I’ve been using zsh with oh-my-zsh for a long time now. If I were to replace it, this is what I’d look for in the future:
cd-ing!$ and !! expansion before execution**If you use virtualenv to isolate your python project’s environment, and want your code tested automatically – read on!
virtualenv makes sure you lockdown your project’s main directory and all subdirectories of it. This ‘lockdown’ means that you never touches your global python binary, or any globally installed libraries (like “sudo pip install flask” ).
Once locked down, you install all packages again, even those you have globally installed. This enables you to have one version of flask globally installed, but another version in your project. All dependencies can be listed in a separate file and validate a precise environment for you to work with. Tightly controlled dependencies is key to a deployment without surprises.
Jenkins is a CI server which means it does a lot of repeating stuff so you can focus on doing more important stuff. More specifically, it listens for changes to your project’s version control system (like git).
When changes are detected, the project is built and the test suite is executed. If any step fails, the CI server tells you that it did.
Jenkins needs some massaging before it handles the hijacked environment of virtualenv. This is how I did it for my local git repository:
1 2 3 4 5 6 7 8 | |
and
1 2 | |
You need to add a git hook which triggers a Jenkins build:
1 2 | |
Add an erroneous test, this will do:
1 2 | |
You should see a new build being queued up in Jenkins within a minute. If that doesn’t work, execute the hook and watch the output for error messages:
1
| |
Now, Jenkins should try to test your project but fail, and report the failure through the GUI. Tada.
Because I always forget the exact commands:
1 2 | |
On OS X, you’ll find both youtube-dl and ffmpeg through homebrew.
Consider this a checklist for your API, should work for REST, CLI or source code!
Have a scope and tell me what it is in a couple of sentences.
Let me learn the general approach and edge cases via a concise and result oriented documentation.
Scream at me when I did something wrong and, if I was close enough (and you have a lot of time), point me in the right direction.
For example, when treating errors you shouldn’t alternate between simultaneous ways of telling your users something went wrong. Pick one and go with it.
How to properly read command line arguments for a bash script and do it well even without getopt:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 | |
I think the syntax is much cleaner than that of getopt’s as well – that is, if you’ve even got a bash version supporting getopt.
Screen could be used as nohup with the additional possibility of attaching to the terminal session again, once your ssh session was aborted (or equivalent).
screen | Create a new screen |
screen -S name | Create a new screen named “name” |
screen -ls | List screens |
screen -D | Detach from screen |
screen -r | Attach to the single existing screen |
screen -r 4262 | Attach to the detached screen with pid 4262 |
screen -x 4262 | Reattach to the attached screen with pid 4262 |
| ctrl+a c | Create a new terminal in the current screen |
| ctrl+a H | Log all output of current screen to a log file |
| ctrl+a space | Change terminals within current screen |
| ctrl+a :sessionname name | Rename current screen to “name” |
| ctrl+a d | Detach from current screen, like exit but keeps the terminal for reattaching later on |
I have all my code stored in ~/Sites and every sub folder represents a project. Each folder is also a git repository. This means there’s history attached to it, which could be much more easily navigated than through the finder or a terminal. Some of the projects are mirrored on github but it’s too messy to keep them all there, although – that kind of overview is nice to have.
My solution to getting an overview was to setup gitweb locally, which makes for a very fast solution. This is how you do it on your OS X-machine:
Find where git’s installed at. I use homebrew, which means git (by default) resides in /usr/local/Cellar/git.
Grab the gitweb cgi, copy it to where your local Apache server can find it:
1 2 | |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 | |
Now, visit & bookmark http://localhost/cgi-bin/gitweb.cgi in your browser and you should be good to go.
A common pattern while editing is to yank anything (i.e. yy), visually select what you want to remove (i.e. vj) and paste (p). This will in turn place the overwritten test in the yank register; but we wanted to paste that same line again!
Type :reg to see all currently populated registers, they’re all preceded by ". This means that you can paste by visually selecting and typing "0p if “0 is the register’s content you want to paste.
Collection of intermediate but handy git tricks. This post will be updated now and then.
git am -3 --signoff <name of patch>.patch
<object id> could be master, 35f1cab or v1.4, for example.
git bisect start
git bisect good <object id>
git bisect bad <object id>
# test the given revision
# say if it's bad or not with
git bisect bad
# continue this way and exit with
git bisect reset
git branch new_branch <ref>
git add -p
# edit
# commit
git format patch origin/master --stdout > my_changes.patch
Where origin/master is the base for references.
git checkout -p
# is hunk too big? 's' to split it
If you want to add some changes that you’d want included in the previous un-pushed commit
git commit -am "just a commit"
# make changes
git commit --amend -a
git rebase -i HEAD~2
# p ...
# e ... <- edit this
#edit files
# git add; git commit --amend; git rebase --continue
git push <remote> :<branch>
git log -p
# for a specific file
git log -p <path>
Going from Wordpress to mynt.
$ ./wp2mynt.php export-file.xml (wp2mynt.php is on github)$ tail -f /usr/local/nginx/logs/error.log I notice a bunch of stuff I missed. Redirections here we go$ ab -n 10000 -c 25 can be run without any problems what so everI’m pretty sure I suck at micromanaging apache & mysql. And that nginx is pretty fly with static files.
Also, a hiccup: I issued a 301 redirect in one case, and it went badly – my nginx non-ninja-config-skils made me issue a recursive redirect loop. 301 responses are heavily cached by browsers, I had to delete the cache after removing the direction out of the config file. Do not make this mistake during ‘live hours’ or with a heavily trafficked page.