You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Kamesh Jayachandran <ka...@collab.net> on 2007/02/14 12:14:04 UTC
[PATCH]Bogus operational logging if scheduled commit has got deletion
of file(with no lock in repo)
Hi All,
Find the attached patch and log.
With regards
Kamesh Jayachandran
Re: [PATCH]Bogus operational logging if scheduled commit has got deletion of file(with no lock in repo)
Posted by Daniel Rall <dl...@collab.net>.
On Wed, 14 Feb 2007, Kamesh Jayachandran wrote:
> Malcolm Rowe wrote:
> >On Wed, Feb 14, 2007 at 06:06:00PM +0530, Kamesh Jayachandran wrote:
> >
> >>>The patch itself looks generally good (although I've not tested it),
> >>>though it did look like it might have caused lines to go over the
> >>>80-character limit.
> >>>
> >>>
> >>It is exactly stops at 80 th character :).
> >>
> >>
> >
> >Well, HACKING does say 'Restrict lines to 79 columns', so technically
> >you probably should have reorganised it slightly to avoid hitting
> >that (probably by assigning the result of svn_path_uri_encode() to a
> >new temporary). But other code in the codebase uses the 80th column,
> >so it's no big deal, and not worth fixing now.
> >
> >
> >>I tested with and without this patch and could observe the fix.
> >>
> >>Committed at r23393.
> >>
> >>
> >
> >I guess my comment above was an approval, though I didn't really mean
> >it to be.
> >
> Sorry I misunderstood it to be.
Patch looks good.
Re: [PATCH]Bogus operational logging if scheduled commit has got
deletion of file(with no lock in repo)
Posted by Kamesh Jayachandran <ka...@collab.net>.
Malcolm Rowe wrote:
> On Wed, Feb 14, 2007 at 06:06:00PM +0530, Kamesh Jayachandran wrote:
>
>>> The patch itself looks generally good (although I've not tested it),
>>> though it did look like it might have caused lines to go over the
>>> 80-character limit.
>>>
>>>
>> It is exactly stops at 80 th character :).
>>
>>
>
> Well, HACKING does say 'Restrict lines to 79 columns', so technically
> you probably should have reorganised it slightly to avoid hitting
> that (probably by assigning the result of svn_path_uri_encode() to a
> new temporary). But other code in the codebase uses the 80th column,
> so it's no big deal, and not worth fixing now.
>
>
>> I tested with and without this patch and could observe the fix.
>>
>> Committed at r23393.
>>
>>
>
> I guess my comment above was an approval, though I didn't really mean
> it to be.
>
Sorry I misunderstood it to be.
With regards
Kamesh Jayachandran
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [PATCH]Bogus operational logging if scheduled commit has got deletion of file(with no lock in repo)
Posted by Malcolm Rowe <ma...@farside.org.uk>.
On Wed, Feb 14, 2007 at 06:06:00PM +0530, Kamesh Jayachandran wrote:
> >The patch itself looks generally good (although I've not tested it),
> >though it did look like it might have caused lines to go over the
> >80-character limit.
> >
>
> It is exactly stops at 80 th character :).
>
Well, HACKING does say 'Restrict lines to 79 columns', so technically
you probably should have reorganised it slightly to avoid hitting
that (probably by assigning the result of svn_path_uri_encode() to a
new temporary). But other code in the codebase uses the 80th column,
so it's no big deal, and not worth fixing now.
> I tested with and without this patch and could observe the fix.
>
> Committed at r23393.
>
I guess my comment above was an approval, though I didn't really mean
it to be.
Regards,
Malcolm
Re: [PATCH]Bogus operational logging if scheduled commit has got
deletion of file(with no lock in repo)
Posted by Kamesh Jayachandran <ka...@collab.net>.
> The log message could use some work. I'd find it easier to read
> something more like:
>
> [[[
> mod_dav_svn: fix the incorrect logging of commits that include deletions
> as 'unlock' operations.
>
> * subversion/mod_dav_svn/lock.c
> (remove_lock): Set SVN-ACTION only when there is a lock token to
> unlock.
> ]]]]
>
> The patch itself looks generally good (although I've not tested it),
> though it did look like it might have caused lines to go over the
> 80-character limit.
>
It is exactly stops at 80 th character :).
I tested with and without this patch and could observe the fix.
Committed at r23393.
Thanks.
With regards
Kamesh Jayachandran
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [PATCH]Bogus operational logging if scheduled commit has got deletion of file(with no lock in repo)
Posted by Malcolm Rowe <ma...@farside.org.uk>.
On Wed, Feb 14, 2007 at 05:44:04PM +0530, Kamesh Jayachandran wrote:
> [[[
> 'Commits' which has some file (without any lock) to be deleted causes bogus
> operational log like , 'unlock /path/for/which/no/lock/exists/in/repo'.
>
> * subversion/mod_dav_svn/lock.c
> (remove_lock):
> Set SVN-ACTION only if 'svn_repos_fs_unlock' is successful.
>
The log message could use some work. I'd find it easier to read
something more like:
[[[
mod_dav_svn: fix the incorrect logging of commits that include deletions
as 'unlock' operations.
* subversion/mod_dav_svn/lock.c
(remove_lock): Set SVN-ACTION only when there is a lock token to
unlock.
]]]]
The patch itself looks generally good (although I've not tested it),
though it did look like it might have caused lines to go over the
80-character limit.
Regards,
Malcolm