and comment system for a blog.
-
-and there vger failed me. maybe it is me failed myself, but i tried all
-possible configuration options. i think, for now vger is not capable of
-running cgi scripts for some parts of the capsule.
-
-in other words, i'm talking about this.
-consider some capsule with tld `gemini://example.com`.
-it serves some static pages on `gemini://example.com/blog/*` and
-`gemini://example.com/about.gmi`.
-gemlikes need three binaries and a `gemlikes.toml` config file.
-they suggest placing them in `/cgi-bin/gemlikes/`.
-i placed them there, but i couldn't make it so it is how it needs to be.
-my vger is serving only cgi or only static pages.
-
-## another try
-
-so i looked though a list again, and found `gmid` there. it seemed like
-a good option too, because it's written in c and openbsd-aware too.
-
-and this was nice expirience, because gmid's config file is very
-similar to other openbsd's tool configs, such as relayd or httpd.
-the only peculiarity with gmid is that it's not yet packaged for
-openbsd, so i had to compile it manually. and also i created the daemon
-script for it in rc.d(8).
-
-here it is, if you need it:
-
-/etc/rc.d/gmid
-```sh
-#!/bin/ksh
-
-daemon="/usr/local/bin/gmid"
-daemon_flags="-c /etc/gmid.conf"
-
-. /etc/rc.d/rc.subr
-
-rc_pre() {
-    ${rcexec} "${daemon} -n ${daemon_flags}"
-}
-
-rc_cmd $1
-```
-
-and of course, my config is in `/etc/gmid.conf`. manpage of gmid contains
-very good descriptions of all the options available.
-
-## thoughts about geminispace
-
-my first impressions of gemini were a little odd. it's a little bit hard
-to read just text, when you are used to graphically overwhelming
-flashing websites with pictures and interactive stuff.
-but after a little bit of time comes the appretiation of the beauty
-of the pure text. 
-
-i can see gemini as a perfect place to host some informative resources,
-personal blogs and also for creative writing.
-and especially the latter, because it is so easy to spin up your own
-instance, i can see at as a great option for writers out there to
-host their content.
diff --git a/_posts/2021-08-18-matrix.aaoth.xyz-is-down.md b/_posts/2021-08-18-matrix.aaoth.xyz-is-down.md
deleted file mode 100644
index d27abd8..0000000
--- a/_posts/2021-08-18-matrix.aaoth.xyz-is-down.md
+++ /dev/null
@@ -1,32 +0,0 @@
----
-title: matrix.aaoth.xyz is down
-date: 2021-08-18T17:41:32+03:00
-author: la-ninpre
-tags: matrix openbsd testing
----
-
-i decided to shut down the matrix server that i launched recently
-(see [previous post][1]).
-here's my thoughts about this brief matrix experience.
-
-<!--more-->
-
-my main intention behind matrix server was to use it with my friends, but
-almost nobody responded, so it's over.
-synapse server is a bit resource heavy, so without it performance will be
-better.
-
-overall, matrix seemed like a good alternative to proprietary messengers to me.
-it's ability to maintain group end-to-end encrypted rooms is very good and
-also it was not so hard to set up.
-
-the biggest con for me was that matrix is spamming all the chats with a huge
-pile of system messages and notices. for example, all status changes of a
-chat room are displayed as separate messages and i find this very annoying.
-also, you can delete messages, but message deletion event is left forever in
-the chat history.
-
-maybe in future i will launch my server again
-(i haven't purged all my configs).
-
-[1]:https://aaoth.xyz/2021/05/23/join-the-test-of-my-matrix-server.html
diff --git a/_posts/2021-11-03-dualboot-linux-and-openbsd-with-grub.md b/_posts/2021-11-03-dualboot-linux-and-openbsd-with-grub.md
deleted file mode 100644
index bc24e2f..0000000
--- a/_posts/2021-11-03-dualboot-linux-and-openbsd-with-grub.md
+++ /dev/null
@@ -1,70 +0,0 @@
----
-title: dualboot linux and openbsd with grub
-date: 2021-11-03T15:23:00+03:00
-author: la-ninpre
-tags: openbsd linux grub tutorial
----
-
-i've been trying to dualboot openbsd with linux using grub on both bios and
-uefi machines and here's a solution that i've come up with.
-
-<!--more-->
-
-there are some guides about this on the internet, but there's no single guide
-that covers both bios and uefi. @rootbsd has a video where he shows how to
-do this, but his solution has one little disadvantage. he's specifying drives
-in a grub config using relative drive and partition numbers, such as
-`(hd0,gpt2)`. since these numbers could be different if one inserts a new drive
-to the computer, or changes drive order, the boot option could fail
-(which happened).
-
-all partitions and drives have their unique identifier -- uuid. there's no
-direct way to specify uuid in grub configuration, but there is a workaround.
-
-grub manual describes the `search` command which has an option to set root
-device if it is found. so we can use it for our purposes.
-
-## steps for dualbooting in bios/legacy mode
-
-1. install linux system on one of your drives
-
-2. reboot and boot from openbsd install media and install openbsd to other drive
-or partition.
-
-3. reboot and login to your linux system
-
-4. open a terminal and run `blkid` or `lsblk -f` to get an output partition
-uuids.
-
-5. write the following at the bottom of `/etc/grub.d/40_custom`:
-
-  ```
-  menuentry 'OpenBSD' {
-      search -su --no-floppy *UUID*
-      chainloader +1
-  }
-  ```
-
-  where *UUID* is the uuid of your openbsd partition (with type 'ufs2')
-
-6. run either `update-grub` or `grub-mkconfig` depending on what distribution
-you are using. consult your distro's wiki to find a way to update your grub
-configuration with recent changes.
-
-7. now reboot and you should see an openbsd's boot option in grub menu.
-
-## steps for uefi system
-
-for boot in uefi mode there are few differences. after installing openbsd
-don't reboot, but choose **shell**. now cd into `/mnt` directory and
-download `BOOTX64.EFI` from your desired openbsd mirror. for example:
-  ```
-  # cd /mnt
-  # ftp https://cdn.openbsd.org/pub/OpenBSD/7.0/amd64/BOOTX64.EFI
-  # reboot
-  ```
-
-after that the only other difference is that `chainloader` directive should
-be `chainloader /BOOTX64.EFI`.
-
-all other steps are the same.