Comments for making things suck less http://mattr.info:8080/blog mattr's blog Wed, 24 Feb 2010 22:10:35 -0600 http://wordpress.org/?v=2.8.6 hourly 1 Comment on Yahoo Webcam Fixes for Kopete by crjackson http://mattr.info:8080/blog/2010/01/11/yahoo-webcam-fixes-for-kopete/comment-page-1/#comment-29543 crjackson Wed, 24 Feb 2010 22:10:35 +0000 http://mattr.info:8080/blog/?p=193#comment-29543 so how do I get this update. my kopete still doesn't send or view webcam. so how do I get this update. my kopete still doesn’t send or view webcam.

]]> Comment on Yahoo Webcam Fixes for Kopete by Paul Fee http://mattr.info:8080/blog/2010/01/11/yahoo-webcam-fixes-for-kopete/comment-page-1/#comment-29537 Paul Fee Tue, 12 Jan 2010 12:49:16 +0000 http://mattr.info:8080/blog/?p=193#comment-29537 I've had problems using Kopete with Yahoo. I could connect and receive messages, but my messages wouldn't reach the recipient. There was no local indication of anything wrong. Has anyone else seen this? Was it a known issue that's now been fixed? Should I try again and raise a new bug if necessary? I've switched to pidgin because of it, but would like to return to kopete once Yahoo IM works for me. I’ve had problems using Kopete with Yahoo. I could connect and receive messages, but my messages wouldn’t reach the recipient. There was no local indication of anything wrong.

Has anyone else seen this? Was it a known issue that’s now been fixed? Should I try again and raise a new bug if necessary?

I’ve switched to pidgin because of it, but would like to return to kopete once Yahoo IM works for me.

]]> Comment on Yahoo Webcam Fixes for Kopete by Venky80 http://mattr.info:8080/blog/2010/01/11/yahoo-webcam-fixes-for-kopete/comment-page-1/#comment-29536 Venky80 Tue, 12 Jan 2010 07:06:28 +0000 http://mattr.info:8080/blog/?p=193#comment-29536 so guess i can close my defect which i opened. so guess i can close my defect which i opened.

]]> Comment on Yahoo Webcam Fixes for Kopete by anne.. http://mattr.info:8080/blog/2010/01/11/yahoo-webcam-fixes-for-kopete/comment-page-1/#comment-29535 anne.. Tue, 12 Jan 2010 05:05:58 +0000 http://mattr.info:8080/blog/?p=193#comment-29535 what do i do with this na? what do i do with this na?

]]> Comment on First version of QtPersistence released by Mario http://mattr.info:8080/blog/2009/12/29/first-version-of-qtpersistence-released/comment-page-1/#comment-29533 Mario Sat, 02 Jan 2010 09:35:57 +0000 http://mattr.info:8080/blog/?p=188#comment-29533 Hi, IMHO that's a very cool project! I have some questions: 1 Do you intend to support already existing databases, which includes fine graned control over QtPersistence? 1a E.g., IMHO it would be useful to have an option for telling QtPersistence, that all columns in all tables which name is USER_ID can be looked up in a table named USER. 1b Would also be useful if one could control e.g. the format of dates (e.g. already existing databases may not use the rdbms' builtin date format, but only some kind of string representation). 2 Do you think that some time in the future it will possible of supporting non-database backend, e.g. saving data in-memory, using Qt-objects only? 3 I have never managed to understand Qt's Interview framework as soon as it comes to QAbstractItemModel. Do you plan to make it possible of "abusing" QtPersistence as a workaround for that? I mean, instead of creating a new class which inherits from QAbstractItemModel one could use QtPersistence? I know that it is already possible to manipulate or view one single database table using Interview, but a table is not a tree... Regards Mario Hi,

IMHO that’s a very cool project! I have some questions:

1 Do you intend to support already existing databases, which includes fine graned control over QtPersistence?

1a E.g., IMHO it would be useful to have an option for telling QtPersistence, that all columns in all tables which name is USER_ID can be looked up in a table named USER.

