On Tue, Jan 22, 2008 at 08:28:25AM +0100, Alex Riesen wrote:The issue is not Windows specific and the problem can be reproduced with different versions of Git including the latest from Git master. In fact, user B does not have to made any changes, it is enough that merge was not fast-forward. In contrast with fast-forward merge, which just update the references, the recursive merge requires the working directory to perform the merge. Because directories are not trucked, there is no way to tell at the end whether an empty directory was created by user before or it became empty as result of merge. Probably, the problem can be solved by remembering the list of empty directories before performing a real merge and then, on success, to remove all empty directories that are not in that list. Here is a script that demonstrates the issue. === CUT === #!/bin/sh set -e rm -rf shared userA userB mkdir shared cd shared git init mkdir -p foo/bar for line in Files not deleted when merging after a rename; do echo $line; done > foo/bar/testfile git add foo/bar/testfile git commit -m 'add foo/bar/testfile' cd .. git clone file://$PWD/shared/.git userA git clone file://$PWD/shared/.git userB cd userA git mv foo/bar bar git commit -m 'mv foo/bar bar' git push cd .. cd userB echo 'more' >> foo/bar/testfile git commit -a -m 'edit foo/bar/testfile' git pull find * === CUT === - To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
| David Miller | Slow DOWN, please!!! |
| Greg Kroah-Hartman | [PATCH 005/196] Chinese: add translation of SubmittingDrivers |
| Andrea Arcangeli | [PATCH 01 of 11] mmu-notifier-core |
| Andrew Morton | 2.6.23-rc3-mm1 |
git: | |
| Carl Worth | Difficulties in advertising a new branch to git newbies |
| Junio C Hamano | Re: [PATCH 3/3] Teach "git branch" about --new-workdir |
| Peter Stahlir | Git as a filesystem |
| Linus Torvalds | Re: irc usage.. |
| Wolfgang Walter | Re: Kernel oops with 2.6.26, padlock and ipsec: probably problem with fpu state ch... |
| Ingo Molnar | Re: iwlwifi: fix build bug in "iwlwifi: fix LED stall" |
| David Woodhouse | Re: [bug?] tg3: Failed to load firmware "tigon/tg3_tso.bin" |
| Cedric Le Goater | Re: [PATCH net-2.6.24 0/3]: More TCP fixes |
| Richard Stallman | Real men don't attack straw men |
| Jason Dixon | Wasting our Freedom |
| bofh | Re: Code signing in OpenBSD |
| no@spam@mgedv.net | Re: HUAWEI not recognized properly (3 modem) |
| high memory | 7 hours ago | Linux kernel |
| semaphore access speed | 10 hours ago | Applications and Utilities |
| the kernel how to power off the machine | 11 hours ago | Linux kernel |
| Easter Eggs in windows XP | 14 hours ago | Windows |
| Shared swap partition | 15 hours ago | Linux general |
| Root password | 15 hours ago | Linux general |
| Where/when DNOTIFY is used? | 17 hours ago | Linux kernel |
| How to convert Linux Kernel built-in module into a loadable module | 19 hours ago | Linux kernel |
| Linux 2.6.24 and I/O schedulers | 20 hours ago | Linux kernel |
| USB Driver -- Interrupt Polling -- A Little Help Please | 1 day ago | Linux general |
