Bug 273783 - Broken again :/
Summary: Broken again :/
Status: RESOLVED FIXED
Alias: None
Product: kio
Classification: Frameworks and Libraries
Component: general (show other bugs)
Version: unspecified
Platform: Compiled Sources Linux
: NOR normal
Target Milestone: ---
Assignee: David Faure
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-05-21 13:30 UTC by Christoph Feck
Modified: 2011-06-18 20:55 UTC (History)
4 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Christoph Feck 2011-05-21 13:30:01 UTC
Version:           unspecified (using Devel) 
OS:                Linux

It is a bit hard to track which commit adds or reverts which regression, but with todays' kdelibs master at 691ae040 I see the following behavior in Konqueror, which already occurred some weeks ago, but got fixed.

Reproducible: Always

Steps to Reproduce:
* go to a bug report
* add a comment


Actual Results:  
It says that someone (=me) already added a comment.

Expected Results:  
My comment is added without that warning.

It easily causes duplicate comments, as seen on bug 269881.
Comment 1 Christoph Feck 2011-05-21 14:54:41 UTC
More regressions seen, most likely introduced by the same commit(s):
- I cannot reassign bugs, no reply
- funny "Please stand by..." message when using https://bugs.kde.org/buglist.cgi?emailreporter1=1&emailtype1=exact&email1=adawit%40kde.org
- Often says "Waiting for reply ..." in status bar, but page is loaded (Stop button disabled).
- Konqueror process doesn't exit when quitting
If any of them is unrelated, I can report them separately.
Comment 2 Anne-Marie Mahfouf 2011-05-21 15:11:03 UTC
Trying to assess the bug report
Comment 3 Anne-Marie Mahfouf 2011-05-21 15:12:53 UTC
OK so I reproduce Christoph described steps.
I edited several bug reports yesterday and this morning before updating master a few hours ago so this regression is very recent.
Comment 4 Anne-Marie Mahfouf 2011-05-21 15:13:03 UTC
OK so I reproduce Christoph described steps.
I edited several bug reports yesterday and this morning before updating master a few hours ago so this regression is very recent.
Comment 5 Dawit Alemayehu 2011-05-21 15:48:44 UTC
I presume this is using konqueror + khtml,  correct ? If so, can anyone of you try reverting commit dffd5950b35ead8c14c5cf3a2495be5f8b7b9a1f in kdelibs and see if the problem goes away ?
Comment 6 Dawit Alemayehu 2011-05-21 15:50:02 UTC
Testing Posting with khtml...
Comment 7 Dawit Alemayehu 2011-05-21 15:53:40 UTC
Well I cannot reproduce this bug. I just added a bogus comment #6 with konqueror + khtml from master and I did not get any warning at all. I am also adding this comment using konqueror + khtml.
Comment 8 Christoph Feck 2011-05-21 16:01:56 UTC
Testing again after restarting KDE with recent master of kdelibs and kdebase to rule out version change (although I did restart kdeinit).
Comment 9 Christoph Feck 2011-05-21 16:07:43 UTC
Nope, it's reproducible. It still gives the warning.

Konqueror/KHML, kdelibs/kdebase recent, KDE restarted.

I will try reverting dffd5950.
Comment 10 Dawit Alemayehu 2011-05-21 16:11:00 UTC
(In reply to comment #1)
> More regressions seen, most likely introduced by the same commit(s):
> - I cannot reassign bugs, no reply
> - funny "Please stand by..." message when using
> https://bugs.kde.org/buglist.cgi?emailreporter1=1&emailtype1=exact&email1=adawit%40kde.org
> - Often says "Waiting for reply ..." in status bar, but page is loaded (Stop
> button disabled).
> - Konqueror process doesn't exit when quitting
> If any of them is unrelated, I can report them separately.

hmm on second though neither one of the two commits I mentioned above can cause all of these issues. Have you tried restarting kdeinit or restarting kde (log out and back in) after installing kdelibs ?
Comment 11 Christoph Feck 2011-05-21 16:15:47 UTC
Testing with dffd5950 reverted.

Dawit, yes, for comment #9 I restarted the complete KDE session, even making sure with "ps ax" and "zypper ps" that no process from KDE is left.
Comment 12 Christoph Feck 2011-05-21 16:16:33 UTC
I can confirm that reverting dffd5950 fixes at least the issue from comment #0. Will check the other issues.
Comment 13 Christoph Feck 2011-05-21 16:22:42 UTC
I only tested quickly, but after reverting dffd5950 all issues mentioned in comment #1 are no longer visible. I was able to reassign a bug, no funny "Please stand by ..." page, and process finished correctly.
Comment 14 Andrea Iacovitti 2011-05-21 16:24:20 UTC
(In reply to comment #1)
> - funny "Please stand by..." message when using
> https://bugs.kde.org/buglist.cgi?emailreporter1=1&emailtype1=exact&email1=adawit%40kde.org
> - Often says "Waiting for reply ..." in status bar, but page is loaded (Stop
> button disabled).

