All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jakub Narebski <jnareb@gmail.com>
To: Jari Aalto <jari.aalto@cante.net>
Cc: git@vger.kernel.org
Subject: Re: [PATCH] config.c: Expand $HOME and tilde character in core.excludesfile
Date: Mon, 28 Jan 2008 14:32:51 -0800 (PST)	[thread overview]
Message-ID: <m3y7a98ttg.fsf@localhost.localdomain> (raw)
In-Reply-To: <y7a9aaem.fsf@blue.sea.net>

Jari Aalto <jari.aalto@cante.net> writes:

> * str_replace(): New function. Generic replace command.
> * str_replace_home(): New function. Substitute $HOME and tilde(~) in string.
> * git_default_config(): Pass core.excludesfile to str_replace_home().
> 
> Signed-off-by: Jari Aalto <jari.aalto AT cante.net>

First, git project does NOT use GNU ChangeLog convention for it's
commit messages.  We rather use descriptive commit messages.

Second, I'm not sure about str_replace... how it fits with strbufs?
AFAIK we try to use strbufs whenever possible and feasible, to avoid
errors in git.

Third, I agree that it is a good idea, but I'd rather have *full*
solution, i.e. for git to expand $HOME (or better yet any
environmental variable) and '~' everywhere, not only for
core.excludesfile, but also for --git-dir and GIT_DIR, for
core.worktree and --work-tree and GIT_WORK_TREE, and for all other
config variables and enviroment variables.

> ---
> 
>  From ac6941f5055b2acdded59627d228bbf03ba0d9fc

What does it mean? A bit cryptic, don't you think?


Comments on code below. One thing: we use tabs for indent, not
spaces.  You use spaces in your code, while context uses tabs.

> +char *str_replace(const char *str, const char *find, const char *replace)
> +{
> +        int maxlen   = strlen(str) + strlen(replace) + 1;
> +        char *start  = strstr(str, find);
> +        char *ptr    = (char *)malloc(maxlen);
> +        int len      = strlen(find);
> +        int llen, rlen;         /* left, right portion length */
> +
> +        if (start == (char *)NULL) {

There is no need to cast NULL. Besides, we write IIRC "if (!start)",
it is common enough idiom.

> +                strcpy( ptr, str);

Style: no space after opening parenthesis: "strcpy(ptr, str)".
Performance: I think it would be better to use stpcpy, although I'm
not quite sure if here too.

> +        }
> +        else{
> +                rlen = strlen(start) - strlen(find);
> +                llen = strlen(str) - strlen(start);
> +                strncpy( ptr, str, llen);
> +                strcat( ptr, replace);
> +                strncat( ptr, start + len, rlen); /* Does not add  '\0' */
> +                strcat( ptr, "");   /* Terminate with null string */

Performance suck here; although this is not time-critical path using
sequence of strcat is just bad style. And 'strcat( ptr, "");' just
takes the cake: use "ptr[len] = '\0'", ehere 'len' is calculated
length of string.

P.S. Junio, when do you think 1.5.4 would be finally released? We have
feature freeze still, isn't it?

-- 
Jakub Narebski
Poland
ShadeHawk on #git

  parent reply	other threads:[~2008-01-28 22:33 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-01-28 21:49 [PATCH] config.c: Expand $HOME and tilde character in core.excludesfile Jari Aalto
2008-01-28 22:28 ` Johannes Schindelin
2008-01-28 22:32 ` Jakub Narebski [this message]
2008-01-29  5:59   ` Miles Bader
2008-01-29  7:25     ` David Symonds
2008-01-29  7:51       ` Miles Bader
2008-01-28 23:52 ` Wayne Davison

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=m3y7a98ttg.fsf@localhost.localdomain \
    --to=jnareb@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=jari.aalto@cante.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.