views:

3325

answers:

64

I mean, we've all done it, making some changes and the checking them in with messages such as "made some changes" or "fixed a bug." Messages so inane, so pointless, you might as well have written "magical fun bus" in their place (of this, I am guilty), as it would be, perhaps, more descriptive. I ask you then, what is the most pointless, most off topic, strangest, or just WORST commit message you have ever authored?

+110  A: 

The worst must be no comment at all. And I am guilty... :-)

Rune Grimstad
+1: exactly what I was about to write...
Fredrik Mörk
Should be community wiki. Also, not the worst. Absolute worst is not using source control. But I swear, I'm changing that!
Chris Lutz
Actually, an incorrect comment is much worse than an emtpy comment. E.g. "fixed comment" when you actually changed a method signature...
sleske
I do exactly the same. I just don't think about it.
Liam
Not the worst, but pretty bad.
Nate Bross
Second worst. The worst of all is 'a'. The dev took the time to bypass the 'Comment must not be null' rule, but didn't take the time to document what the heck just went into the system.
glasnt
A: 

[insert friends name] is the bomb.

when you fix one of their errors. Became some sort of hype among friends after one started it when fixing useless stupid errors. Though the projects were not that big :).

bastijn
+4  A: 

Just checked git log and I found Corrected mistakes (committed by me!) which I think is not a good commit message as it does not shed light on what mistakes have been corrected. I should have added a description.

Alan Haggai Alavi
I once had a coworker who used only one commit message in two years: "updates" ;)
Aleksandar Ivanisevic
That is shocking! :-)
Alan Haggai Alavi
@Aleksandar: I didn't know we had worked together..
Chris Lively
+7  A: 

This is something like:

Updated

mouviciel
+14  A: 

Switched off unit test X because the build had to go out now and there was no time to fix it properly.

hkdk3107
+1 - Assert.IsTrue(1 == 1) would have also worked ...
Martin
+2  A: 
  • Committed some changes
  • Some bugs fixed
  • Minor updates
  • Added missing file in previous commit

I feel really sorry most of them were in school projects I promise :-(

victor hugo
at least you used a vcs on school projects :)
SnOrfus
"Added missing file in previous commit" - dont think this is too bad if previous commit was descriptive enough.
Owen
+4  A: 

My usual nondescript commit message:

oops

This usually results from committing more than was intended with svn commit. However, now that I use Git, this essentially never happens. Not only do I not generally make the mistake of committing more than was intended (I don't use git commit -a), but Git allows me to fix the bad commit before pushing the commit anywhere else!

Greg Hewgill
The good thing about "oops" is that usually looking at the commit immediately before that one tells you what it's all about.
Joachim Sauer
+4  A: 

"fixed errors in the previous commit"

Says it all really.. i.e. stuffed up one commit, then failed to even describe what was wrong (and what was corrected)..

RobS
+23  A: 

Much worse than no message at all is the message that looks right but is wrong in subtle and misleading ways.

Ed Guiness
Fully agree. There are times when no message, or something like "bug fixes", works reasonably well. If you don't have an issue tracking system (and I don't at home), you can't cross-reference to issues. If it's some minor bugfixes, you can see what happened easily enough from the version diff.
David Thornley
+4  A: 

STUFF!!!!!11

For my master thesis' latex repository.

bayer
+30  A: 

"." or "changes"

Patrick Cornelissen
+1. "." was always a favorite of mine :-)
paxdiablo
Surely you mean "................" as there's a trigger which checks for a minimum of 16 characters
Pete Kirkham
Pete, this depends on your setup. :-)(And the VCS you use)
Patrick Cornelissen
I prefer "x" (but in SO you have to type 15 chars)
erikkallen
In this case check Petes "solution" ;-)
Patrick Cornelissen
+1 from my muscle memory, which fondly remembers git commit -a -m '.'
ide
+11  A: 

Can't take full credit for this one, but one of the web designers was tasked with adding corporate advertising to the internal homepage. Check-in message:

Added a banner to the default admin page. Please have mercy on me =(

Mark Glorie
I thought you were going to say there was an ad in a commit message; "Fixed user login but --This commit brought to you by Burger King."
chap
A: 

'hmmmmmm'

Added that once when I couldn't even be bothered to think of something crap to write as a comment. very bad.

