[#21039] Happy new year and... moving Ruby development to Git? — Michael Klishin <michael.s.klishin@...>
Happy new year everyone.
On Jan 1, 2009, at 6:42 AM, Michael Klishin wrote:
On Thu, Jan 1, 2009 at 11:22 AM, James Gray <james@grayproductions.net> wrote:
brabuhr@gmail.com writes:
Quoting Michael Klishin <michael.s.klishin@gmail.com>:
Hi,
-----BEGIN PGP SIGNED MESSAGE-----
T24gRnJpLCBKYW4gMiwgMjAwOSBhdCAxMjoxOCBQTSwgRmxvcmlhbiBHaWxjaGVyIDxmbG9AYW5k
My opinion:
My two cents:
Eust=E1quio Rangel wrote:
On Sat, Jan 3, 2009 at 21:40, M. Edward (Ed) Borasky <znmeb@cesmail.net> wrote:
Nikolai Weibull wrote:
On Sat, Jan 3, 2009 at 10:39 PM, M. Edward (Ed) Borasky
Giuseppe Bilotta wrote:
On Sun, Jan 4, 2009 at 12:14, Joel VanderWerf <vjoel@path.berkeley.edu>wrote:
Michael Klishin wrote:
On Mon, Jan 5, 2009 at 12:29 PM, Michael Klishin
On 05.01.2009, at 16:10, Rados=C5=82aw Bu=C5=82at wrote:
On Mon, Jan 05, 2009 at 10:22:49PM +0900, Michael Klishin wrote:
-----BEGIN PGP SIGNED MESSAGE-----
On Mon, Jan 5, 2009 at 10:06 AM, Florian Gilcher <flo@andersground.net>wrote:
T24gTW9uLCBKYW4gNSwgMjAwOSBhdCA4OjA4IEFNLCBNZWlucmFkIFJlY2hlaXMKPG1laW5yYWQu
On Jan 1, 2009, at 04:42 AM, Michael Klishin wrote:
On Sat, Jan 3, 2009 at 00:34, Eric Hodel <drbrain@segment7.net> wrote:
On Sat, Jan 3, 2009 at 4:48 AM, Michael Klishin
On Sat, Jan 03, 2009 at 12:48:09PM +0900, Michael Klishin wrote:
-----BEGIN PGP SIGNED MESSAGE-----
On Mon, Jan 5, 2009 at 4:59 PM, Florian Gilcher <flo@andersground.net> wrote:
-----BEGIN PGP SIGNED MESSAGE-----
On Mon, Jan 5, 2009 at 6:45 PM, Florian Gilcher <flo@andersground.net> wrote:
-----BEGIN PGP SIGNED MESSAGE-----
On Jan 2, 2009, at 17:25 PM, Nikolai Weibull wrote:
> I think I'm entitled to an opinion on the subject because I am a
On Jan 6, 2009, at 12:16 PM, Brent Roman wrote:
On Wed, Jan 07, 2009 at 04:03:12AM +0900, James Gray wrote:
Hi,
[#21094] [Bug #975] ruby curses extension does not support multibyte characters — Michal Suchanek <redmine@...>
Bug #975: ruby curses extension does not support multibyte characters
[#21097] [Bug #977] caller for all threads patch — Roger Pack <redmine@...>
Bug #977: caller for all threads patch
I made a patch to Thread#caller(lev=1). It may be more flexible than
One thing I think might be cool is rather than raising an error for a
On Mon, Jun 8, 2009 at 5:06 PM, SASADA Koichi<ko1@atdot.net> wrote:
[#21180] how to use git to work with MBARI patches — Stephen Bannasch <stephen.bannasch@...>
At 3:16 AM +0900 1/7/09, Brent Roman asked in a previous thread about git:
[#21181] ruby_version changes in 1.9.1-rc1 — "Luis Lavena" <luislavena@...>
Hello All,
[#21238] [Bug #996] IRB doesn't work anymore with ruby-1.9.1-rc1 on MinGW — Chauk-Mean PROUM <redmine@...>
Bug #996: IRB doesn't work anymore with ruby-1.9.1-rc1 on MinGW
[#21244] [Bug #999] SSL & ZIP missing from ruby-1.9.1-preview1-i386-mswin32 — William Mason <redmine@...>
Bug #999: SSL & ZIP missing from ruby-1.9.1-preview1-i386-mswin32
Issue #999 has been updated by Luis Lavena.
