playing with ansible

Makefile

@@ -38,6 +38,7 @@ $(databases): .dbs/%.rec: %/ | .dbs/
 
 # This two-variable read can only happen because of the quotes in the titles.
 db.rec: $(databases)
+	$(warning rebuilding from $? )
 	printf '%s\n' '%rec: guide' > $@
 	printf '%s\n' '%key: title' >> $@
 	printf '%s\n' '%type: requires rec guide' >> $@

README.md

@@ -44,6 +44,8 @@ They should not assume the reader knows much beyond common terminal commands, an
 People should be able to read an article from the beginning, then keep going until the end, and then stop.
 Articles should not take a detour through a chain of other articles of unknown size.
 
+[Do not Jaquays documentation](https://splint.rs/posts/no_links)
+
 ## Be Opinionated
 
 - Guides should not ask the reader to select options half-way through.

data/git/basics.md

@@ -6,36 +6,40 @@ tags: [ "data" ]
 
 ## New Machines
 
-```bash
+```sh
 git config --global user.email "$YOUR_EMAIL"
 ```
 
-```bash
+```sh
 git config --global user.name "$YOUR_NAME"
 ```
 
 # New Git
 
-Start a git in directory `$DIR`:
+Decide on algorithm:
 
-```bash
+- If you're scared of insecure hash-sums, go with `hash=sha256`.
-mkdir $DIR && cd $DIR
+- If you don't know what a hash sum is, go with `hash=sha1`.
-```
 
-```bash
+## Init the Git
-git init
+
+Start a git in directory `${DIR}`:
+
+```sh
+git init --object-format=${hash} ${DIR}
+cd ${DIR}
 ```
 
 Make a file explaining what the project does, and tell `git` to track it:
 
-```bash
+```sh
-echo "I hereby solemnly swear never to commit a binary." > README.md
+echo "I hereby solemnly swear never to commit a binary file." > README.md
 git add README.md
 ```
 
 Then make the initial commit, explaining the change you just made:
 
-```bash
+```sh
 git commit
 ```
 
@@ -43,17 +47,17 @@ git commit
 
 Once you make a change to some file, add it and make a commit explaining it.
 
-```bash
+```sh
 git add $FILE
 ```
 
-```bash
+```sh
 git commit -m"change $FILE"
 ```
 
 Check your history:
 
-```bash
+```sh
 git log
 ```
 
@@ -64,20 +68,20 @@ Give it the same name as the `$DIR` directory, above.
 
 Add this as a remote:
 
-```bash
+```sh
 REMOTE=gitlab
 git remote add $REMOTE  https://gitlab.com/$USERNAME/$DIR
 ```
 
 Tell git you're pushing the branch "master" to the remote repo "origin": 
 
-```bash
+```sh
 git push -u master origin
 ```
 
 If someone makes a change on the remote, pull it down with:
 
-```bash
+```sh
 git pull
 ```
 
@@ -86,31 +90,31 @@ git pull
 A branch is a full copy of the project to test additional ideas.
 You can make a new branch called 'featurez' like this:
 
-```bash
+```sh
 git branch $FEATURE_BRANCH
 ```
 
 Have a look at all your branches:
 
-```bash
+```sh
 git branch
 ```
 
 Switch to your new branch:
 
-```bash
+```sh
 git checkout $FEATURE_BRANCH
 ```
 
 And if your changes are rubbish, checkout the "master" branch again, then delete "featurez":
 
-```bash
+```sh
 git branch -D $FEATURE_BRANCH
 ```
 
 Or if it's a good branch, push it to the remote:
 
-```bash
+```sh
 remote=origin
 git push $remote $FEATURE_BRANCH
 ```
@@ -119,13 +123,13 @@ git push $remote $FEATURE_BRANCH
 
 Once you like the feature, merge it into the main branch.  Switch to master then merge it:
 
-```bash
+```sh
 git merge $FEATURE_BRANCH
 ```
 
 And delete the branch, as you've already merged it:
 
-```bash
+```sh
 git branch -d $FEATURE_BRANCH
 ```
 
