|
i don't understand why anyone wouldn't use git (or *sigh* mercurial) locally regardless of what central vcs the project uses, literally why would you handicap yourself like that
|
# ? Jan 20, 2013 00:11 |
|
|
# ? May 9, 2024 18:54 |
|
why would you add an extra step between you and the server on purpose
|
# ? Jan 20, 2013 00:17 |
|
svn workflow: write code, commit. git workflow: write code, commit, push.
|
# ? Jan 20, 2013 00:18 |
|
rotor posted:I've administered several and have never had an issue. how did they get hosed rotor posted:why would you add an extra step between you and the server on purpose
|
# ? Jan 20, 2013 00:20 |
|
rotor posted:svn workflow: write code, commit. svn workflow: write code, commit, wait, commit, wait, commit git workflow: write code, commit, commit, commit, push (i work in chunks that i want to commit separately as they're separate thoughts)
|
# ? Jan 20, 2013 00:22 |
|
yeah I mentioned that if you're poor that git is a good idea.
|
# ? Jan 20, 2013 00:24 |
|
if you don't have the money to set up a good svn machine on a fast connection, or if you're in swaziland and have to push things out over a cell modem then it totes makes sense.
|
# ? Jan 20, 2013 00:25 |
|
rotor posted:why would you add an extra step between you and the server on purpose
|
# ? Jan 20, 2013 00:28 |
|
isn't git on windows a bit rube goldberg though
|
# ? Jan 20, 2013 00:28 |
|
Gazpacho posted:are all the code changes you work on less than 10 lines or something commit early, commit often
|
# ? Jan 20, 2013 00:31 |
|
Base Emitter posted:isn't git on windows a bit rube goldberg though although it is a real butthead about newlines, i admit
|
# ? Jan 20, 2013 00:31 |
|
Base Emitter posted:isn't git on windows a bit rube goldberg though yes, and also everywhere else
|
# ? Jan 20, 2013 00:31 |
|
rotor posted:commit early, commit often
|
# ? Jan 20, 2013 00:33 |
|
Gazpacho posted:are all the code changes you work on less than 10 lines or something who gives a poo poo?
|
# ? Jan 20, 2013 00:33 |
|
i was joking about the source control bit but welp
|
# ? Jan 20, 2013 00:33 |
|
rotor posted:commit early, commit often and committing often works better with git
|
# ? Jan 20, 2013 00:33 |
|
Gazpacho posted:aka waste time with code reviews of things that don't even work yet, degrading the review process into a rubberstamp yeah I guess if you work somewhere with weird inflexible rules regulating every action you take that might be a problem.
|
# ? Jan 20, 2013 00:34 |
|
Cocoa Crispies posted:and committing often works better with git not really
|
# ? Jan 20, 2013 00:35 |
|
i tried using mercurial and it warned me that one of my files would take 221 megs of memory to manage and i was like wtf is this poo poo
|
# ? Jan 20, 2013 00:36 |
|
here's how it works: you start a feature branch. this is roughly analogous to a local repo. you commit your work here when you want to. if you need to switch out to another task, you do so. then, when you're ready to commit it to trunk/rc/whatever, that's when you do the code review and merge. its very similar to the git process except there's no local repo to get in your way.
|
# ? Jan 20, 2013 00:38 |
|
Lord help you if you ever try to maintain more than a single branch in SVN.
|
# ? Jan 20, 2013 00:38 |
|
Progressive JPEG posted:Lord help you if you ever try to maintain more than a single branch in SVN. its actually dead simple but sure I guess if you're a drooling simpleton it might be difficult
|
# ? Jan 20, 2013 00:39 |
what are some of you guys working on right now? i'm trying to figure out what to do next, spent some time last night adding/cleaning up some website poo poo with the one site i've been building. this is kind of my short list right now - write some actual testing for the site. this requires reading rails docs on doing so and it seems both boring and annoying as gently caress because i have to go back and read a bunch of poo poo and then figure out how to write the tests on top of my pre-existing code. this is mostly to appease people who are in a hiring position who feel that testing is apparently the only meaningful thing to judge someone by. this is probably due in part because it's rails people - maybe work on a corresponding app and experiment with pulling in json and poo poo to display information cause i've never done it. i almost have more of an urge to do this in android/java than ios/objc, idk why though. i really need to learn c proper. - do the online databases course stanford has going on because i suck at them also i've noticed that for some big of the projects, people on github seem slow as gently caress at responding to pull requests. took a couple days to start getting responses to a pull request i put in on rails for someone's issue.
