headers
David D. Kilzer | 1 Feb 2006 07:23

Removing keywords from resolved bugs

Is there a reason for removing keywords from Resolved bugs?  Is it  
just so that the Keywords page on Bugzilla will reflect an "accurate"  
count of unresolved bugs with that keyword, and to make it easier to  
search for unresolved bugs with that keyword?

In the example below, the comment talked about removing the  
"Regression" keyword alone, but ended up wiping out every keyword in  
that bug.

Could we modify the Keywords page to list and query for unresolved  
bugs?  Perhaps provide two columns for each keyword: one of  
unresolved bugs and one of total bugs?

I'd be happy to provide a patch to Bugzilla to do that if it would  
stop this keyword removal.  Note that others have expressed a  
displeasure in this removal on #webkit.

Dave

Begin forwarded message:

> From: bugzilla-daemon <at> opendarwin.org
> Date: January 31, 2006 11:20:55 PM CST
> To: ddkilzer <at> kilzer.net
> Subject: [Bug 6550] REGRESSION: window.open('myurl', '_self') fails  
> to load myurl in current window
>
> http://bugzilla.opendarwin.org/show_bug.cgi?id=6550
>
>
(Continue reading)

Permalink | Reply | 6 comments |
headers
Eric Seidel | 1 Feb 2006 07:56
Picon
Favicon

Re: Removing keywords from resolved bugs

I think that in many cases ("Regression" and "HasReduction" being  
two) it doesn't make much sense for resolved bugs to carry keywords  
("Regression" from what?).

That said, I had intended to only remove the "Regression" keyword to  
clean up the "known regressions list", and mistakenly removed all of  
them.

I'm not sure what the "official" policy on keywords-attached-to- 
resolved bugs is.  My impression was that we generally removed them  
after resolution.  Perhaps it's something which needs to be discussed  
more generally?  Do you have a suggested policy?

I'm not currently on #webkit (not feeling particularly well), but  
would be happy to discuss this with you or others at a later time.

-eric

On Jan 31, 2006, at 10:23 PM, David D. Kilzer wrote:

> Is there a reason for removing keywords from Resolved bugs?  Is it  
> just so that the Keywords page on Bugzilla will reflect an  
> "accurate" count of unresolved bugs with that keyword, and to make  
> it easier to search for unresolved bugs with that keyword?
>
> In the example below, the comment talked about removing the  
> "Regression" keyword alone, but ended up wiping out every keyword  
> in that bug.
>
> Could we modify the Keywords page to list and query for unresolved
(Continue reading)

Permalink | Reply | 4 comments |
headers
Joost de Valk | 1 Feb 2006 08:36
Picon

Re: Removing keywords from resolved bugs

My vote would be for adapting Bugzilla. Though I tend to agree with Eric 
and have removed keywords from bugs before, David Kilzer's idea seems 
easier to implement and less work to maintain :) . As i am also an 
Opendarwin admin, i'll be happy to look into it :)

Joost

Eric Seidel wrote:
> I think that in many cases ("Regression" and "HasReduction" being two) 
> it doesn't make much sense for resolved bugs to carry keywords 
> ("Regression" from what?).
>
> That said, I had intended to only remove the "Regression" keyword to 
> clean up the "known regressions list", and mistakenly removed all of 
> them.
>
> I'm not sure what the "official" policy on 
> keywords-attached-to-resolved bugs is.  My impression was that we 
> generally removed them after resolution.  Perhaps it's something which 
> needs to be discussed more generally?  Do you have a suggested policy?
>
> I'm not currently on #webkit (not feeling particularly well), but 
> would be happy to discuss this with you or others at a later time.
>
> -eric
>
> On Jan 31, 2006, at 10:23 PM, David D. Kilzer wrote:
>
>> Is there a reason for removing keywords from Resolved bugs?  Is it 
>> just so that the Keywords page on Bugzilla will reflect an "accurate"
(Continue reading)

Permalink | Reply | 0 comments |
headers
Maciej Stachowiak | 1 Feb 2006 09:03
Picon
Favicon

Re: Removing keywords from resolved bugs


On Jan 31, 2006, at 10:56 PM, Eric Seidel wrote:

> I think that in many cases ("Regression" and "HasReduction" being  
> two) it doesn't make much sense for resolved bugs to carry keywords  
> ("Regression" from what?).