@@ -133,7 +137,7 @@ git branch -d $FEATURE_BRANCH
 
 ## Pulling another git repo into a subtree
 
-```bash
+```sh
 git subtree add -P config git@gitlab.com:bindrpg/config.git master
 ```
 
@@ -141,27 +145,27 @@ git subtree add -P config git@gitlab.com:bindrpg/config.git master
 
 ## Delete All History
 
-```bash
+```sh
 git checkout --orphan temp
 ```
 
-```bash
+```sh
 git add -A
 ```
 
-```bash
+```sh
 git commit -am "release the commits!"
 ```
 
-```bash
+```sh
 git branch -D master
 ```
 
-```bash
+```sh
 git branch -m master
 ```
 
-```bash
+```sh
 git push -f origin master
 ```
 
@@ -169,21 +173,21 @@ Gitlab requires more changes, such as going to `settings > repository` and switc
 
 ## Clean up Bloated Repo
 
-```bash
+```sh
 git fsck --full
 ```
 
-```bash
+```sh
 git gc --prune=now --aggressive
 ```
 
-```bash
+```sh
 git repack
 ```
 
 ## Find Binary Blobs
 
-```bash
+```sh
 git rev-list --objects --all \
 | git cat-file --batch-check='%(objecttype) %(objectname) %(objectsize) %(rest)' \
 | sed -n 's/^blob //p' \

data/git/git-lfs.md

@@ -1,6 +1,7 @@
 ---
 title: "git-lfs"
 tags: [ "data", "git" ]
+requires: [ "git" ]
 ---
 
 Git Large File Storage ('LFS') needs to change your `~/.gitconfig` to check out those binary files:

data/git/git_stash.md

@@ -0,0 +1,69 @@
+---
+title: "git stash"
+tags: [ "data", "git" ]
+requires: [ "git" ]
+---
+
+Save file-changes without committing anything.
+
+Change a file:
+
+```sh
+file=README.md
+fortune >> ${file}
+git diff
+git stash save
+```
+
+List which stashes you have:
+
+
+```sh
+git stash list
+
+    stash@{1}: WIP on master: c21f102 init git
+```
+
+Make a new file, then stash it:
+
+
+```sh
+otherfile=file.log
+fortune > ${otherfile}
+git add ${otherfile}
+stashname=logfile
+git stash save ${stashname}
+```
+
+Now you can see two stashed changes, and the most recent has a name:
+
+```sh
+git stash list
+
+    stash@{0}: On master: logfile
+    stash@{1}: WIP on master: c21f102 init git
+```
+
+You can delete a stash by referring to its index number, or name (if it has one).
+
+```sh
+choice=1
+git stash drop ${choice}
+
+choice=${stashname}
+git stash drop ${choice}
+```
+
+Or just run `git stash drop` to remove the most recent (labelled `{0}`).
+
+Return stashed changes with an index number (or the most recent).
+
+```sh
+git stash pop ${choice}
+```
+
+Delete all stashes:
+
+```sh
+git stash clear
+```

data/json.md

@@ -0,0 +1,13 @@
+---
+title: "ijq"
+tags: [ "data", "json", "TUI" ]
+---
+
+Analyse `json` easier with `ijq`.
+
+```sh
+column -J -ts: -H PW,GID,shell -N User,PW,UID,GID,Description,Home,shell /etc/passwd > host.json
+ijq !$
+```
+
+If you get stuck, try adding `.[]`.

data/task/contexts.md

@@ -0,0 +1,34 @@
+---
+title: "Taskwarrior Contexts"
+tags: [ "data", "task" ]
+requires: [ "Taskwarrior" ]
+---
+
+# Contexts
+
+Set three contexts by their tags:
+
+```sh
+task context define work +sa or +hr
+```
+
+```sh
+task context define study +ed or +void or +rat
+```
+
+```sh
+task context define home -sa -hr -ed -void -rat
+```
+
+Change to the first context.
+
+```sh
+task context work
+```
+
+Then stop.
+
+```sh
+task context none
+```
+

data/task/task.md

@@ -1,6 +1,6 @@
 ---
-title: "task"
+title: "Taskwarrior"
-tags: [ "organization" ]
+tags: [ "data", "organization" ]
 ---
 