Someone else
+5  A: 

Not the worst, but

oops, forgot to add the file

is close enough.

hmemcpy
+1  A: 

"work in progress"

kay.herzam
+2  A: 

i see these:

bug fix

(the file has been there for 1 year and gone through 60 revisions... so "bug fix" doesn't mean anything).

typo

(when i check the diff, there actually is a bug fix. so i wonder, is "typo" just to make the mistake look not so bad).

[blank]

(the CEO and CTO like to use empty comment).

動靜能量
+1  A: 

A cursory look at a project I've been doing the past one year where I did not use git seriously for the initial three quarters of the project year, revealed these among other commit messages:

  • minor changes
  • some brief changes
  • assorted changes
  • lots and lots of changes
  • another big bag of changes
  • lots of changes after a lot of time
  • LOTS of changes. period
ayaz
+4  A: 

"Nothing to see here, move along" - When there were actually some major canges within the project. I just couldnt be bothered writing something useful.

I've used this after initially committing something that was broke then quickly committing a fix.
Chris Lively
A: 

“Obligatory placeholder commit message”.

Yeah, I wasn't lazy to type it out, but I was lazy enough to think of something better.

apostlion
+2  A: 

Alex made me do it

Omry
A: 

A long time ago, in a galaxy far far away...

kevchadders
+1  A: 

"Test commit. Please ignore"

jhs
I don't think I ever committed that to a real project but that's the one that drives me crazy. One of the benefits of distributed SCM is that you can test commit on your own and then uncommit before going public.
jhs
A: 

"fix" for a 7 kb commit ... oops :-)

none
+17  A: 

Whooohoo Revision 666!

On a pointless change of course.

Ranieri
Hehe, I did something similar on revision 1337 once :P
cwap
A: 

"committed changes"

A: 

"Fixed the build"

While that does get to the point of why I'm commiting the change, it doesn't say anything about what I did change. That's why we have file diffs, right?

JB King
+5  A: 

"put code that worked where the code that didn't used to be"

Mike Two
+1  A: 

"I'm just a grunt. Don't blame me for this awful PoS."

Eduardo León
+1  A: 

Revision 1024!!

thephpdeveloper
+55  A: 

I feel this is just too good not to share. It was a programming assignment for our algorithms class. Apparently, we did most of the work after midnight; presented below are the more random/profane git commit messages:

  • JOSH, WE WENT OVER THIS. EXPANDTAB.
  • STEVE, WE WENT OVER THIS. C++ IO SUCKS.
  • Some shit.
  • add actual words
  • I CAN HAZ COMMENTZ.
  • giggle.
  • Whatever.
  • Finished fondling.
  • FONDLED THE CODE
  • this is how we generate our shit.
  • unh
  • don't be retarded.
  • It works!
  • unionfind is no longer being molested.
  • Well, it's doing something.
  • I'M PUSHING.
  • Whee.
  • Whee, good night.
  • It'd be nice if type errors caused the compiler to issue a type error
  • Fucking templates.
  • I hate this fucking language.
  • marks
  • that coulda been bad
  • hoo boy
  • Fucking egotistical bastard. adds expandtab to vimrc
  • if you're not using et, fuck off
  • WHO THE FUCK CAME UP WITH MAKE?
  • fix some fucking errors
  • first blush
  • uhhhhhh
  • forgot we're not using a smart language
  • include shit
  • things occurred
  • i dunno, maybe this works
  • 8==========D
  • whooooooooooooooooooooooooooo
  • clarify further the brokenness of C++. why the fuck are we using C++?
jleedev
"why the fuck are we using C++" best one :)
romkyns
Some of my comments look like this.
tster
Is this *supposed* to be one big block?
Joachim Sauer
Putting them on separate lines would have taken up more space, and I thought it was a funnier like this.
jleedev
+12  A: 

the IT guy noticed revision 666 was coming up and wanted to make it special.