[#21259] Do I need a special build arg to get irb to accept utf characters on OSX — Dave Thomas <dave@...>
I'm seeing very strange behavior at the irb prompt with ruby 1.9.1 =20
On 1/11/09 1:06 PM, Dave Thomas wrote:
[#21269] [Bug #1005] Strange behaviour of File.directory? under Windows — Alex Fortuna <redmine@...>
Bug #1005: Strange behaviour of File.directory? under Windows
[#21289] MRI 1.8.6 bug - exit is ignored if ensure clause throws exception which is caught by outer rescue — Shri Borde <Shri.Borde@...>
The following code snippet does print "after foo" when using MRI 1.8.6 on W=
[#21302] [PATCH] drastically improve rb_waitpid for short-lived children — Evan Phoenix <evan@...>
This evening I wrote this patch while helping a friend diagnose why,
[#21309] Current status of 1.9.1 RC2 — "Yugui (Yuki Sonoda)" <yugui@...>
-----BEGIN PGP SIGNED MESSAGE-----
[#21315] Proc-as-Binding is out again? — Charles Oliver Nutter <charles.nutter@...>
1.8 can use a proc as a binding:
It seems to be more interface based rather than disallowing the access
[#21339] [Bug #1010] Ruby-1.9's rake sh doesn't work on Windows (but fix provided) — Chauk-Mean Proum <redmine@...>
Bug #1010: Ruby-1.9's rake sh doesn't work on Windows (but fix provided)
Issue #1010 has been updated by Chauk-Mean Proum.
[#21345] [Bug #1011] Chinese GB18030 transcoding is breaking build at r21509 — Luis Lavena <redmine@...>
Bug #1011: Chinese GB18030 transcoding is breaking build at r21509
[#21383] [Bug #1019] irb/xmp fails because StringInputMethod doesn't support #encoding — Dave Thomas <redmine@...>
Bug #1019: irb/xmp fails because StringInputMethod doesn't support #encoding
[#21399] Proposal: Module#copy_method — Yehuda Katz <wycats@...>
I'd like it to be possible to copy methods from one module to another. The
2009/1/18 Yehuda Katz <wycats@gmail.com>:
Excerpts from Pit Capitain's message of Sun Jan 18 11:23:48 +0200 2009:
Yehuda Katz wrote:
AOP requires being able to hook into method dispatch, which either requires
Yehuda Katz wrote:
So I guess the question is... what's wrong with the implementation ;)
Yehuda Katz wrote:
Just FYI, the define_method version fails if your method requires the use of
I believe that methods (and blocks defined as methods via method_defined) a=
Hi,
Hi,
Yukihiro Matsumoto wrote:
Hi,
For my use-case, it would be ok to support copying methods to ANY module,
On Tue, Jan 20, 2009 at 10:15, Yehuda Katz <wycats@gmail.com> wrote:
[#21444] ruby 1.9 hashes, etc. — "Roger Pack" <rogerdpack@...>
I noticed in the NEWS file that 1.9 has more compacted hashes--does it then
On Mon, Jan 19, 2009 at 10:20 PM, Roger Pack <rogerdpack@gmail.com> wrote:
[#21448] [PATCH] Allow building ruby with mingw — Alexey Borzenkov <snaury@...>
---
[#21454] [Feature #1027] Dir.home — Kazuhiro NISHIYAMA <redmine@...>
Feature #1027: Dir.home
Hi,
Excerpts from Nobuyoshi Nakada's message of Wed Jan 21 05:19:23 +0200 2009:
[#21458] Repost: Evaluation order of block pass versus arguments — Charles Oliver Nutter <charles.nutter@...>
I posted this nearly a year ago and never got any replies. I'm reposting
[#21460] Array concat loops sometimes taking large amounts of "system" time. — Ron Mayer <rm_rails@...>
On both ruby1.8 and ruby1.9, about half the time when I run loops like this:
[#21473] Time.gm give me ArgumentError out of range (ruby 1.8.6 (2007-09-24 patchlevel 111) [i486-linux]) — Claudio Fiorini <claudio@...>
Hi all,
[#21492] 2 small warning fixes for 1.9.1 — Marcus Rueckert <darix@...>
hi,
[#21495] [Bug #1033] wrong documentation for Kernel#catch — Patrik Wenger <redmine@...>
Bug #1033: wrong documentation for Kernel#catch
[#21550] [Feature #1046] request: ability to run without specifying .rb — Roger Pack <redmine@...>
Feature #1046: request: ability to run without specifying .rb
Issue #1046 has been updated by Roger Pack.
Hi,
>
Hi,
> |Oh sorry--I meant that you would probably find my answer [that I am lazy and
[#21552] [Feature #1047] request: getters, setters for the GC — Roger Pack <redmine@...>
Feature #1047: request: getters, setters for the GC
I agree that something in this direction should be done.
Issue #1047 has been updated by Yusuke Endoh.
[#21570] [Bug #1057] ripper does not compile with mingw — Charlie Savage <redmine@...>
Bug #1057: ripper does not compile with mingw
[#21598] [Bug #1060] mkmf refuses to find 3rd party extensions - ruby 1.9.1 trunk — Charlie Savage <redmine@...>
Bug #1060: mkmf refuses to find 3rd party extensions - ruby 1.9.1 trunk
[#21605] Couple questions about testing Ruby interpreters — Michael King <kingmt@...>
I am trying to apply Brent's MBARI patches to Ruby 1.8.7 p72 to Ruby 1.8.6
[#21613] [Bug #1063] in `write': Not enough space - <STDOUT> (Errno::ENOMEM) on Windows XP — Nick Gorbikoff <redmine@...>
Bug #1063: in `write': Not enough space - <STDOUT> (Errno::ENOMEM) on Windows XP
Issue #1063 has been updated by Tinco Andringa.
[#21618] UID Too Big, A Bug? — James Gray <james@...>
Is this a bug in Ruby?
[#21640] [Bug #1068] Ruby Cannot Handle Some UIDs — James Gray <redmine@...>
Bug #1068: Ruby Cannot Handle Some UIDs
On Wed, Jan 28, 2009 at 05:00:05PM +0100, James Gray wrote:
Hi,
On Thu, Jan 29, 2009 at 08:53:38AM +0100, Nobuyoshi Nakada wrote:
Ondrej Bilka wrote:
On Jan 29, 2009, at 3:43 AM, Urabe Shyouhei wrote:
On Thu, Jan 29, 2009 at 8:33 AM, James Gray <james@grayproductions.net>wrote:
On Jan 29, 2009, at 7:57 AM, Rocky Bernstein wrote:
On Thu, Jan 29, 2009 at 9:04 AM, James Gray <james@grayproductions.net>wrote:
One small edit I'm sorry I can resist since after all this *is* a
[#21701] [Feature #1081] add File::write() convenience method — Suraj Kurapati <redmine@...>
Feature #1081: add File::write() convenience method
Issue #1081 has been updated by Yusuke Endoh.