 Set up the configuration file:
@@ -9,6 +9,18 @@ Set up the configuration file:
 task
 ```
 
+Taskwarrior published a new feature to synchronize tasks others, but the feature was not ready.
+The server's default installation instructions assume that users pay for hosting services.
+All listed providers run proprietary software and actively support genocide.
+
+
+To ignore the synchronization, tell the configuration file to use a local synchronization file.
+
+```
+task config sync.local.server_dir
+task config data.location ~/.local/state/
+```
+
 Add a task:
 
 ```sh
@@ -44,18 +56,16 @@ task add project:house buy potted plant
 task add proj:house.repair buy screwdriver
 task add proj:house.repair buy shelf brackets
 task add pro:house.paint buy white paint
-task add pro:house.paint buy red paint
+
-task add pro:house.paint buy black paint
+for t in "buy red paint" "buy black paint" "buy brushes" ; do
-task add pro:house.paint buy brushes
+  task add pro:house.paint $t
+done
 ```
 
 ## Summary
 
 ```sh
 task pro:house sum
-```
-
-```sh
 task burndown.daily pro:house
 ```
 
@@ -63,44 +73,12 @@ The summaries will show how fast a project is being completed, and when you can
 
 # Tags
 
+
 ```sh
 task add +buy toothbrush
-```
-
-You can then see only tasks which involve buying something with:
-
-```sh
 task +buy
 ```
 
-# Contexts
-
-Set three contexts by their tags:
-
-```sh
-task context define work +sa or +hr
-```
-
-```sh
-task context define study +ed or +void or +rat
-```
-
-```sh
-task context define home -sa -hr -ed -void -rat
-```
-
-Change to the first context.
-
-```sh
-task context work
-```
-
-Then stop.
-
-```sh
-task context none
-```
-
 # Review
 
 View list of tasks completed in the last week:
@@ -111,21 +89,17 @@ task end.after:today-1wk completed
 
 # User Defined Attributes
 
-Make a UDA 'size'.
+Define a new attribute for tasks called 'size'.
+The 'user defined attribute' (UDA) needs a `type` and `label`.
 
 ```sh
 task config uda.size.type string
-```
-
-```sh
 task config uda.size.label Size
 ```
+You can also ensure task tasks can only be `large`, `medium`, or `small`, then set a default.
 
