someone later on -m "removed single quotes, added double quotes"
More commits == More money!!!
Real talk: I have worked at a company where an exec briefly had the bright idea to rank developers on number of commits made.
You can imagine how that turned out.
He wasnβt an exec for very long.
lmaoo, I can't believe someone actually did that and was serious about it.
I'm already imagining commits piling up with comments like:
"add a newline"
"move function up"
"rename variable users to user_list"
"Revert previous commit"
"no John, it's user_list and that's it, don't you dare change it again"
He wasnβt an exec for very long.
Good riddance I guess
Oh, no, we can do better.
Every single line of every single PR is an individual commit. And never squash-commit.
Talk about atomic commits
If you really want to nuke it from orbit: do it for every character.
It is absolutely technically feasible to go further than that, but I think thatβs the point of diminishing returns in terms of effort.
Hah, I knew it would come to that
but I think thatβs the point of diminishing returns in terms of effort
Not if you make another application to automatically split commits in that way... application which will need its own commits, it's never-ending
Why do u prefer single quotes?
Not Shall330, but double quotes can imply variables in the string in some languages.
Which languages are these?
PHP for one. "$x" will interpret the variable $x. '$x' does not.
Well there's ya problem ur using php.
Also bash
And Perl.
I have no preference! It was to do with Spring/yaml config and some really strange conflict which required the use of single quotes. I'm still a total noob in the world of software dev, so I wouldn't be able to explain why it worked π€£
"Update"
It was my dotfiles, ok?
You're going to regret not adding the second sentence to the commit message some day.
Bold of you to assume there is a cohesive theme to dotfile commits. Usually it's just been long enough that I don't want to lose my latest tweaks
Squash me later
My commit messages have gotten extremely lazy since I start squashing all my commits down to one. I just describe the PR on the first commit message and write nonsense in all the others.
I've been adding the commit shas + messages to the final description, if anyone wants to see my exploratory work.
I know that if you are on the local repository where the commits were originally created they'll remain accessible through recovery methods but AFAIK orphaned commits aren't synced to other machines.
That's correct. This is for work, which uses GitHub. The dangling commits remain accessible via their sha through the web ui, so I can link them in the PR description. I don't put them in the actual commit message.
I think these are garbage collected eventually, but no idea on cadence. It's long, anyway.
"Changed consent banner text to reflect from legal, even though they are wrong"
"circumvented GDPR despite calling out strong ethical objections, because legal & management are forcing me to"
Update coffee list and tex files.
...and yes we have our office coffee list in a git repo.
Do new hires have to submit a pull request to get in on it? If so, I'm guessing the non-technical staff don't get to drink a lot of coffee.
It's actually just a paper list where people make ticks for coffee. That is then manually synced to an hledger file in git.
For what? To keep track of who's drinking coffee? Are you charging for coffee?
I work in academia... There is no free coffee here.
Mfw I git push -f
Always force-with-lease
Do you have any plans to open source your coffee list? I want to add this feature where I get an iced coffee in the morning
"fix for real this time"
"28496 - there, it's fucking fixed you twat waffle."
Ticketed bug bosses son found. Dude nagged his dad who nagged us until it got fixed. Boss doesn't review code. And for the sake of a half dozen coworkers, I hope he never does.
"I'm sorry, guys."
";"
Previous commit was some stupid easy fix I didn't even bother compiling. Well, I should have, because it was the first time in recent memory I committed some code with a missing semicolon...
Before anyone asks : no, we don't do reviews Β―\_(γ)_/Β―
Join me and use ~ to mean "shit changed just check the diff"
For those I just commit with the message "ngl there's a lot of changes in here"
Tweaked some things
+2543; -5642;
"Moved code base from 2 spaces to 4."
Hrfg7ygfrvvby
git commit -m "if this doesn't fix it I'm looking up availabilities at my nearest maccas"
wip: good luck you monday fuck hahahaha
"My laptop died before I could push these"
Not a good commit name I know, but my laptop stopped booting and I just really needed to continue from my desktop.
It's a student project anyway.
Fix for vendor lies.
Fixed dead link (fr)
French dead link?
Converted more tests to use the correct handle.
A
Because who cares for commit messages to my private repo?
You when you mess something up and only figure it out later?
I do Ansible playbooks so not too had to see when I ballzed it up.
Whatever works for you, but I love writing a quality commit message. It's like my reward for doing all that work - wrapping up the changes in a nice little summary.
Sometimes I just scroll through my commits when I'm feeling blue...
hoho
I had just figured out how to fix a bug that broke our in house reminder system. My smug ass put hoho as the commit message for figuring something out that the previous guy couldn't. Of course it came back to bite me in the ass THIS WEEK when the system broke again in a similar fashion, but I couldn't remember what I had done.