^^^^
Same issue was reported in bug #212043
Comment 15 Dawit Alemayehu 2011-05-21 16:31:00 UTC
(In reply to comment #14)
> (In reply to comment #1)
> > - funny "Please stand by..." message when using
> > https://bugs.kde.org/buglist.cgi?emailreporter1=1&emailtype1=exact&email1=adawit%40kde.org
> > - Often says "Waiting for reply ..." in status bar, but page is loaded (Stop
> > button disabled).
> 
> ^^^^
> Same issue was reported in bug #212043

huh ? What is that report got to do with this ?? That was reported against KDE 4.3.2. This is about a commit that only exists in the yet unreleased KDE 4.7.
Comment 16 Dawit Alemayehu 2011-05-21 16:33:21 UTC
(In reply to comment #13)
> I only tested quickly, but after reverting dffd5950 all issues mentioned in
> comment #1 are no longer visible. I was able to reassign a bug, no funny
> "Please stand by ..." page, and process finished correctly.

Ok. I will revert even though I am unable to reproduce the original issue at all. I am still using konqueror + khtml. Anyhow, someone from khtml group can address the bug that change was supposed to address.
Comment 17 Dawit Alemayehu 2011-05-21 16:43:46 UTC
Git commit 02a68ff2b6e247d308efac1727a0401f12a1db0a by Dawit Alemayehu.
Committed on 21/05/2011 at 16:39.
Pushed by adawit into branch 'master'.

Revert commit dffd5950b35ead8c14c5cf3a2495be5f8b7b9a1f because it causes regression for some people.

CCBUG: 273239
CCBUG: 273783

M  +1    -1    kparts/browserrun.cpp     

http://commits.kde.org/kdelibs/02a68ff2b6e247d308efac1727a0401f12a1db0a
Comment 18 Christoph Feck 2011-05-21 16:46:54 UTC
I will now leave for weekend and cannot do further testing for the next 50+ hours. If this is indeed a blocker, someone else needs to make sure the tarballs for beta1 don't have this bug.
Comment 19 Andrea Iacovitti 2011-05-21 16:49:20 UTC
(In reply to comment #15)
> (In reply to comment #14)
> > (In reply to comment #1)
> > > - funny "Please stand by..." message when using
> > > https://bugs.kde.org/buglist.cgi?emailreporter1=1&emailtype1=exact&email1=adawit%40kde.org
> > > - Often says "Waiting for reply ..." in status bar, but page is loaded (Stop
> > > button disabled).
> > 
> > ^^^^
> > Same issue was reported in bug #212043
> 
> huh ? What is that report got to do with this ?? That was reported against KDE
> 4.3.2. This is about a commit that only exists in the yet unreleased KDE 4.7.

May be to exactly suggests that the commit in question is unrelated (given that
it is an old observed issue) ?
Comment 20 Anne-Marie Mahfouf 2011-05-21 17:50:22 UTC
OK the bug is gone after your revert, weird that you could not reproduce it.
Comment 21 Dawit Alemayehu 2011-05-21 19:19:01 UTC
(In reply to comment #20)
> OK the bug is gone after your revert, weird that you could not reproduce it.

Yes, it is rather baffling that I cannot reproduce it at all. Unfortunately the revert will cause KIO slaves not to be reused under certain conditions, as in bug# 273239, when using konqueror + khtml. Oh well...
Comment 22 Christoph Feck 2011-05-23 03:48:14 UTC
Regarding the "Please stand by ..." mentioned in comment #19, I was able to trigger it on today's trunk, filed a separate bug 273931.
Comment 23 Dawit Alemayehu 2011-06-18 20:55:40 UTC
Git commit a3d65e2bf927c79fa4426116e9ca9169c0441758 by Dawit Alemayehu.
Committed on 17/06/2011 at 01:45.
Pushed by adawit into branch 'master'.

When a request is not handled, instead of discarding any KIO slave that might
have been put-on-hold, publish it so that it can be reused.

This fixes bug 273239 without introducing the regression reported in bug 273783.

BUG: 273239
CCBUG: 273783
FIXED-IN: 4.7.0
REVIEW: 101648

M  +1    -1    kparts/browserrun.cpp     

http://commits.kde.org/kdelibs/a3d65e2bf927c79fa4426116e9ca9169c0441758