It is useful to know that a fix was for a regression.

> That said, I had intended to only remove the "Regression" keyword  
> to clean up the "known regressions list", and mistakenly removed  
> all of them.
>
> I'm not sure what the "official" policy on keywords-attached-to- 
> resolved bugs is.  My impression was that we generally removed them  
> after resolution.  Perhaps it's something which needs to be  
> discussed more generally?  Do you have a suggested policy?

I don't think we should remove any keywords on resolution, unless  
whatever they say about the bug is no longer true. A fixed regression  
was still a regression, a bug that had a recution still does. Bugs  
get reopened sometimes if they are not verified as fix, or if they  
pop up again later, it is valuable to save all the information about  
them.

Regards,
Maciej
Permalink | Reply | 1 comment |
headers
Trey Matteson | 1 Feb 2006 18:38
Picon

Re: Removing keywords from resolved bugs

I totally agree.  This idea of removing keywords on resolution makes  
no sense to me (and is a loss of useful info).  By that logic, why  
not delete the whole record from the DB since they are no longer bugs  
in the code?

trey

On Feb 1, 2006, at 12:03 AM, Maciej Stachowiak wrote:

>
> On Jan 31, 2006, at 10:56 PM, Eric Seidel wrote:
>
>> I think that in many cases ("Regression" and "HasReduction" being  
>> two) it doesn't make much sense for resolved bugs to carry  
>> keywords ("Regression" from what?).
>
> It is useful to know that a fix was for a regression.
>
>> That said, I had intended to only remove the "Regression" keyword  
>> to clean up the "known regressions list", and mistakenly removed  
>> all of them.
>>
>> I'm not sure what the "official" policy on keywords-attached-to- 
>> resolved bugs is.  My impression was that we generally removed  
>> them after resolution.  Perhaps it's something which needs to be  
>> discussed more generally?  Do you have a suggested policy?
>
> I don't think we should remove any keywords on resolution, unless  
> whatever they say about the bug is no longer true. A fixed  
> regression was still a regression, a bug that had a recution still
(Continue reading)

Permalink | Reply | 0 comments |
headers
bradley.morrison | 1 Feb 2006 19:55
Picon

RE: SVN over HTTP


Hello,

I'm looking at providing a number of developers access to the
subversion repository from behind a firewall. WebSVN unfortunately
only provides a read-only view into the repository and not full
checkout access.

Is it possible for HTTP proxy support to be turned on, ideally
alongside WebSVN but maybe even replacing it?

Many thanks
Bradley

From: webkit-dev-bounces <at> opendarwin.org
[mailto:webkit-dev-bounces <at> opendarwin.org] On Behalf Of ext David D.
Kilzer

> SVN over HTTP is not set up on that server.  It looks like 
> something called WebSVN is set up instead:
>  http://anoncvs.opensource.apple.com/svn/
> It would be nice if the repository was accessible via HTTP
> as well, though.

On Tue, Jan 10, 2006 at 06:11:40PM -0500, David.Carson <at> nokia.com wrote:

> Hi,
> Those of us behind a firewall, I thought that I would have a go at 
> configuring subversion to operate over a proxy. This would save us 
> from having to go to public WLAN areas to download the source.
(Continue reading)

Permalink | Reply | 1 comment |
headers
David Hyatt | 1 Feb 2006 22:32
Picon
Favicon

Re: Removing keywords from resolved bugs

I don't see any reason to remove the regression keyword after a bug  
is resolved and don't really understand the logic behind suddenly  
doing this to all the regression bugs.

It's still useful even after a bug is fixed to know that it was a  
regression from an earlier build.  What if the bug isn't really fixed  
and has to be reopened?