I convinced him that: "needs more cow bell" would be ideal.

a few months later the head of production was rather distraught because she thought the client had requested it.

SketchBookGames
I put an Ascii picture of the USS Enterprise (Original) in for revision 1701.
Alister Bulman
+3  A: 

One of my worst was quite recent:

I am even stupider than I thought

robotadam
+1  A: 

PROJ-1254: Some reasonable message

Turns out I made 4-5 commits with a completely wrong ticket id, it should have been PROJ-2154

Even worse: our issue tracking software looks up related code changes from the repository. Therefore, if somebody ever looks at that 6-month-old issue they will find a lot of random code commits that don't even make sense tied to it.

Oops.

TM
This is why I'm glad I'm also in charge of our Subversion server. If I mess up a comment badly enough, I'll temporarily modify the Subversion hooks so I can `svn propedit` my commit messsage ;-)
Mike Spross
@Mike Spross: you don't need to use `svn propedit` if you're the admin, you can use `svnadmin setlog`
rmeador
@rmeador: I didn't know that. Time to re-read the SVN book ;-)
Mike Spross
+1  A: 

Ready for it?.....

I did it for the lulz!

We checked in an interface:

public interface ICanHasState

...this is in a respectable business package, no less. If it weren't for obfuscation, someone might actually find that.

Tullo
+17  A: 

The following are all real commits made by me on one project:

  • A fix I believe, not like I tested or anything
  • Useful text
  • pgsql is being a pain
  • pgsql is more strict, increase the hackiness up to 11
  • c&p fail
  • syntax
  • ericflo sucks
  • ERICFLO SUCKS
  • fix
  • someone fails and it isn't me
  • Gross hack because eric doesn't know how to code
  • totally more readable
  • better grepping
  • fix
  • fix bug, for realz
  • fix /sigh
  • Does this work
  • MOAR BIFURCATION
  • bifurcation
  • REALLY FUCKING FIXED
  • FIX
  • better ignores
  • More ignore
  • more ignores
  • more ignores
  • more ignores
  • more ignores
  • more ignores
  • more ignored words
  • more fixes
  • really ignore ignored worsd
  • fixes
  • /sigh
  • fix
  • fail
  • pointless limitation
  • eppic fail eric
  • ajax-loader hotness, oh yeah
  • ericflo broke the regex, lame
  • small is a real HTML tag, who knew.
  • WTF is this.
  • Do things better, faster, stronger
  • Fixed a bug cause brosner said to
  • Use a real JS construct, WTF knows why this works in chromium.
Alex Gaynor
+1 for "more ignores". And again...and again....and 'crap, I forgot to ignore that directory as well'...
richsage
Apparently this Eric fella is holding you guys down.
Baddie
A: 

WARNING xyz.php doesn't work in this checkout.

Pekka
+8  A: 

Back in 90's when we didn't have a revision control system, we were used to put change log entries at the top of the source code. Below are some of my entries from our GUI library. Jury shall consider that the defendant was 17 years old and a beginner in English back then :)

// 25th Sep 93 - 01:17 - I must sleep... it's working... in just three hours...
// 26th May 94 - 02:05 - SSG DID IT AGAIN!!!! THE ULTIMATE CHANGELINE
//                       DETECTION ROUTINES!!!!...
// 19th Aug 95 - 18:26 - I was wrong...
//  2nd Apr 94 - 21:12 - Completed with no bugs...
//  6th Aug 94 - 20:34 - Fixed a little bug...
// 10th Dec 94 - 02:11 - Fixed a bug in NoteLineCount... not seriously...
// 27th Aug 97 - 23:58 - woa!! this one was really HARD!
// 29th Jun 97 - 00:48 - Made it to compile...
// 13th Dec 94 - 01:04 - changed things...
// 26th Dec 94 - 00:22 - touched... (cs=2626)
// 31st Dec 94 - 00:21 - i think i fixed a bug... (cs=2605)
//  7th Oct 96 - 14:25 - perfect...
// 29th Nov 93 - 02:50 - Moved something to somewhere... goodnight...
ssg
Moved something to somewhere...LOL!
flq
The dates aren't even in order - he just inserts a new comment 'somewhere' at the top.
Kirk Broadhurst
Kirk, this is a compilation from multiple sources. Dates were in order for each source, at least :)
ssg
+2  A: 

Detailed description of an other change, which I made earlier to another file, simply because my IDE remembers the message I entered for that other commit. IMO this is much worse than a meaningless or empty message, since it's misleading.

ammoQ
+2  A: 

"I'll explain this when I'm sober .. or revert it" .. thanks to git, I can't give a link.

