git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCHv2 maint] git-svn: Fix git svn log --show-commit

From
Michael J Gruber <git@drmicha.warpmail.net>
Date
May 21, 2011, 09:30 UTC
Message-ID
<4DD7863D.2010002@drmicha.warpmail.net>
In-Reply-To
<7vk4dl4a9f.fsf@alter.siamese.dyndns.org>
Junio C Hamano venit, vidit, dixit 20.05.2011 18:17:
Show 17 quoted lines
> Michael J Gruber <git@drmicha.warpmail.net> writes:
> 
>> git svn log --show-commit had no tests and, consequently, no attention
>> by the author of
>>
>> b1b4755 (git-log: put space after commit mark, 2011-03-10)
>>
>> who kept git svn log working only without --show-commit.
>>
>> Introduce a test and fix it.
>>
>> Reported-by: Bernt Hansen <bernt@norang.ca>
>> Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
>> ---
>> git svn scares me.
> 
> Sorry about this breakage. 

I have to be sorry, I broke it! Which is why I took responsibility to fix it asap.

Show 26 quoted lines
>> diff --git a/git-svn.perl b/git-svn.perl
>> index a5857c1..0cee0e9 100755
>> --- a/git-svn.perl
>> +++ b/git-svn.perl
>> @@ -5735,7 +5735,7 @@ sub cmd_show_log {
>>  	my $esc_color = qr/(?:\033\[(?:(?:\d+;)*\d*)?m)*/;
>>  	while (<$log>) {
>>  		if (/^${esc_color}commit (- )?($::sha1_short)/o) {
>> -			my $cmt = $1;
>> +			my $cmt = $2;
> 
> Even more defensive approach would be not to grab the grouping by doing:
> 
> -  		if (/^${esc_color}commit (- )?($::sha1_short)/o) {
> +  		if (/^${esc_color}commit (?:- )?($::sha1_short)/o) {
> 
> and not to change anything else. I should have noticed the $1 reference
> that was immediately on the next line when I saw and applied your patch,
> but if there were more references in the scope that is outside of the
> patch context, the same bug would be likely to have gone unnoticed.
> 
> I do not have enough bandwidth to read every single line of the patch from
> everybody, so small bugs in patches from known to be good people (you
> included) can slip through, unless marked with "I am not familiar with
> this codepath" or "I am not strong in Perl regexp" or somesuch, in which
> case I try to allocate more time to give it another pass of eyeballing.

So, for the record: I don't grok perl. I don't grok perl regexps I don't like either ;)

I use perl when I have to (fixing others' cgi, hacking git-svn) and try my best, or ask for help. Here, three things came together: I did not really comprehend the full regexp with the ($::thingy), probably because /foo bar/ looks separated to my eyes, but was not aware of that lack of understanding; there was no test for the codepath which used $cmt (which ends up being used as $foo->c somewhere else); I didn't even know about "--show-commit" (and that codepath).

> In any case, thanks for the fix. I think being defensive with (?:) would
> be a better idea, so I'll tweak the patch before applying with your test.
Inserting a few more "?" or even "???" definitely describes the issue well..

Turns out that "non-capturing groups" are a regexp feature which is not even specific to perl.

So, thanks Bernt for teaching me a new "git svn log" option and Junio and Andreas for a new regexp feature ;)

Michael
Previous: Junio C Hamano
Message 6 of 6 in “git-svn: Fix git svn log --show-commit”
  1. git-svn: Fix git svn log --show-commitMichael J Gruber, May 20, 2011
  2. Andreas SchwabMay 20, 2011
  3. [PATCHv2 maint] git-svn: Fix git svn log --show-commitMichael J Gruber, May 20, 2011
  4. [PATCHv2 maint] git-svn: Fix git svn log --show-commitBernt Hansen, May 20, 2011
  5. Junio C HamanoMay 20, 2011
  6. Michael J GruberMay 21, 2011

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.