Diff
I know I know. I figured out what I did before I thought of that... which was just now when you said it.
Revert this later
Yikes
Prevent subprocess from killing itself until finished.
WIP
You copycat.
init: add readme
βInitial commit, come back laterβ
Fix 3
Lol test
At a PV node, do not cut off from any TT hit.
The SPRT result of this may be different than before thanks to LMR.
Bench: 5212899
I wonder what the last commit message of my career will be.
I wonder what my last commit at each job was. I'll bet it was boring. About 10% of my commit messages are genuinely interesting.
For some definition of interesting β¦. The quashed commit whose message is a Jira ticket number and title?
This is what I hate about squashing. I love the fact that my crazy journey to a finished feature is there in the log for everyone to see, including myself lol. But squashing makes me feel empty when I write commit messages. Where's the fun... Just a straight line of perfectionism. Maybe someone has a crazy but cool idea that would be worth picking. π
Maybe. I do more DevOps these days, so tend to have many small changes that canβt even be tested without checking them in and running in CI. Iβd have hundreds of βfix unit testsβ commits alone
I've been thinking about that when talking to my DevOps colleagues, that there's gotta be a better way to test CI before committing. The whole change-commit-test dance would kill me if on a daily basis. So cumbersome.
We squash. I'm not really interesting in your local journey to land the change. It's sometimes useful during review, but after that it's mostly the state of the main branch I care about. It's what I need to bisect anyway.
I don't like commits that are just references to issues. Copy the issue into the commit message so git blame tells you something useful. Unless it's just closing a simple big. Then the title and issue reference are plenty.
Depends on the project I imagine.
"it's possible i buried a self delete timebomb in one of our repos years ago. enjoy looking for it just incase assholes"
haaaands
Fuck it I quit
Update
.
"Merge pull request #8 from [branch name]"
Not the most exciting but hey, someone has to do it.
Merge commits π
" "
Mod: Added delay to let switches settle
Something along the lines of "I don't know what this does, I wrote this code two weeks ago"
Formatting fixes
Merge remote-tracking branch 'origin/master'
Ew - rebase all the things!
Fixes
Refactoring for the EU region.
Reusing Terraform projects for the win.
Changed index.html
Added zip code to the reseller portal when batch-creating transactions in
I try to start off descriptive, in case I want to diff a few changes back. After that, sometimes I write notes or general reflections on life in there... Until yesterday I thought no one was reading them
This is pretty funny, but now I want to watermark all of my future work with outdated "Made With ..." Stickers.
"Made With Printshop Deluxe!"
"Made With Microsoft Works!"
"Sent From my Sidekick!"
"cleanup" lol
Change port to 8080 because stupid container locks down 80
Fixed tests.
"update doc to reflect reality still more"
chore: replaced println() with proper logging util
I forgot to replace them water finishing debugging.
m
usually the commits at the end of the day when I haven't finished a task yet. It will be squashed and disappear eventually.
Lint issue
Uhh it's on a foss project so imma not dox myself lmao
My commits are a little too specific to post here, lmao. I likely put too much info into them, Iβm still trying to get a good balance.
I think long messages are a good habit. Start with something readable in the history, past that who cares? Most people rarely read past the preview, and if they do they want details
I think it's great because it makes you reflect on what the goal was and what you did. I sometimes stop to make a quick change as I'm writing, or just collect my thoughts before mentally dismissing the task
Update
"Add ability to load sources which are not a nginx file listing."
I try to be descriptive while keeping it short.
βSome appropriate message.β
Quack quack
"Couldn't count all the stuff"
?
WIP
"Monte Carlo code changes. BLAME: notme"
fix this fucking gitignore
latest cargo crates updated
Minor text fixes
Stuff
Beta 0.9.9.9.9.9
Lots of updates (it's a private repo and I'm bad at doing smaller individual commits!)
some fixes
Fixinβ thangz
feature suggestion but "hyprcursor"
CPEUI-20102: It's working but still clean up to do.
removed /32s in favour of /24s. Small layout changes.
Fix
h
1.1
Options for donation and contact links
Nothing important
Rolling back demo commit
Add Exporters MVP
Comment
github-sync: local install (WIP, split and rebase)
changelog
somebody renamed the interface but didnβt rename the monitoring interface 2
first release
I wrote a brief manifesto on my hatred of Python because some refactoring ended up having a comma at the end of a line, which screwed me over for about an hour until I happened to noticed it.
Implement add headers
Oh i don't need to add bugs...
Come to think of it, i think i'll try emojis on my next commit message... just to test...
Removed double quotes, added single quotes
someone later on -m "removed single quotes, added double quotes"
More commits == More money!!!
