Re: [PATCH 8/9] Use the early config machinery to expand aliases

> Hi Brandon,
> 
> On Wed, 7 Jun 2017, Brandon Williams wrote:
> 
> > I like seeing chunks of old code being deleted :D
> 
> Me, too. In particular as hacky code as this one. It caused quite a couple
> of gray hairs here.
> 
> > On 06/07, Johannes Schindelin wrote:
> > 
> > > @@ -245,36 +201,37 @@ static int handle_options(const char ***argv, int *argc, int *envchanged)
> > >  
> > >  static int handle_alias(int *argcp, const char ***argv)
> > >  {
> > > +	struct strbuf worktree_dir = STRBUF_INIT;
> > >  	int envchanged = 0, ret = 0, saved_errno = errno;
> > >  	int count, option_count;
> > >  	const char **new_argv;
> > >  	const char *alias_command;
> > >  	char *alias_string;
> > > -	int unused_nongit;
> > > -
> > > -	save_env_before_alias();
> > > -	setup_git_directory_gently(&unused_nongit);
> > >  
> > >  	alias_command = (*argv)[0];
> > > -	alias_string = alias_lookup(alias_command, NULL);
> > > +	alias_string = alias_lookup(alias_command, &worktree_dir);
> > >  	if (alias_string) {
> > >  		if (alias_string[0] == '!') {
> > >  			struct child_process child = CHILD_PROCESS_INIT;
> > >  
> > > +			if (worktree_dir.len)
> > > +				setup_git_directory();
> > 
> > So if there is a worktree then we run the setup code explicitly.  I
> > assume that is to ensure that all envvars are setup properly before
> > running the alias.  Just interesting to note that the actual value of
> > worktree_dir is never used, its just used essentially as a boolean
> > indicator of there being a worktree/gitdir/repository.  I'm not
> > suggesting to change from what you have here, its just food for thought.
> 
> Indeed.
> 
> That is what I tried to indicate by this paragraph in the commit message:
> 
> 	Rather than just chdir()ing into the indicated directory in case of an
> 	alias expanding to a shell command, we simply set up the .git/ directory
> 	so that e.g. GIT_PREFIX is set as expected.