Hi Junio, On Mon, 25 Jan 2016, Junio C Hamano wrote: > Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes: > > >> Is the fact that $global_excludes is specified using an absolute > >> path significant to the correctness of this test script? > > > > Apparently not. So I followed your suggestion and made this independent > > of the OS: > > > > https://github.com/dscho/git/commit/0b9eb308 > > I see many "cd" in the script, some of which come before the global > exclude file is enabled, but others ("--stdin from subdirectory", > for example) do run from a subdirectory. It turns out that this > file is used as the value of core.excludesfile. By the time that > file is read in setup_standard_excludes(), presumably we have > already done setup_git_directory() and moved back to the root of the > working tree, so I guess such a change would be harmless. Yeah, it appears that after the entire test ran, there is only one global-excludes file, and none of the `rm` commands removes any global-excludes file. > It somehow makes me feel dirty, though. But that is not a specific > problem to mingw port. Well, I found the entire file a little hard to read ;-) But yeah, I do not think that is specific to the Windows port. Ciao, Dscho -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html