 ```sh
 task config uda.size.values large,medium,small
-```
-
-```sh
 uda.size.default=medium
 ```
 

data/task/taskwarrior_configuration.md

@@ -0,0 +1,20 @@
+---
+title: "Taskwarrior Configuration"
+tags: [ "data", "task" ]
+requires: [ "Taskwarrior" ]
+---
+
+Show your current config:
+
+
+```sh
+task show
+```
+
+Use machine-readable output to make a config file with all configuration keys shown, then make it your configuration file.
+
+```sh
+task _show > ${file}
+mv ${file} ~/.config/task/taskrc
+```
+

networking/port_scan.md

@@ -0,0 +1,25 @@
+---
+title: "Port Scan"
+tags: [ "networking" ]
+repo: 'https://github.com/mrjackwills/havn/'
+---
+
+`havn` scans ports.
+It's not in many repos, but if you can `cargo install havn`.
+
+
+```sh
+havn -h
+havn -a
+domain=splint.rs
+havn -p 19-90 ${domain}
+```
+
+That last command doesn't work reliably, so increase the number of retries (`-r`), or decrease concurrent requests (`-c`).
+
+
+```sh
+havn -p 19-443 -r 6 ${domain}
+havn -p 1-1000 -c 500 -r 5 ${domain}
+```
+

system/ansible/ansible_with_docker.md

@@ -0,0 +1,114 @@
+---
+title: "Ansible with Docker"
+tags: [ "system", "ansible", "docker" ]
+requires: [ "Docker" ]
+---
+
+Set up two containers: `deb` and `arch`, add them to an `ansible` hosts file, then do a 'ping' to see if they respond.
+
+## Required Packages
+
+- `ansible`
+- `jq`
+- `docker`
+
+## Debian Container
+
+```sh
+docker run -di --rm --name deb --hostname deb debian
+docker exec -it deb sh -c 'apt update && apt -y install openssh-server python3 sudo'
+```
+
+Generate the host's ssh keys, then start the ssh daemon:
+
+```sh
+docker exec -it deb sh -c 'ssh-keygen -A'
+docker exec -d deb /usr/sbin/sshd -D
+```
+
+## Arch Linux Container
+
+```sh
+docker run -di --rm --name arch --hostname arch archlinux
+docker exec -it arch sh -c 'pacman -Syu --noconfirm python sudo openssh'
+docker exec -it arch sh -c 'ssh-keygen -A'
+docker exec -d arch /usr/sbin/sshd -D
+```
+
+## `ssh` Keys
+
+Copy across your public ssh key to the container's `authorized_keys` file:
+
+```sh
+pubkey=~/.ssh/id_rsa.pub
+for hostname in arch deb; do
+    docker cp $pubkey $hostname:/root/.ssh/authorized_keys
+    docker exec -it $hostname sh -c "chown -R root:root /root/.ssh/"
+    docker exec -it $hostname sh -c "chmod -R 700 /root/.ssh/"
+done
+```
+
+## Hosts File
+
+Find name of containers' IPv4 addresses.
+
+```sh
+docker network inspect bridge
+```
+
+The output is awful.
+Use `jq` to parse the `json`:
+
+```sh
+docker network inspect bridge  | jq -r '.[].Containers | .[].IPv4Address'
+```
+
+Now put those into a host file:
+
+```sh
+docker_hosts=hosts.txt
+echo '[containers]' > $docker_hosts
+
+docker network inspect bridge  | \
+    jq -r '.[].Containers | .[] | "root@" + .IPv4Address' | \
+    cut -d/ -f1 >> $docker_hosts
+```
+
+You may need to add those host keys to your known hosts file.
+Either connect interactively, or (for scripts):
+
+```sh
+hosts="$(docker network inspect bridge  | jq -r '.[].Containers | .[] | .Name + " " + .IPv4Address' | \
+    cut -d/ -f1)"
+
+echo "$hosts"
+
+echo "$hosts" | while read hostname ip; do
+    printf "%s" "$ip"
+    key="$(docker exec $hostname cat /etc/ssh/ssh_host_ed25519_key.pub)"
+    echo "$ip $key" >> ~/.ssh/known_hosts
+done
+```
+
+Check if they ping:
+
+```sh
+ansible -i $docker_hosts all -m ping
+```
+
+This command produces an irritating warning about the python interpreter (i.e., `python3`).
+