Hi,
Hi,
Is it intended that the length of the leading nulls are not included
Hi,
>> Is it intended that the length of the leading nulls are not included
Hi,
Teamwork. :-)
2010/3/6 Run Paint Run Run <runrun@runpaint.org>:
[#21702] [Feature #1082] add Object#singleton_class method — Suraj Kurapati <redmine@...>
Feature #1082: add Object#singleton_class method
Issue #1082 has been updated by Suraj Kurapati.
Hi,
> We haven't met any consensus of a name for the method.
+1 for Eigenclass
+1 for metaclass, as is it compatible with ActiveSupport, Rubinius and _why=
Hi,
Hi,
Hi,
On Mon, Jan 4, 2010 at 9:41 PM, Yukihiro Matsumoto <matz@ruby-lang.org> wro=
Hi --
Hi,
On 23.02.10 08:14, Shugo Maeda wrote:
Hi,
Issue #1082 has been updated by Yusuke Endoh.
Hi,
[ruby-core:21089] Re: Happy new year and... moving Ruby development to Git?
On 03.01.2009, at 4:25, Nikolai Weibull wrote: > Or at least get our facts straight and present valid cases for the > SCMs that we prefer? I use Git and Mercurial every day: Git for work and open source projects, and Mercurial for all my personal documents, emacs.d, tools, configuration files. I happened to use Darcs for about a month. Now here is my perspective. First, don't trust me when I say something about Windows support below,, simply because I do not use Windows. I can't say if msysGit still has significant problems or not, I'd trust Luis Lavena on this. Another note: I am 10000% ok with using Hg instead of Subversion or Bzr instead of Subversion, but because this thread moved to "git vs hg vs bzr vs darcs vs git-is-for-nerds-well-not-really vs freebsd-is- not-linux-you-mofos, I just want to waste some time and tell my personal experience with them. if you are really interested in a summary, read Learning curve section and the last section. Others are mostly noise. Learning curve vs. Feature richness: Git has learning curve, but once you get it, it *is* as simple to use as any other VCS. But when you hit the wall and you need an "advanced" feature (from "what commit introduced this issue" to "what subcomponents of the project change more often and thus need more attention"), it literally can do wonderful things. Mercurial is easier to learn, and it feels more like home for Subversion users. Even commands are pretty much the same, though bringing in some new concepts and trying to tie them to existing vocabulary of the SVN world leads to more confusion in certain cases (ex.: update/merge). It has pretty much any feature Git has, but sometimes you have to go and install plugins to get them. And some plugins are plain stupid, hg log does not use pager by default, but a pager plugin forces *every* single command to go through a pager, which is not really what you want. But in general, I have no plans of moving my personal stuff from Mercurial. Won't pretend to know the Bazaar situation, but I think it is on par with Mercurial, and is quite friendly, and has everything you possibly may need either built in or as plugins. Darcs is quite friendly and interactive, which looks very nice at first, but later when you try to merge with the rest of the team, and screaw up with all this "pick a hunk" interactive awesomeness, you either try to avoid all this interactive stuff, or it brings as much irritation as pleasure initially. So Darcs is the most nerdy, not Git. Branching and merging: Git stores all the local branches in a single working copy, and you can switch between them, delete them, rename them, etc in one physical directory on your HD. I always preferred this approach to say what Mercurial has, though it is not crucial. Since branching/merging is essential part of the process Git was created for, it has all the features in the world related to branching, and more: fetch/ pull to confuse people who barely know how "push" is different from "commit"; rebase so you have a way to screaw up even if you think of yourself as of advanced user; and other fun things ;) But 95% of the time you can do the job without them, and Git brancing/ merging can be distilled to like 3-4 commands you can easily explain in a contributors guide. Darcs has idea of "every branch is a separate directory" (or used to have it). I can't say it is a problem but you will end up having 3-5 copies of the pretty much the same code, and if you repository is 600 megs, not all movies you try to download from iTunes may fit into your HD. Mercurial has local repository branches, but you... cannot delete them without touching your nose with yout foot first! There are special plugins that try to provide this sort of functionality, and none of them really