Tim Post
A: 

got few times comments in foreign languages,,, good ones are the ones in which programmer speaks to the code ie. "got it bitch" ...

eugeneK
+1  A: 

when i had improved performance of slow code,i changed engine started as ferrari engine started

Alexander
A: 
oops

Looking back in the commit log, I did this 3 times in a row, about a month ago :-/

David Oneill
A: 

fixed a [insert guilty coworker's name here] bug....

Jordan
A: 
JuanZe
A: 

git commit -m "Inital Commit"

No description, and too often used... something more useful/creative?

CodeJoust
A: 

here are some choice ones from our commit log:

  • missing functions
  • tightened up the java in my service call
  • adding my
  • story commit to release lock for joe
DefyGravity
+51  A: 
(\ /)
(O.o)
(> <) Bunny approves these changes.
Eduramiba
+1 ha ha ha ha! (had to add 3 ha's to be enough for a valid comment)
Chris Ballance
Are you sure Bunny approves? He looks a wee bit angry/scared there... o_O
glasnt
+2  A: 

"good stuff only" .. but it's better in Swedish: "bara bra grejjor".

You know, like, it's good stuff, like. That's why I committed it, like. ;)

Marcus Lindblom
+2  A: 

Classic for me and the other developer in the office used to be:

"Continued development..."
Gav
A: 

The worst I’ve personally done (a few times) was the following:


That’s not actually blank because a commit message was required, so I entered a non-breaking space (&nbsp / 0160).

I know, I know; it’s cheating. But to be fair, I’m the only one using this system, and the changes were really trivial.

Synetech inc.
+5  A: 

"Does anyone read this? I'll be at the coffee shop accross the street."

Brian Leahy
+3  A: 
That's just how I roll
Chris Ballance
A: 
Joel - 20/07/09

And the most prolific commit message I've used -

.
gamers2000
+1  A: 

Besides the obligatory "Fixed some bugs" I had some "Final version" messages in my bachelor thesis svn repository logs. There were about 30 of them, evenly distributed among the last two days before the final deadline.

Kitto
+1  A: 

This isn't one I did, but one I saw:

D

Although this looks like a typo, what it really meant is "Converted source from using Guardian C system calls to using our proprietary D-series procedure wrappers."

orbfish
+1  A: 

-Various changes

Pat
A: 
(m)

Which means "minor change".

bugfix

then

Why has the +"*ç%&/ Visual Studio not added the file?

Although I have to admit that I don't regard any of these as bad: In the log, an "(m)" usually indicates: You don't need to look into this revision, the significant changes have been done before or after. The "Why has +"*ç...." indicates: This change has actually been explained in the last commit, so don't be surprised not to find it there.

chiccodoro
+2  A: 

No matter how often I asked, educated & pleaded, I used to work with someone who kept entering:

I change some files

She left before I killed her.

Kris Jenkins
A: 
(This space intentionally left blank)
Ed Daniel
A: 

On revision 2013 of my current project:

We made it past 2012 and the world did not end!

Jay
+1  A: 

fortune -o | cvs ci -m

Martin Broadhurst
A: 

"lol it almost works sharks with lazerbeams on their foreheads"

stefan_g
A: 

Sometimes I run these for shits and giggles:

git log --oneline | grep WTF
git log --oneline | grep thingy
git log --oneline | grep oops

It's usually good for a laugh.

George Mandis
Needs moar rebase.
rpkelly
A: 

My favorite: Frist!

Chris Lively