git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Michael G Schwern <schwern@pobox.com>
To: Jonathan Nieder <jrnieder@gmail.com>
Cc: git@vger.kernel.org, gitster@pobox.com, robbat2@gentoo.org,
	bwalton@artsci.utoronto.ca, normalperson@yhbt.net
Subject: Re: [PATCH 2/7] Change canonicalize_url() to use the SVN 1.7 API when available.
Date: Sat, 28 Jul 2012 12:01:20 -0700	[thread overview]
Message-ID: <50143700.80900@pobox.com> (raw)
In-Reply-To: <20120728135018.GB9715@burratino>

On 2012.7.28 6:50 AM, Jonathan Nieder wrote:
>> --- a/perl/Git/SVN/Utils.pm
>> +++ b/perl/Git/SVN/Utils.pm
> [...]
>> @@ -100,6 +102,20 @@ API as a URL.
>>  =cut
>>  
>>  sub canonicalize_url {
>> +	my $url = shift;
>> +
>> +	# The 1.7 way to do it
>> +	if ( defined &SVN::_Core::svn_uri_canonicalize ) {
>> +		return SVN::_Core::svn_uri_canonicalize($url);
>> +	}
>> +	# There wasn't a 1.6 way to do it, so we do it ourself.
>> +	else {
>> +		return _canonicalize_url_ourselves($url);
>> +	}
>> +}
>> +
>> +
>> +sub _canonicalize_url_ourselves {
>>  	my ($url) = @_;
>>  	$url =~ s#^([^:]+://[^/]*/)(.*)$#$1 . canonicalize_path($2)#e;
> 
> Leaves me a bit nervous.

As it should, SVN dumped a mess on us.


> What effect should we expect this change to have?  Is our emulation
> of svn_uri_canonicalize already perfect and this change just a little
> futureproofing in case svn_uri_canonicalize gets even better, or is
> this a trap waiting to happen when new callers of canonicalize_url
> start relying on, e.g., %-encoding of special characters?

This change is *just* about sliding in the SVN API call and seeing if git-svn
still works with SVN 1.6.  It should have no effect on SVN 1.6.  These patches
are a very slow and careful refactoring doing just one thing at a time.  Every
time I tried to do too many things at once, tests broke and I had to tease the
patch apart.

At this point in the patch series the code is not ready for canonicalization.
 Until 3/8 in the next patch series, canonicalize_url() basically does nothing
on SVN 1.6 so the code has never had to deal with the problem.  3/8 deals with
improving _canonicalize_url_ourselves() to work more like
svn_uri_canonicalize() and thus "turns on" canonicalization for SVN 1.6 and
deals with the breakage.


> If I am reading Subversion r873487 correctly, in ancient times,
> svn_path_canonicalize() did the appropriate tweaking for URIs.  Today
> its implementation is comforting:
> 
> 	const char *
> 	svn_path_canonicalize(const char *path, apr_pool_t *pool)
> 	{
> 	  if (svn_path_is_url(path))
> 	    return svn_uri_canonicalize(path, pool);
> 	  else
> 	    return svn_dirent_canonicalize(path, pool);
> 	}
> 
> It might be easier to rely on that on pre-1.7 systems.

I didn't know about that.  I don't know what your SVN backwards compat
requirements are, but if that behavior goes back far enough in SVN to satisfy
you folks, then canonicalize_url() should fall back to
SVN::_Core::svn_path_canonicalize().  But try it at the end of the patch
series.  The code has to be prepared for canonicalization first.  Then how it
actually does it can be improved.


-- 
Defender of Lexical Encapsulation

  reply	other threads:[~2012-07-28 19:02 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-07-28  9:38 Canonicalize the git-svn path & url accessors Michael G. Schwern
2012-07-28  9:38 ` [PATCH 1/7] Move the canonicalization functions to Git::SVN::Utils Michael G. Schwern
2012-07-28  9:38 ` [PATCH 2/7] Change canonicalize_url() to use the SVN 1.7 API when available Michael G. Schwern
2012-07-28 13:50   ` Jonathan Nieder
2012-07-28 19:01     ` Michael G Schwern [this message]
2012-07-28 19:30       ` Jonathan Nieder
2012-07-28 19:51         ` Michael G Schwern
2012-07-28 19:57           ` Jonathan Nieder
2012-07-28 20:02             ` Jonathan Nieder
2012-07-28 20:24               ` Michael G Schwern
2012-10-14 11:42       ` [PATCH/RFC 0/2] " Jonathan Nieder
2012-10-14 11:45         ` [PATCH 1/2] git svn: do not overescape URLs (fallback case) Jonathan Nieder
2012-10-14 11:48         ` [PATCH 2/2] git svn: canonicalize_url(): use svn_path_canonicalize when available Jonathan Nieder
2012-10-23 22:58           ` Eric Wong
2012-07-28  9:38 ` [PATCH 3/7] Extract, test and enhance the logic to collapse ../foo paths Michael G. Schwern
2012-07-30 19:51   ` Eric Wong
2012-07-30 20:46     ` Michael G Schwern
2012-09-26 19:45       ` Jonathan Nieder
2012-09-26 20:58         ` Eric Wong
2012-09-26 21:38           ` Jonathan Nieder
2012-09-26 21:54             ` Eric Wong
2012-09-26 22:43               ` Jonathan Nieder
2012-09-27  0:15                 ` Eric Wong
2012-09-27  2:11                   ` Jonathan Nieder
2012-07-28  9:38 ` [PATCH 4/7] Add join_paths() to safely concatenate paths Michael G. Schwern
2012-09-26 20:51   ` Jonathan Nieder
2012-07-28  9:38 ` [PATCH 5/7] Remove irrelevant comment Michael G. Schwern
2012-07-28  9:38 ` [PATCH 6/7] Switch path canonicalization to use the SVN API Michael G. Schwern
2012-07-28 13:55   ` Jonathan Nieder
2012-07-28 19:07     ` Michael G Schwern
2012-07-30 20:04       ` Eric Wong
2012-08-02 21:51         ` Eric Wong
2012-08-02 23:18           ` Michael G Schwern
2012-10-05  7:04       ` [PATCH] git-svn: keep leading slash when canonicalizing paths (fallback case) Jonathan Nieder
2012-10-05 23:12         ` Eric Wong
2012-10-06  5:36           ` Junio C Hamano
2012-07-28  9:38 ` [PATCH 7/7] Make Git::SVN and Git::SVN::Ra canonicalize paths and urls Michael G. Schwern
2012-07-28 14:11   ` Jonathan Nieder
2012-07-28 19:15     ` Michael G Schwern

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

  List information: http://vger.kernel.org/majordomo-info.html

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

  git send-email \
    --in-reply-to=50143700.80900@pobox.com \
    --to=schwern@pobox.com \
    --cc=bwalton@artsci.utoronto.ca \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=jrnieder@gmail.com \
    --cc=normalperson@yhbt.net \
    --cc=robbat2@gentoo.org \
    /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.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).