did what Git has when I tried. So, for long living branches you end up having directory copies, just like with Darcs, and like I said, it sounds like an ugly thing to me. Again, we are sacrificing that hot new Angelina Jolie movie downloaded by iTunes/Miro, so maybe it is a big deal for some. I cannot really judge bazaar here, but looks like again, it is directory copies here and there. Maybe this thing is some golden standard of DVCS world Git developers had no idea about. Sharing of repositories: I only mean technical aspect here, GitHub vs. the rest of the world discussion comes at the end of this message. All git, darcs, bzr, hg can use SSH and HTTP to sync with remote repositories, and have some sort of web UI if you really need it. What else would you wnat? Really useful features vs Crap you never use: bisection, a *very* useful feature, is supported by each and every of them. Stashed commits, squashed commits (merge 100 commits as one large) and rebases, things that advanced users like for a real reason, are supported by every system I think. The rest is something 99% of the population does not use. Performance: I don't know why people who maintain projects even of Rails size care so much about Git being super fast, it should be a problem for KDE and gcc and not Ruby libraries. But. Git performs the best, Mercurial and Bazaar are very fast, too. Mercurial, in fact, is so fast I used it on a very old (5 years or so) machine and never though that it is written, to large extent, in a very high level language. Darcs is probably slowest, but not too slow anyway. So unless MRI grows to the size of gcc, which I hope will never happen, they all perform reasonably well or better. UI front ends and tools support: A note for Emacs users: vcs-mode backends are available for at least git/hg/bzr, and the keymap is the same. So, you may not even notice that you have hg and not git under the hood. magit, gitsum and a bunch of Hg modes available for those who need it. Win! Vim users may chill out and stop screaming "we have it too, and better!!!", we all know ViViVi is The Devil's Editor, and one day the higher power will remind us of our sins, but I am sure vim has a counter part for every Emacs feature ;) NetBeans has support for Hg out of the box, Eclipse has plugins I cannot really judge... not sure how many MRI contributors use those behemoths. Git has like 3 or 4 standalone UI programs for OS X, a couple for Linux and probably has no UI tools on Windows. Mercurial said to have a good Windows tool (TortoiseHg), but on OS X and Linux it is a completely different story. I never really used a UI tool on OS X, but I know some people who were looking for one, and what they found was miserable. I have no idea what Bazaar has but it probably does better than Hg since Canonical cares about less technical users. Darcs seems to be lacking UI tools altogether. There will be bugs: If something goes wrong with Git, you have a very vibrant community support, and bug fixing process is pretty fast. Mercurial and Darcs are in Python, which means, you even can even try to fix the problem yourself in a reasonable amount of time. Darcs is in Haskell, and has the smallest community of them all, so unless you are functional programming geek, the first time you come around a bug, it is probably gonna be the last time you used Darcs. Availability of Github: Now, to the crucial part. Everything you read above simply does not really matter. Because GitHub is only available for Git. Hg, bzr, darcs all lack github. They may have launchpads, space ships and invented a way to store bits in a bitbucket, but they lack GitHub. And GitHub totally changes the way people work on open source, the way it should be. Open source is about project ecosystem and how parts (developers, companies) of that ecosystem interact. With GitHub you don't have to invent ways of collaboration, tools to visualize what you want to see, and simply poking that guy who created that cool thing but introduced a bug that you fixed and want to push upstream. Bitbucket and Launchpad may be both great code silos, provide wikis, trackers and so forth, but they miss the crucial point: forking, pull requests and merging should be so simple 13 y/o can do it, and GitHub gets this part right. Seriously, what made a lot of projects in the Ruby community switch to Git? The fact that it is "new hotness"? Performance? The fact that Hg and bzr are in Python? Linus Torvalds fanatism? Marking efforts of Chris, Tom, PJ and co? No. Ease of collaboration that is not really possible without GitHub at this moment. MK