|
|
# ? Jan 20, 2013 00:41 |
|
I'd say about 80% of the desire to use git lies with gits ability to hide your shameful dumbass intermediate commits from the public eye.. which is a lovely reason imo.
|
# ? Jan 20, 2013 00:41 |
|
rotor posted:did you get paid for writing it? then its good code. well it's not so much the quality of my own code that i'd like to accurately measure, but rather the exact amount of badness in someone else's code
|
# ? Jan 20, 2013 00:42 |
|
github is pretty cool for open-sores development where you want to pull in random code from wherever like some kind of whore.
rotor fucked around with this message at 00:45 on Jan 20, 2013 |
# ? Jan 20, 2013 00:43 |
|
Win8 Hetro Experie posted:well it's not so much the quality of my own code that i'd like to accurately measure, but rather the exact amount of badness in someone else's code other peoples code is always bad. you can measure how bad by how many times you say 'gently caress!' while integrating it.
|
# ? Jan 20, 2013 00:45 |
|
rotor posted:its actually dead simple but sure I guess if you're a drooling simpleton it might be difficult so in git it's git checkout master git checkout -b feature_branch # do things git commit -m "fix some poo poo" src/us/bf1c/some_app/asdf/fartz/FileIChanged.java # lmao java filenames git push origin feature_branch # this is the part where i make the pull request the only command there that takes time is "git push" what's it like in svn?
|
# ? Jan 20, 2013 00:48 |
|
rotor posted:other peoples code is always bad. you can measure how bad by how many times you say 'gently caress!' while integrating it. I think I have a tendency to take those as a challenge
|
# ? Jan 20, 2013 00:52 |
|
It's like the story of that IBM dev who secretly installed a tool chain on his pc during the "cleanroom" fad and instantly became the team's star programmer, you are the IBM manager in this story
|
# ? Jan 20, 2013 00:58 |
|
rotor posted:other peoples code is always bad. you can measure how bad by how many times you say 'gently caress!' while integrating it. if you were management tier you'd say "other peoples code is always good 'cause you didn't have to write it yourself"
|
# ? Jan 20, 2013 00:58 |
|
we have a github business plan, which i insisted on when everything was being set up and it has never let us down ever and is totally painless. it makes the dumb people uncomfortable (mercurial users) and the attention deficit disorder people angry (wanted to set up a gitlab server).
|
# ? Jan 20, 2013 01:01 |
|
we have the $200/month github plan and i think github's waffled on and off about using our software so that's kind of cool github is one of the three companies i'd leave current job for (github, 37signals, valve)
|
# ? Jan 20, 2013 01:02 |
|
Cocoa Crispies posted:so in git it's pretty much the same except without the push at the end
|
# ? Jan 20, 2013 01:45 |
|
writing multithreaded c++ with the power of c++11 and the Microsoft Visual C++ compiler
|
# ? Jan 20, 2013 01:50 |
|
Cocoa Crispies posted:so in git it's In svn making a branch is a cp -R from /trunk/ into /branches/poo poo/. svn posted:Subversion has no internal concept of a branch—only copies. When you copy a directory, the resulting directory is only a “branch” because you attach that meaning to it. You may think of the directory differently, or treat it differently, but to Subversion it's just an ordinary directory that happens to have been created by copying. Merging code/applying changes across branches in this system is exactly as awful as you'd expect.
|
# ? Jan 20, 2013 01:55 |
|
Also as a side-effect the branch carries the same revision number as the main tree since it's all just thrown into one big one-dimensional repository/timeline, so if you're including the svn revision number in your versioning get ready for interleaved numbering across branches, resulting in a bizzaro world of "1.2-15832" and "1.0-18586".
|
# ? Jan 20, 2013 02:00 |
|
But svn can be fine in situations where your workforce is incapable of understanding branching, in which case svn isn't your bottleneck since it doesn't understand it either.
|
# ? Jan 20, 2013 02:03 |
|
Progressive JPEG posted:Also as a side-effect the branch carries the same revision number as the main tree since it's all just thrown into one big one-dimensional repository/timeline, so if you're including the svn revision number in your versioning get ready for interleaved numbering across branches, resulting in a bizzaro world of "1.2-15832" and "1.0-18586". and this is a big change from git, where version numbers are oh wait
|
# ? Jan 20, 2013 02:06 |
|
|
# ? May 9, 2024 18:54 |
|
branching is maginally nicer with git due to the per-line vs per-file diff
|
# ? Jan 20, 2013 02:08 |