{"thread":{"id":"13549","subject":"partial globbing in fetch refspecs broken in v1.5.5","startedAt":"2008-05-16T21:28:34Z","lastAt":"2008-05-17T12:23:05Z","messageCount":3,"participants":["Nikolaus Schulz","Jeff King"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"77159","messageId":"20080516212833.GA5366@penelope.zusammrottung.local","threadId":"13549","inReplyTo":null,"subject":"partial globbing in fetch refspecs broken in v1.5.5","fromName":"Nikolaus Schulz","fromEmail":"microschulz@web.de","sentAt":"2008-05-16T21:28:34Z","receivedAt":"2008-05-16T21:28:34Z","isPatch":false,"sender":{"key":"microschulz@web.de","avatar":null},"body":"Hi, \n\nthe new refspec parser in v1.5.5 has also broken fetch uses like\n\n    git fetch <url> +refs/heads/<foo>*:refs/remotes/<bar>*. \n\nSuch a refspec works like a charm with git 1.5.4.  In case the usefulness of\nsuch a refspec isn't obvious, here's my use case.  I want to track several\nremotes, and in particular a selection of branches from one remote repository,\nwhich has the following branches[1]: \n\n  HEAD\n  candidate\n  candidate+patches\n  debian-experimental\n  debian-sarge\n  debian-sid\n  maint\n  master\n  release\n  release+patches\n\nWith git 1.5.4, I can add the following to .git/config\n\n[remote \"debian\"]\n        url = http://smarden.org/git/git.git\n        fetch = +refs/heads/debian-*:refs/remotes/debian/*\n\n...and it will work just like '%'-style pattern matching in Makefiles, e.g.\nthe remote branch debian-sid becomes remotes/debian/sid.  I find that very\nnatural; but git v1.5.5 rejects this refspec as invalid.  \n\nI am just a beginner with git, and I do not know if there is any problem with\nsuch a use of refspecs.  But if not, I would like to beg the maintainers to\nconsider restoring this functionality.\n\nNikolaus\n"},{"id":"77177","messageId":"20080517052212.GC18876@sigill.intra.peff.net","threadId":"13549","inReplyTo":"20080516212833.GA5366@penelope.zusammrottung.local","subject":"Re: partial globbing in fetch refspecs broken in v1.5.5","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2008-05-17T05:22:12Z","receivedAt":"2008-05-17T05:22:12Z","isPatch":false,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Fri, May 16, 2008 at 11:28:34PM +0200, Nikolaus Schulz wrote:\n\n> the new refspec parser in v1.5.5 has also broken fetch uses like\n> \n>     git fetch <url> +refs/heads/<foo>*:refs/remotes/<bar>*. \n\nThis was intentionally changed in ef00d150 (Tighten refspec processing).\nThere is some associated discussion in this thread:\n\n  http://mid.gmane.org/7vwso2ieuu.fsf@gitster.siamese.dyndns.org\n\nBut I don't see any particular reason why this syntax should not be\nallowed (besides, \"it was never meant to be supported, so it was causing\nother breakage\").\n\n-Peff\n"},{"id":"77188","messageId":"20080517122304.GA3745@penelope.zusammrottung.local","threadId":"13549","inReplyTo":"20080517052212.GC18876@sigill.intra.peff.net","subject":"Re: partial globbing in fetch refspecs broken in v1.5.5","fromName":"Nikolaus Schulz","fromEmail":"microschulz@web.de","sentAt":"2008-05-17T12:23:05Z","receivedAt":"2008-05-17T12:23:05Z","isPatch":false,"sender":{"key":"microschulz@web.de","avatar":null},"body":"On Sat, May 17, 2008 at 01:22:12AM -0400, Jeff King wrote:\n> On Fri, May 16, 2008 at 11:28:34PM +0200, Nikolaus Schulz wrote:\n> \n> > the new refspec parser in v1.5.5 has also broken fetch uses like\n> > \n> >     git fetch <url> +refs/heads/<foo>*:refs/remotes/<bar>*. \n> \n> This was intentionally changed in ef00d150 (Tighten refspec processing).\n> There is some associated discussion in this thread:\n> \n>   http://mid.gmane.org/7vwso2ieuu.fsf@gitster.siamese.dyndns.org\n\nAh.  I had found that commit, but not this discussion, thanks. \nI don't understand what the problem with packing the OP ran into with\nhis setup, though (see [1]).\n\n> But I don't see any particular reason why this syntax should not be\n> allowed (besides, \"it was never meant to be supported, so it was causing\n> other breakage\").\n\nWhich breakage?\n\nNikolaus\n\n[1] http://article.gmane.org/gmane.comp.version-control.git/77381\n"}]}