Real talk: I have worked at a company where an exec briefly had the bright idea to rank developers on number of commits made.
You can imagine how that turned out.
He wasnβt an exec for very long.
lmaoo, I can't believe someone actually did that and was serious about it.
I'm already imagining commits piling up with comments like:
users
touser_list
"user_list
and that's it, don't you dare change it again"Good riddance I guess
Oh, no, we can do better.
Every single line of every single PR is an individual commit. And never squash-commit.
Talk about atomic commits
If you really want to nuke it from orbit: do it for every character.
It is absolutely technically feasible to go further than that, but I think thatβs the point of diminishing returns in terms of effort.
Hah, I knew it would come to that
Not if you make another application to automatically split commits in that way... application which will need its own commits, it's never-ending
Why do u prefer single quotes?
Not Shall330, but double quotes can imply variables in the string in some languages.
Which languages are these?
PHP for one. "$x" will interpret the variable $x. '$x' does not.
Well there's ya problem ur using php.
Also bash
And Perl.
I have no preference! It was to do with Spring/yaml config and some really strange conflict which required the use of single quotes. I'm still a total noob in the world of software dev, so I wouldn't be able to explain why it worked π€£
"Update"
It was my dotfiles, ok?
You're going to regret not adding the second sentence to the commit message some day.
Bold of you to assume there is a cohesive theme to dotfile commits. Usually it's just been long enough that I don't want to lose my latest tweaks
Squash me later
My commit messages have gotten extremely lazy since I start squashing all my commits down to one. I just describe the PR on the first commit message and write nonsense in all the others.
I've been adding the commit shas + messages to the final description, if anyone wants to see my exploratory work.
I know that if you are on the local repository where the commits were originally created they'll remain accessible through recovery methods but AFAIK orphaned commits aren't synced to other machines.
That's correct. This is for work, which uses GitHub. The dangling commits remain accessible via their sha through the web ui, so I can link them in the PR description. I don't put them in the actual commit message.
I think these are garbage collected eventually, but no idea on cadence. It's long, anyway.
Tests: Add a few test files.
I thought we lost you ... Jia Tan!
Initial commit
grats on formalising your (new) project
Initial commit #463
God read my code and it made Him cry
"Changed consent banner text to reflect from legal, even though they are wrong"
"circumvented GDPR despite calling out strong ethical objections, because legal & management are forcing me to"
Update coffee list and tex files.
...and yes we have our office coffee list in a git repo.
Do new hires have to submit a pull request to get in on it? If so, I'm guessing the non-technical staff don't get to drink a lot of coffee.
It's actually just a paper list where people make ticks for coffee. That is then manually synced to an hledger file in git.
For what? To keep track of who's drinking coffee? Are you charging for coffee?
I work in academia... There is no free coffee here.
Mfw I
git push -f
Always force-with-lease
Do you have any plans to open source your coffee list? I want to add this feature where I get an iced coffee in the morning
"fix for real this time"
"28496 - there, it's fucking fixed you twat waffle."
Ticketed bug bosses son found. Dude nagged his dad who nagged us until it got fixed. Boss doesn't review code. And for the sake of a half dozen coworkers, I hope he never does.
"I'm sorry, guys."
";"
Previous commit was some stupid easy fix I didn't even bother compiling. Well, I should have, because it was the first time in recent memory I committed some code with a missing semicolon...
Before anyone asks : no, we don't do reviews Β―\_(γ)_/Β―
Join me and use
~
to mean "shit changed just check the diff"For those I just commit with the message "ngl there's a lot of changes in here"
+2543; -5642;
"Moved code base from 2 spaces to 4."
Hrfg7ygfrvvby
git commit -m "if this doesn't fix it I'm looking up availabilities at my nearest maccas"
"My laptop died before I could push these"
Not a good commit name I know, but my laptop stopped booting and I just really needed to continue from my desktop.
It's a student project anyway.
Fix for vendor lies.
French dead link?
Converted more tests to use the correct handle.
A
Because who cares for commit messages to my private repo?
You when you mess something up and only figure it out later?
I do Ansible playbooks so not too had to see when I ballzed it up.
Whatever works for you, but I love writing a quality commit message. It's like my reward for doing all that work - wrapping up the changes in a nice little summary.
Sometimes I just scroll through my commits when I'm feeling blue...
I had just figured out how to fix a bug that broke our in house reminder system. My smug ass put hoho as the commit message for figuring something out that the previous guy couldn't. Of course it came back to bite me in the ass THIS WEEK when the system broke again in a similar fashion, but I couldn't remember what I had done.
Diff
I know I know. I figured out what I did before I thought of that... which was just now when you said it.
