Stefan Beller <sbeller@xxxxxxxxxx> writes: > On Wed, Mar 30, 2016 at 9:05 AM, Stefan Beller <sbeller@xxxxxxxxxx> wrote: >> On Wed, Mar 30, 2016 at 2:03 AM, Norio Nomura <norio.nomura@xxxxxxxxx> wrote: >>> Hi, >>> >>> `git submodule update --init --recursive` stores `gitdir` in full path into `.git` of nested submodules. >>> So, working directory is not portable to another directory. >> >> Are you reporting a regression bug? (Is that a new thing or has it >> always been that way and you just discover that it is unfortunate?) >> Which versions did you test with? > > ➜ 15:34:32 git:(master) git --version > git version 2.8.0 > > at the end of your gist. > The same happens when using 2.7.4, it doesn't happen when using 2.6.6 though. > > It turns out ee8838d (2015-09-08, submodule: rewrite `module_clone` > shell function in C) > broke it. > > I'll look into fixing it. Thanks. This may be an unrelated tangent, but somewhere in "submodule init" codepath seem to turn paths into absolute too aggressively. Merging your recent "path related fix" to 'pu' seems to break some test by showing absolute paths when the test expects to see relative ones. -- 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