Re: [PATCH] commit: configure submodules

Junio C Hamano [off-list ref] writes:

Jens Lehmann [off-list ref] writes:

quoted

quoted

Jens, what do you think?  I see no reason for anybody other than
"submodule init" to call gitmodules_config() that reads from the
in-tree .gitmodules file.

I think the copying on init is not what we should do here because
it sets the user's customization to what ever happened to be in
.gitmodules at the time he initialized the submodule.

Hrm, why does the user have submodule.$name.$whatever customized
before saying "submodule init $whategver" for that copying to be
problematic?

quoted

So I think Orgad's change is sane and should go in.

Matching what cmd_commit() does to what cmd_status() does, i.e. grab
submodule.$name.ignore from somewhere, is not something I questioned.
The patch is a good change to make them consistent.

What I was wondering was if that is a consistently wrong thing to do
to read from .gitmodules not $GIT_DIR/config.

In any case, the log message I suggested in the review needs to be
updated in the reroll to make it clear that this is about reading
from .gitmodules, not "configuration".  AFAICS, gitmodule_config()
does not even read from $GIT_DIR/config, right?

OK.  gitmodule_config() does not read $GIT_DIR/config, but
cmd_status() and cmd_commit() call git_diff_basic_config() that is
called from git_diff_ui_config() to read submodule.$name.ignore from
it.  So Orgad's patch is _only_ about submodule.$name.ignore that is
in in-tree .gitmodule; the log message shouldn't mention "config",
as setting configuration variables work for both status and commit
just fine.

Am 24.09.2012 18:27, schrieb Junio C Hamano:

Junio C Hamano [off-list ref] writes:

quoted

In any case, the log message I suggested in the review needs to be
updated in the reroll to make it clear that this is about reading
from .gitmodules, not "configuration".  AFAICS, gitmodule_config()
does not even read from $GIT_DIR/config, right?

OK.  gitmodule_config() does not read $GIT_DIR/config, but
cmd_status() and cmd_commit() call git_diff_basic_config() that is
called from git_diff_ui_config() to read submodule.$name.ignore from
it.  So Orgad's patch is _only_ about submodule.$name.ignore that is
in in-tree .gitmodule; the log message shouldn't mention "config",
as setting configuration variables work for both status and commit
just fine.

Yes, I was just checking that call path too to make sure the user
settings from $GIT_DIR/config really override those found in
.gitmodules. And of course you are right, while the change to the
code is sane the commit message still needs some work.