# git branch -a prefixes origin/ branches with remotes/ 7 messages from 2010-03-31 to 2010-04-01. Participants: layer, Jakub Narebski, Jay Soffian, Jeff King, Andreas Schwab, Michael J Gruber. Thread: https://gitlist.dev/t/23273 ## layer, 2010-03-31 17:26 Subject: git branch -a prefixes origin/ branches with remotes/ Message-ID: <18137.1270056387@relay.known.net> URL: https://gitlist.dev/e/18137.1270056387%40relay.known.net ``` In git 1.6.6.1 & 1.6.3.3, "git branch -a" output is like this * master remotes/origin/foo ... instead of like this for 1.6.1.3 * master origin/foo ... Is this a feature or bug? I searched the mailing list archives and announcements couldn't find any reference to this change. This is important to me because I use the output of "git branch -a" in scripts. Thanks. Kevin ``` ## Jakub Narebski, 2010-03-31 17:52 Subject: Re: git branch -a prefixes origin/ branches with remotes/ Message-ID: URL: https://gitlist.dev/e/m3aatoidl9.fsf%40localhost.localdomain In-Reply-To: <18137.1270056387@relay.known.net> ``` layer writes: > In git 1.6.6.1 & 1.6.3.3, "git branch -a" output is like this > > * master > remotes/origin/foo > ... > > instead of like this for 1.6.1.3 > > * master > origin/foo > ... > > > Is this a feature or bug? I searched the mailing list archives and > announcements couldn't find any reference to this change. > > This is important to me because I use the output of "git branch -a" in > scripts. Don't use git-branch output in scripts; git-branch is porcelain and its output is meant for end user and is subject to change! Use either git-show-ref, or powerfull git-for-each-ref. -- Jakub Narebski Poland ShadeHawk on #git ``` ## Jay Soffian, 2010-03-31 21:59 Subject: Re: git branch -a prefixes origin/ branches with remotes/ Message-ID: URL: https://gitlist.dev/e/k2r76718491003311459u4bbaeb4ax5b2e550acdd14c88%40mail.gmail.com In-Reply-To: <18137.1270056387@relay.known.net> ``` On Wed, Mar 31, 2010 at 1:26 PM, layer wrote: > In git 1.6.6.1 & 1.6.3.3, "git branch -a" output is like this > > * master >  remotes/origin/foo >  ... > > instead of like this for 1.6.1.3 > > * master >  origin/foo >  ... > > Is this a feature or bug?  I searched the mailing list archives and > announcements couldn't find any reference to this change. Feature. Introduced in 1.6.3: http://article.gmane.org/gmane.comp.version-control.git/116404 See "builtin-branch: improve output when displaying remote branches" > This is important to me because I use the output of "git branch -a" in > scripts. > > I should add that the scripts are merely testing for the existence of > remote tracking branches in the local repo. If there's a better way > to do this, I'm game. Thanks. git branch is a so-called porcelain command, and as you've now been bitten by, the output of porcelain's is subject to change. You probably want git for-each-ref. Aside, the only list of plumbing vs porcelain commands I'm aware of is in the bash-completion script under contrib. See __git_list_porcelain_commands here: http://git.kernel.org/?p=git/git.git;a=blob;f=contrib/completion/git-completion.bash Those are generally commands whose output you do not want to be parsing. All the other git commands installed under git-core are mostly plumbing upon which you can build additional functionality. (Note to git list: there should be a man page that has a single-line summary of each command and states whether the command is plumbing or porcelain. Maybe there is and I just don't know it.) j. ``` ## Jeff King, 2010-03-31 22:03 Subject: Re: git branch -a prefixes origin/ branches with remotes/ Message-ID: <20100331220340.GA4564@coredump.intra.peff.net> URL: https://gitlist.dev/e/20100331220340.GA4564%40coredump.intra.peff.net In-Reply-To: ``` On Wed, Mar 31, 2010 at 05:59:49PM -0400, Jay Soffian wrote: > Aside, the only list of plumbing vs porcelain commands I'm aware of is > in the bash-completion script under contrib. See > __git_list_porcelain_commands here: > > http://git.kernel.org/?p=git/git.git;a=blob;f=contrib/completion/git-completion.bash Try command-list.txt, which is meant to be authoritative. > (Note to git list: there should be a man page that has a single-line > summary of each command and states whether the command is plumbing or > porcelain. Maybe there is and I just don't know it.) Try "man git", which splits the command list by porcelain and plumbing (and further by subcategory). -Peff ``` ## Andreas Schwab, 2010-03-31 22:04 Subject: Re: git branch -a prefixes origin/ branches with remotes/ Message-ID: URL: https://gitlist.dev/e/m2iq8c2ln9.fsf%40igel.home In-Reply-To: ``` Jay Soffian writes: > (Note to git list: there should be a man page that has a single-line > summary of each command and states whether the command is plumbing or > porcelain. Maybe there is and I just don't know it.) git(1) Andreas. -- Andreas Schwab, schwab@linux-m68k.org GPG Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5 "And now for something completely different." ``` ## Jay Soffian, 2010-03-31 22:09 Subject: Re: git branch -a prefixes origin/ branches with remotes/ Message-ID: URL: https://gitlist.dev/e/q2y76718491003311509z9c2f83c2t4a217f7c68b4ff4f%40mail.gmail.com In-Reply-To: ``` On Wed, Mar 31, 2010 at 6:04 PM, Andreas Schwab wrote: > Jay Soffian writes: > >> (Note to git list: there should be a man page that has a single-line >> summary of each command and states whether the command is plumbing or >> porcelain. Maybe there is and I just don't know it.) > > git(1) Yes, I'm an idiot. Anyone else wanna pile on? :-) j. ``` ## Michael J Gruber, 2010-04-01 09:17 Subject: Re: git branch -a prefixes origin/ branches with remotes/ Message-ID: <4BB464B4.1080109@drmicha.warpmail.net> URL: https://gitlist.dev/e/4BB464B4.1080109%40drmicha.warpmail.net In-Reply-To: ``` Jay Soffian venit, vidit, dixit 01.04.2010 00:09: > On Wed, Mar 31, 2010 at 6:04 PM, Andreas Schwab wrote: >> Jay Soffian writes: >> >>> (Note to git list: there should be a man page that has a single-line >>> summary of each command and states whether the command is plumbing or >>> porcelain. Maybe there is and I just don't know it.) >> >> git(1) > > Yes, I'm an idiot. Anyone else wanna pile on? :-) Just blame it on April fool's day, you fooled everyone who took your remark seriously ;) Michael ```