Re: [PATCH v3 03/10] config: Make git_config() more flexible

On Wed, Feb 18, 2009 at 11:13 PM, Junio C Hamano [off-list ref] wrote:

Felipe Contreras [off-list ref] writes:

quoted

Currently git_config() returns an error if there is no repo config file
available (cwd is not a git repo); it will correctly parse the system
and global config files, but still return an error.

Currently it does???

Here is what I get from using the command from master and next:

   $ cd /var/tmp
   $ ls -d /.git /var/.git /var/tmp/.git /var/tmp/config
   ls: /.git: No such file or directory
   ls: /var/.git: No such file or directory
   ls: /var/tmp/.git: No such file or directory
   ls: /var/tmp/config: No such file or directory
   $ git config -l >/dev/null ; echo $?
   0
   $ git config alias.co; echo $?
   checkout
   0

This patch doesn't alter the behavior of 'git config', it's for
git_config() (the function). At this point cmd_config() will return an
exit code of 0 even if there's an error on git_config() when no
exclusive file is specified.

I have $HOME/.gitconfig (where the alias comes from) and no system wide
configuration file.

Also the patch is mistitled.  Whatever you are trying to say about the
current problem which I do not seem to get, and whatever different
behaviour from the current one you are trying to implement (which is not
quite clear from the above log message), it is not about making it more
flexible.

The patch text suggests you are trying to change the function's exit
status so the title would be "git-config: report errors correctly in its
exit status", but it is unspecified in your commit log message what
definition of "correctly" you are using in this patch.

What I'm trying to do is "Make git_config() don't return an error when
there's no repo config file if there are other config files
available"... I thought there was no short way to say that =/

-- 
Felipe Contreras