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

Re: 1.5.0-rc6 up for signing/testing

From: Hyrum K. Wright <hyrum_wright_at_mail.utexas.edu>
Date: Sat, 24 May 2008 19:05:23 -0700

Mark Phippard wrote:
> On Sat, May 24, 2008 at 7:59 PM, Mark Phippard <markphip_at_gmail.com> wrote:
>> Well, the tests were all looking good for me. All the bindings tests
>> passed and I had gotten through all the ra_local. Earlier I had run
>> serf and neon. But then on my Windows box I accidentally stumbled on
>> a bug I created in JavaHL. I am creating a fix now that I will
>> propose for backport.
>>
>> There is another thread about that.
>>
>> Quick summary: We made a change to JavaHL last week to synch it with
>> other API changes. When you run svn log in JavaHL, you now get back
>> all of the revprops in a Map. Where as we used to get author, date
>> and message as explicit fields. Anyway, we added a class to parse the
>> svn:date string into a Java date. The bug was that the parsing code
>> was assuming the date was in the current timezone, and I believe it is
>> actually in UTC timezone. No one has actually confirmed that for me,
>> and if it is not true, then only the test that is failing is wrong. I
>> am going to commit code that assumes the date is in UTC as that is
>> what it looks like when I am testing it.
>
> I committed and nominated this for backport. Once I used the log API
> from Subclipse this bug also became very obvious. All the Subversion
> commit activity had a time that was 4 hours into the future. With
> this patch applied the dates are right.
>
> Sorry to screw up the RC6, but hopefully not many people had started testing it.

I'll review the change, and pending a favorable review, roll RC 7 tonight.

-Hyrum

Received on 2008-05-25 04:05:44 CEST

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