+Make the warning shut-up:
+
+```sh
+echo '
+[containers:vars]
+ansible_python_interpreter=/usr/bin/python3.13' >> $docker_hosts
+```
+
+Now the ping is cleaner:
+
+```sh
+ansible -i $docker_hosts all -m ping
+```
+

system/makefiles/patterns.md

@@ -1,4 +1,3 @@
-
 ---
 title: "Makefile Patterns"
 tags: [ "system", "make" ]

system/makefiles/python_projects.md

@@ -0,0 +1,98 @@
+---
+title: "Python Projects with Makefiles"
+tags: [ "tutorial", "system", "makefiles", "graphviz", "python" ]
+requires: [ "Makefiles" ]
+---
+
+If you have a python script which requires a packages - e.g. `graphviz` - you can automate the setup with a `Makefile`.
+The `Makefile` will:
+
+1. Make three copies of an 8-line installer script.
+2. Use one of the installer script to install local packages.
+3. Install a symbolic link to python.
+3. Make a script called `activate`, which tells python to use 
+
+# Context
+
+Python coders don't like updating their projects, they just expect everyone to install the same version of everything that they have.
+Historically, people dealt with this by installing only half a dozen copies of `graphviz`; but now each python project uses a local environment, with a local copy of `graphviz`, which means everyone gets to install a new copy of `graphviz` every time they try out a project.
+
+Downloading 40MB of software for each 40-line script you write is called 'virtual environments' because it sounds cool.
+We can make it even cooler with `make`, but not yet, because python - like the fae of old - will not fetch anything until you know its true name.
+
+# Setup
+
+```sh
+command -v python
+realpath `!!`
+```
+
+You must reveal that true path, because `python` is always a relative symbolic link, to an absolute symbolic link, which leads to a shortcut.
+We can finally let `make` know how to invoke python, and where it will install `graphviz`.
+
+If your python's version is '3.14', then python needs its packages placed in `${somewhere}/lib/python3.14/site-packages/`.
+You must create a new, local, name for these packages, because - like the fey of old - python demands a private name in return for revealing its true name.
+
+I'll call mine `camelot`, because the path is long and arduous.
+Set up the Makefiles like this:
+
+```make
+py_link != command -v python
+py != realpath $(py_link)
+version != basename $(py)
+
+virtenv = camelot
+```
+
+Now you can ask for a local `pip` script, which can install the python packages:
+
+```make
+    [...]
+
+$(virtenv)/bin/pip:
+    $(py) -m venv $(virtenv)
+```
+
+Finally, list the packages you want in `requirements.txt`, and make `pip` install from it.
+
+
+```sh
+ppkg=graphviz
+echo ${ppkg} > requirements.txt
+```
+
+```make
+    [...]
+
+pkgs = $(virtenv)/lib/$(version)/site-packages/
+
+$(pkgs): $(virtenv)/bin/pip
+$(pkgs): requirements.txt
+    $(virtenv)/bin/pip install -r $<
+```
+
+The complete Makefile looks like this:
+
+
+```make
+all: .default
+
+py_link != command -v python
+py != realpath $(py_link)
+version != basename $(py)
+
+virtenv = camelot
+
+$(virtenv)/bin/pip:
+    $(py) -m venv $(virtenv)
+
+pkgs = $(virtenv)/lib/$(version)/site-packages/
+
+$(pkgs): $(virtenv)/bin/pip
+$(pkgs): requirements.txt
+    $(virtenv)/bin/pip install -r $<
+
+.PHONY: .default
+.default: $(pkgs)
+```
+

system/managing_groups.md

@@ -17,10 +17,10 @@ Remove yourself from all groups, and add yourself back to only `wheel`, `audio`,
 sudo usermod --groups wheel,audio,$USER
 ```
 
