[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: r931209 - /subversion/trunk/subversion/libsvn_ra/ra_loader.c

From: C. Michael Pilato <cmpilato_at_collab.net>
Date: Wed, 07 Apr 2010 09:25:02 -0400

Bert Huijben wrote:
>
>> -----Original Message-----
>> From: cmpilato_at_apache.org [mailto:cmpilato_at_apache.org]
>> Sent: dinsdag 6 april 2010 18:25
>> To: commits_at_subversion.apache.org
>> Subject: svn commit: r931209 -
>> /subversion/trunk/subversion/libsvn_ra/ra_loader.c
>>
>> Author: cmpilato
>> Date: Tue Apr 6 16:25:21 2010
>> New Revision: 931209
>>
>> URL: http://svn.apache.org/viewvc?rev=931209&view=rev
>> Log:
>> Add compat code to work around pre-1.7 servers' messy handling of
>> mergeinfo paths.
>>
>> * subversion/libsvn_ra/ra_loader.c
>> (svn_ra_get_mergeinfo): Strip leading slashes from catalog keys
>> returned by the RA provider.
>
> Shouldn't we handle this in the specific RA layers instead of globally for all ra-layers?
>
> That would allow removing the code for future code paths (like we did for HTTPv2), while this code would hide the issues if we ever reintroduce the bug?
>
> I'm pretty sure you fixed ra_local, so that is at least one ra layer that doesn't need this fix.

So, I made this change with 1.6.x in mind. 1.6.x benefits universally from
this change because a) there's only codepath to debug if something is wrong,
b) 1.6.x code will never be able to optimize out the correction (because we
won't likely teach it to recognize newer servers), and c) I'm not proposing
the server fix for 1.6.x.

But sure, I can move the corrective behavior into specific RA layers for trunk.

-- 
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Received on 2010-04-07 15:25:44 CEST

This is an archived mail posted to the Subversion Dev mailing list.