1b Would also be useful if one could control e.g. the format of dates (e.g. already existing databases may not use the rdbms’ builtin date format, but only some kind of string representation).

2 Do you think that some time in the future it will possible of supporting non-database backend, e.g. saving data in-memory, using Qt-objects only?

3 I have never managed to understand Qt’s Interview framework as soon as it comes to QAbstractItemModel. Do you plan to make it possible of “abusing” QtPersistence as a workaround for that? I mean, instead of creating a new class which inherits from QAbstractItemModel one could use QtPersistence? I know that it is already possible to manipulate or view one single database table using Interview, but a table is not a tree…

Regards
Mario

]]> Comment on First version of QtPersistence released by Matt http://mattr.info:8080/blog/2009/12/29/first-version-of-qtpersistence-released/comment-page-1/#comment-29532 Matt Thu, 31 Dec 2009 19:00:37 +0000 http://mattr.info:8080/blog/?p=188#comment-29532 @Jay: exactly right. Not such a great name, but the name can always be changed. I suck at naming things, so I'd love to see any suggestions. @Albert: eventually, yes. I'm sure there will be a refactoring that will need to be done to support those, and the necessary evils around them. :) I just saw Wt::Dbo the other day in fact, and intend to look at it more soonish. Thanks for the comments everybody! @Jay: exactly right. Not such a great name, but the name can always be changed. I suck at naming things, so I’d love to see any suggestions.

@Albert: eventually, yes. I’m sure there will be a refactoring that will need to be done to support those, and the necessary evils around them. :) I just saw Wt::Dbo the other day in fact, and intend to look at it more soonish.

Thanks for the comments everybody!

]]> Comment on First version of QtPersistence released by Albert http://mattr.info:8080/blog/2009/12/29/first-version-of-qtpersistence-released/comment-page-1/#comment-29531 Albert Thu, 31 Dec 2009 15:02:58 +0000 http://mattr.info:8080/blog/?p=188#comment-29531 Nice to see you working on this! By the way, I took a look at the code and do you plan tu support object relations? If so, I think you'll find the .save() function in persistent objects is not the most appropiate way to handle them. That's why I used a manager and commit() and rollback() functions in Kandau. Also, you may be interested in Wt::Dbo, which has been released recently: http://www.webtoolkit.eu/wt#/blog/2009/11/26/wt__dbo__an_orm__c___style/more Nice to see you working on this! By the way, I took a look at the code and do you plan tu support object relations? If so, I think you’ll find the .save() function in persistent objects is not the most appropiate way to handle them. That’s why I used a manager and commit() and rollback() functions in Kandau.

Also, you may be interested in Wt::Dbo, which has been released recently:

http://www.webtoolkit.eu/wt#/blog/2009/11/26/wt__dbo__an_orm__c___style/more

]]> Comment on First version of QtPersistence released by Jay http://mattr.info:8080/blog/2009/12/29/first-version-of-qtpersistence-released/comment-page-1/#comment-29530 Jay Wed, 30 Dec 2009 23:14:06 +0000 http://mattr.info:8080/blog/?p=188#comment-29530 Persistence != ORM Persistence != ORM

]]> Comment on kde bugzilla ui changes: feedback wanted by Max Kanat-Alexander http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29527 Max Kanat-Alexander Tue, 24 Nov 2009 09:47:22 +0000 http://mattr.info:8080/blog/?p=184#comment-29527 You know, as far as backtrace duplication and analysis goes, I wrote a Bugzilla Extension for bugzilla.gnome.org, called traceparser, that does all sorts of fancy things with stack traces in Bugzilla, including automatically finding identical and similar traces (even when filing a bug). It doesn't work in Bugzilla 3.4 without modification, but once Bugzilla 3.6 comes out, I could probably make it work pretty nicely if you guys wanted to use it on bugs.kde.org. Why does bugs.kde.org run under mod_cgi, by the way? That's the speed problem, for sure. -Max You know, as far as backtrace duplication and analysis goes, I wrote a Bugzilla Extension for bugzilla.gnome.org, called traceparser, that does all sorts of fancy things with stack traces in Bugzilla, including automatically finding identical and similar traces (even when filing a bug). It doesn’t work in Bugzilla 3.4 without modification, but once Bugzilla 3.6 comes out, I could probably make it work pretty nicely if you guys wanted to use it on bugs.kde.org.

