From: Andreas Ericsson <ae@op5.se>
Date: Thu, 16 Mar 2006 11:55:55 GMT
Subject: Re: [PATCH] Invoke git-repo-config directly.
Message-ID: <4419524B.1030405@op5.se>
In-Reply-To: <7vmzfq8zmr.fsf@assigned-by-dhcp.cox.net>

Junio C Hamano wrote:
> 
> And the second issue is the last point in the "implications"
> list above.  You are right, and I stand corrected.  Our scripts
> should consistently use dash form.
> 
> One thing that bothers me is that we need to keep encouraging
> users to use dashless form from the command line, while we
> update our scripts to use dash form.  What a contradicting and
> confusing situation X-<.
> 

One of the two reasons for rewriting the git wrapper in C was the 
performance penalty that came from having it as a shell-script while it 
was desirable from a porcelainish standpoint to use the dash-less form 
since we thought even then that "git" would always be in PATH while 
"git-foo" was to be moved to the still-imaginary GIT_EXEC_PATH.

The prepending of the GIT_EXEC_PATH to PATH was a laziness workaround 
for scripts that use the dashed form until we'd had time to change 
those, although I see from the commit-message that it's a far cry from 
abundantly clear (reading it now I even think it's clear I meant the 
other way around, although I remember I didn't). Anyways, the relative 
parts are these, from commit 8e49d50388211a0f3e7286f6ee600bf7736f4814

---8<---8<---8<---
     The location of the GIT_EXEC_PATH (name discussion's closed,
     thank gods) can be obtained by running

     	git --exec-path

     which will hopefully give porcelainistas ample time to adapt their
     heavy-duty loops to call the core programs directly and thus save
     the extra fork() / execve() overhead, although that's not really
     necessary any more.

     The --exec-path value is prepended to $PATH, so the git-* programs
     should Just Work without ever requiring any changes to how they call
     other programs in the suite.

     Some timing values for 10000 invocations of git-var >&/dev/null:
     	git.sh: 24.194s
     	git.c:   9.044s
     	git-var: 7.377s

---8<---8<---8<---

 From the timing values there I think the performance issues of using 
the dash-less form can just be ignored. Very rarely will a porcelainish 
wrapper do 10000 invocations of git commands where less than 2 seconds 
will be a large percentage of the overall runtime.

-- 
Andreas Ericsson                   andreas.ericsson@op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231