Ok, let's bring something interesting to /prog/, version control. Some repository of scripts, on some version control system.
The cheapest way is to get a svn on google code (and publish user/pass), unless someone wants to host something else that integrates better with this public/anonymous system.
You do realize there is no way to recover the admin access now if someone goes and clicks on the reset button right?
Name:
Anonymous2008-03-18 22:17
And nothing of value was lost
Name:
Anonymous2008-03-18 22:20
>>14
So? prog.git has a mob branch that's editable by anyone. It's the most democratic repo possible.
Name:
Anonymous2008-03-18 22:21
updating 'refs/remotes/origin/mob'
from 0000000000000000000000000000000000000000
to 2118d66c9e7ceb38c3f4acfc1206cf2fdd448033
Generating pack...
Done counting 0 objects.
Writing 0 objects...
Total 0 (delta 0), reused 0 (delta 0)
The mob user can push only to the 'mob' branch, sorry
error: hooks/update exited with error code 1
error: hook declined to update refs/remotes/origin/mob
ng refs/remotes/origin/mob hook declined
error: failed to push to 'git+ssh://mob@repo.or.cz/srv/git/prog.git/'
how do i pushed
Name:
Anonymous2008-03-18 22:22
I fucked up the repo so that when you clone it, you don't get the mob branch automatically. You need to use gitk/giggle to create a mob branch, then commit from that. Sorry. If someone could fix that, that'd be awesome.
$ git clone git://repo.or.cz/prog.git Initialized empty Git repository in /home/anon/stuff/prog.git/prog/.git/ fatal: no matching remote head fetch-pack from 'git://repo.or.cz/prog.git' failed.
failed
oh well, maybe some other time ...
Name:
Anonymous2008-03-19 0:08
OK, let's try this again.
# get a copy of the repo
git clone git://repo.or.cz/4chanprog.git
# checkout the latest, mob-controlled branch.
git checkout origin/mob
# make a local branch based on the remote one.
git branch mob
# change files, commit to your local mob branch, etc.
#...
# push your changes to the mob branch.
git push git+ssh://mob@repo.or.cz/srv/git/4chanprog.git mob
# get latest changes in the mob branch.
git pull git://repo.or.cz/4chanprog.git mob
I'll periodically reset the merge mob into master. Hopefully this works.
Name:
Anonymous2008-03-19 1:40
>>4
How about fibonacci instead?
Yes, there's actually something somewhat worthwhile on the Git repo now. Get it, and start contributing!
>>24 lets begin with something trivial
true.c
------
int main() {
return 0;
}
false.c
-------
int main() {
return 1;
}
HAHAHA
in case he doesn't get it, *that's* system specific and will fail in other systems.
You need to include <stdlib.h> and return EXIT_FAILURE or EXIT_SUCCESS
1 might not indicate failure on a *lot* of systems. morons.
Here's your code improved.
/* true.c
cc -DANON_FALSE true.c for false.c
*/
A value of zero (or EXIT_SUCCESS, which is required to be zero) for the argument status conventionally indicates successful termination. This corresponds to the specification for exit() in the ISO C standard.
—POSIX.1
>>28
Correct, but did I say something differend? 1 might not indicate failure on a *lot* of systems. morons.
I did not say anything about 0. Which implies that 0 *does* indicate success (like EXIT_SUCCESS, but 0 and EXIT_SUCCESS need not to be equal)
and you know what? fuck POSIX. We are talking about ISO C here which covers an awful lot more systems. fuck you.
>>29
You said it was wrong to not use EXIT_SUCCESS. As for your main point, I think it's helpful to assume !success -> failure, regardless of the specific value. Returning the same EXIT_FAILURE for all types of failures would suck.
And ignoring POSIX leaves a very small base to work with considering the scope of the project.
commit 8614cd9838480cadd51c9009802bb2d5f89e4d9f
Author: Simon Peyote Joints <simonpj@world4chan.org>;
Date: Wed Mar 19 22:09:41 2008 +0100
Add an implementation of the factorial algorithm in Haskell.
Name:
Anonymous2008-03-19 17:45
I merged mob into master. Keep it up!
Name:
Anonymous2008-03-19 17:52
>>46
It keeps incremental backups of the versions of your files. You can have the entire source code history of your project, stored in a very space-efficient manner because it only stores the differences, not whole files.
Name:
Anonymous2008-03-19 18:05
Use darcs.
Name:
Anonymous2008-03-19 18:19
Can anyone recommend a good introduction to git? I'm really interested in this whole distributed version control system thingy.
Name:
Anonymous2008-03-19 18:19
>>52
Why? Because it's written in Haskell? Git is a great program, and the public Git repo is a great site.
>>83
All this freedom and public domain stuff is quite nice and such... BUT DON'T FUCK WITH MY SOURCE!
Name:
Anonymous2008-03-20 22:59
>>84
When someone forks a free project, it doesn't just supersede the original. This shit is gay.
Name:
Anonymous2008-03-20 23:19
I've tried Darcs and Git. I like the interactiveness of Darcs and patch-based setup. Git seems pretty similar, from a distance. I might switch to that seeing as it's supposedly faster, and I'm too lazy to conduct real tests. If anyone can provide me with benchmarks that are believable, I'd appreciate it.
Name:
Anonymous2008-03-20 23:23
GIT Is Tits
Name:
Anonymous2008-03-21 6:31
>>86
The Mozilla project evaluated quite a few distributed SCM tools and picked Mercurial.
Name:
Anonymous2008-03-21 7:00
>>86
Check out two projects, one using Git and one using Darcs.
Check the one with Git out first, because you will be waiting forever for the second project to check out.
Name:
Anonymous2008-03-21 7:30
>>88
They only chose Mercurial because it had a good Windows port at the time the evaluated DSCMs, but there's a very good Git port now, though it's not official yet.
Name:
Cudder2008-03-21 8:12
I should write a script to keep the prog.git, the rechan thread, and the rechan local archive in sync.