If the problem is search querying, just refine the query to include  
only bugs that are still open.

dave

On Jan 31, 2006, at 10:56 PM, Eric Seidel wrote:

> I think that in many cases ("Regression" and "HasReduction" being  
> two) it doesn't make much sense for resolved bugs to carry keywords  
> ("Regression" from what?).
>
> That said, I had intended to only remove the "Regression" keyword  
> to clean up the "known regressions list", and mistakenly removed  
> all of them.
>
> I'm not sure what the "official" policy on keywords-attached-to- 
> resolved bugs is.  My impression was that we generally removed them  
> after resolution.  Perhaps it's something which needs to be  
> discussed more generally?  Do you have a suggested policy?
>
> I'm not currently on #webkit (not feeling particularly well), but  
> would be happy to discuss this with you or others at a later time.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Darin Adler | 1 Feb 2006 22:33
Picon
Favicon

Re: SVN over HTTP

On Feb 1, 2006, at 10:55 AM, bradley.morrison <at> nokia.com wrote:

> I'm looking at providing a number of developers access to the  
> subversion repository from behind a firewall. WebSVN unfortunately  
> only provides a read-only view into the repository and not full  
> checkout access.
>
> Is it possible for HTTP proxy support to be turned on, ideally  
> alongside WebSVN but maybe even replacing it?

We'd really like to do that, but there are some obstacles. The  
svn.opensource.apple.com server is a Mac OS X machine onsite at  
Apple. Here are some of the issues I discovered today investigating  
this:

     - Subversion-over-HTTP requires Apache 2.0, which isn't  
officially supported on Mac OS X.
     - The version of Subversion on that server, built by a group  
here at Apple, does not include Subversion-over-HTTP support.
     - Some experts here at Apple have heard reports that Subversion- 
over-HTTP is unreliable on Tiger, aborting during large commits,  
possibly due to problems with poll(2) and similar issues.

At the moment, the folks running the server don't have the time to  
maintain what would be an experimental configuration.

Maybe we can explore some other options?

     -- Darin
(Continue reading)

Permalink | Reply | 0 comments |
headers
Mark Rowe | 1 Feb 2006 23:27
Picon

Re: SVN over HTTP

Hi all,

A little while back I set up a Trac instance at <http:// 
trac.webkit.org/> to provide a web-based view of the SVN repository.   
One requirement of Trac is local access to the SVN repository, so I  
have set up a mirror of anonsvn using SVN::Mirror.  The mirrored  
repository is available over HTTP at <http://trac.webkit.org/svn/>  
with one caveat: the mirrored repository has revision numbers one  
higher than the corresponding anonsvn revisions.  This is an artifact  
of the mirroring process that I've not been able to hide.

Hopefully this helps,

Mark

On 11/01/2006, at 12:30, David D. Kilzer wrote:

> You don't have anything configured wrong.  SVN over HTTP is not set up
> on that server.  It looks like something called WebSVN is set up
> instead:
>
>   http://anoncvs.opensource.apple.com/svn/
>
> It would be nice if the repository was accessible via HTTP as well,
> though.
>
> Dave
>
>
> On Tue, Jan 10, 2006 at 06:11:40PM -0500, David.Carson <at> nokia.com
(Continue reading)

Permalink | Reply | 0 comments |
headers
Timothy Hatcher | 4 Feb 2006 00:04

Project Changes (rename to Debug and Release)

I will be landing a patch today that changes the configuration names to: Debug, Release and Production (from Development, Deployment and Default). All scripts and makefiles will also change to take into account the new configuration names. If you have any shell profiles that set up the environment referencing the old names you will need to change them. Let me know if you can't get something to work after this change.


— Timothy Hatcher

 

_______________________________________________
webkit-dev mailing list
webkit-dev <at> opendarwin.org
http://www.opendarwin.org/mailman/listinfo/webkit-dev
Permalink | Reply | 0 comments |

Gmane