Revert this later
Yikes
Prevent subprocess from killing itself until finished.
WIP
You copycat.
init: add readme
βInitial commit, come back laterβ
Fix 3
Lol test
hey, you did ask
modified: index.html
Rip
"render trans chunks after entities/sel box etc "
Very progressive.
"Debug"
sq
chore: ensure it's God's divine intellect
I wonder what the last commit message of my career will be.
I wonder what my last commit at each job was. I'll bet it was boring. About 10% of my commit messages are genuinely interesting.
For some definition of interesting β¦. The quashed commit whose message is a Jira ticket number and title?
This is what I hate about squashing. I love the fact that my crazy journey to a finished feature is there in the log for everyone to see, including myself lol. But squashing makes me feel empty when I write commit messages. Where's the fun... Just a straight line of perfectionism. Maybe someone has a crazy but cool idea that would be worth picking. π
Maybe. I do more DevOps these days, so tend to have many small changes that canβt even be tested without checking them in and running in CI. Iβd have hundreds of βfix unit testsβ commits alone
I've been thinking about that when talking to my DevOps colleagues, that there's gotta be a better way to test CI before committing. The whole change-commit-test dance would kill me if on a daily basis. So cumbersome.
We squash. I'm not really interesting in your local journey to land the change. It's sometimes useful during review, but after that it's mostly the state of the main branch I care about. It's what I need to bisect anyway.
I don't like commits that are just references to issues. Copy the issue into the commit message so
git blame
tells you something useful. Unless it's just closing a simple big. Then the title and issue reference are plenty.Depends on the project I imagine.
"it's possible i buried a self delete timebomb in one of our repos years ago. enjoy looking for it just incase assholes"
haaaands
Fuck it I quit
Update
.
"Merge pull request #8 from [branch name]"
Not the most exciting but hey, someone has to do it.
Merge commits π
" "
Mod: Added delay to let switches settle
Something along the lines of "I don't know what this does, I wrote this code two weeks ago"
Formatting fixes
Merge remote-tracking branch 'origin/master'
Ew - rebase all the things!
Fixes
Reusing Terraform projects for the win.
Changed index.html
Added zip code to the reseller portal when batch-creating transactions in
I try to start off descriptive, in case I want to diff a few changes back. After that, sometimes I write notes or general reflections on life in there... Until yesterday I thought no one was reading them
This is pretty funny, but now I want to watermark all of my future work with outdated "Made With ..." Stickers.
"Made With Printshop Deluxe!" "Made With Microsoft Works!" "Sent From my Sidekick!"
"cleanup" lol
Change port to 8080 because stupid container locks down 80
Fixed tests.
"update doc to reflect reality still more"
chore: replaced println() with proper logging util
I forgot to replace them water finishing debugging.
m
usually the commits at the end of the day when I haven't finished a task yet. It will be squashed and disappear eventually.
Lint issue
Uhh it's on a foss project so imma not dox myself lmao
My commits are a little too specific to post here, lmao. I likely put too much info into them, Iβm still trying to get a good balance.
I think long messages are a good habit. Start with something readable in the history, past that who cares? Most people rarely read past the preview, and if they do they want details
I think it's great because it makes you reflect on what the goal was and what you did. I sometimes stop to make a quick change as I'm writing, or just collect my thoughts before mentally dismissing the task
Update
"Add ability to load sources which are not a nginx file listing."
I try to be descriptive while keeping it short.
βSome appropriate message.β
Quack quack
"Couldn't count all the stuff"
?
WIP
"Monte Carlo code changes. BLAME: notme"
fix this fucking gitignore
latest cargo crates updated
Minor text fixes
Stuff
Beta 0.9.9.9.9.9
Lots of updates (it's a private repo and I'm bad at doing smaller individual commits!)
some fixes
Fixinβ thangz
feature suggestion but "hyprcursor"
CPEUI-20102: It's working but still clean up to do.
I'll never get time to do that "clean up"
WIP
Simplify link
Add links to navbar
curl -s https://whatthecommit.com/index.txt
Fixed boolean logic and bump version.
removed /32s in favour of /24s. Small layout changes.
Fix
h
1.1
Options for donation and contact links
Nothing importantRolling back demo commit
Add Exporters MVP
Comment
github-sync: local install (WIP, split and rebase)
changelog
somebody renamed the interface but didnβt rename the monitoring interface 2
first release
I wrote a brief manifesto on my hatred of Python because some refactoring ended up having a comma at the end of a line, which screwed me over for about an hour until I happened to noticed it.
Implement add headers
Oh i don't need to add bugs...
Come to think of it, i think i'll try emojis on my next commit message... just to test...
"Added backoffs and logging for rate limiting"
Offset controls that are grouped together
Submodule update