-Add yourself to the `wheel` group:
+Add yourself to the `docker` group:
 
 ```sh
-su root -c "usermod --append --groups wheel $USER"
+su root -c "usermod --append --groups docker $USER"
 ```
 Add yourself to the `network` group:
 
@@ -28,5 +28,11 @@ Add yourself to the `network` group:
 sudo usermod -aG network $USER
 ```
 
-The changes will not take effect until you log in again, so reboot or log into `localhost` with [ssh](../networking/ssh.md).
+The changes have not taken effect, so log into your own account again with `su`:
 
+
+```sh
+groups
+sudo su $USER
+groups
+```

virtualization/docker.md

@@ -1,6 +1,7 @@
 ---
 title: "Docker"
 tags: [ "documentation", "virtualization" ]
+requires: [ "Managing Groups" ]
 ---
 ```sh
 sudo pacman -S docker

vision/ffmpeg.md

@@ -10,7 +10,7 @@ ffmpeg -i [input file] output_file.mkv
 
 The input file might be a device, such as a camera.
 
-#Record screen
+# Record screen
 
 Take the format as 'grab the x11 screen'.
 
@@ -30,7 +30,8 @@ or maybe just...
 ffmpeg -f x11grab -s "$(xdpyinfo | grep dimensions | awk '{print $2}')" -i :1.0 out.mkv
 ```
 
-#Add default pulse audio
+# Add default pulse audio
+
 ```bash
 ffmpeg -f x11grab -s [screensize] -i :0.0 -f alsa -i default out.mkv
 ```

vision/make_a_gif.md

@@ -0,0 +1,22 @@
+---
+title: "Make a gif"
+tags: [ "vision" ]
+---
+
+Split your video into frames.
+
+```sh
+d=$(mktemp -d)
+cp $vid.mp4 $d
+cd $d
+
+ffmpeg -i $vid.mp4 -vf "fps=10, scale=360:-1" frame%04d.png
+gifski --fps 10 -o ${out}.gif frame*.png
+```
+
+Optimize:
+
+```sh
+gifsicle --optimize=3 --lossy=100 -o ${final}.gif ${out}.gif
+```
+

writing/tex/calendar.md

@@ -0,0 +1,12 @@
+---
+title: "Calendar"
+tags: [ "writing", "tex", "fun" ]
+requires: [ "LaTeX Packages" ]
+---
+
+```sh
+yplan $(date +%Y) > ${file}.tex
+pdflatex -output-directory=/tmp/ ${file}.tex
+mv /tmp/${file}.pdf .
+```
+

writing/tex/tex_packages.md

@@ -1,5 +1,5 @@
 ---
-title: "LaTeX Setup the Hard Way"
+title: "LaTeX Packages"
 tags: [ "writing" ]
 ---
 
@@ -21,7 +21,7 @@ ls /opt/texlive/${YEAR}/texmf-dist/scripts/texlive/tlmgr.pl
 ```
 
 Double-check the year.
-It should *not* match the real year, it should match the `texlive`.
+It should *not* match the current year, it should match the `texlive`.
 
 ## Problems along the Path
 

writing/vim.md

@@ -3,13 +3,14 @@ title: "vim basics"
 tags: [ "vim", "basic" ]
 ---
 
-1. Insert text by pressing `i`.
+1. Insert text by pressing `a`.
 1. Stop inserting text by pressing `Ctrl+[`.
 1. Exit with `ZZ`.
 1. Congratulations, you now know `vim`.
 
 ## Extras
 
+- [Learning Vim](vim/vi.md)
 - [Navigation](vim/navigate.md)
 - [Completion](vim/completion.md)
 - [Search](vim/search.md)

writing/vim/vi.md

@@ -0,0 +1,46 @@
+---
+title: "How to Learn `vim`"
+tags: [ "vim", "learning" ]
+requires: [ "vim basics" ]
+---
+
+1. Uninstall `vim`.
+1. Install `vi`.
+1. Write a few blog posts.
+
+The [ancient wisdom](https://stackoverflow.com/questions/1218390/what-is-your-most-productive-shortcut-with-vim/1220118#1220118) is correct - everyone should start with `vi`.
+But the standard wisdom fails to mention writing.
+So people just 'try to remember', then they edit a couple of configuration files with `vim`, then stop, because it's a pain to think about keystrokes while thinking about how the configuration file works.
+Nobody can learn like that.
+
+Learn with a blog, or writing emails in `vim`, or writing anything, as long as it's *real writing*.
+Write a journal or some recipes.
+Write literally anything which is written in paragraphs.
+
+And while you're writing, remember one rule: if you want something, `vi` can do it, you just need to look up how.
+
+Blog first, then try out some of these commands:
+
+- `hjkl`
+- `x`
+- `:!grep very %`
+- `onew line<Esc>...`
+- `Onew line<Esc>...`
+- `~`
+- `~~~~`
+- `$r!`
+- `w %-2`
+- `x!`
+- `0RTipex`
+- `_4rX`
+- `bc2e`
+- `c2E`
+- `cw`
+- `cW`
+- `ci'`
+- `zz`
+- `:set number`
+- `:set nonumber`
+- `:set relativenumber`
+- `:set number relativenumber`
+