Why does bugs.kde.org run under mod_cgi, by the way? That’s the speed problem, for sure.

-Max

]]> Comment on kde bugzilla ui changes: feedback wanted by John http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29526 John Sat, 21 Nov 2009 21:53:08 +0000 http://mattr.info:8080/blog/?p=184#comment-29526 My main issues: 1) Speed! It really is frustrating waiting on bko, but I guess that's more a hosting issue. 2) Reassigning to a different Product, you have to wait for the next screen to then choose the right Component, couldn't we do it in one go? 3) When closing as fixed should be able to say what release it was fixed in. That would also need some work in the svn BUG: script, perhaps to automatically assign the next release for the branch the commit was on? 4) Target milestone doesn't seem to work? 5) In the new layout, could some of the lesser used fields like URL, Keywords, Depends, Blocks be moved to the right? 6) Looks like the layout issues that konqi had with width of bug lists has been fixed? 7) Documentation or flowcharts for the bug lifecycle, how the status is supposed to change, what each value means, etc. We provide absolutely no help to new users or devs to find their way around (or even old ones). My main issues:

1) Speed! It really is frustrating waiting on bko, but I guess that’s more a hosting issue.
2) Reassigning to a different Product, you have to wait for the next screen to then choose the right Component, couldn’t we do it in one go?
3) When closing as fixed should be able to say what release it was fixed in. That would also need some work in the svn BUG: script, perhaps to automatically assign the next release for the branch the commit was on?
4) Target milestone doesn’t seem to work?
5) In the new layout, could some of the lesser used fields like URL, Keywords, Depends, Blocks be moved to the right?
6) Looks like the layout issues that konqi had with width of bug lists has been fixed?
7) Documentation or flowcharts for the bug lifecycle, how the status is supposed to change, what each value means, etc. We provide absolutely no help to new users or devs to find their way around (or even old ones).

]]> Comment on kde bugzilla ui changes: feedback wanted by Diederik van der Boor http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29525 Diederik van der Boor Sat, 21 Nov 2009 10:20:01 +0000 http://mattr.info:8080/blog/?p=184#comment-29525 I can second what Milian Wolff says. - I skip the wizard by using the "report bug" menu item. The current wizard makes me type redundant input. E.g. I selected I used SuSE, next it asks how I got my packages, format they are, and whether I'm using Linux/Windows/MacOS as OS. Especially the last question is just annoying. With the report feature, Bugzilla asks whether the app+version are correct. Once you uncheck the version number, you also need to specify which app this is about, etc.. - Searching for duplicates does not work. I report a plasma crash, and bugzilla also shows wishes for plasma features.. I report an Akonadi problem, and bugzilla suggests bug reports about Kate too. (imaginary example, but silly things like these do happen). So I report the bug, spend time writing an explaination, and next there is a developer to close it as DUPLICATE. I rather would have figured that out myself already :) With these things fixed, I think bugzilla could - at least for users - get a lot better already. I don't think the software is really that bad at all, but could need some polishing. I can second what Milian Wolff says.

- I skip the wizard by using the “report bug” menu item.

The current wizard makes me type redundant input. E.g. I selected I used SuSE, next it asks how I got my packages, format they are, and whether I’m using Linux/Windows/MacOS as OS. Especially the last question is just annoying.

With the report feature, Bugzilla asks whether the app+version are correct. Once you uncheck the version number, you also need to specify which app this is about, etc..

- Searching for duplicates does not work.

I report a plasma crash, and bugzilla also shows wishes for plasma features.. I report an Akonadi problem, and bugzilla suggests bug reports about Kate too. (imaginary example, but silly things like these do happen).

So I report the bug, spend time writing an explaination, and next there is a developer to close it as DUPLICATE. I rather would have figured that out myself already :)

With these things fixed, I think bugzilla could – at least for users – get a lot better already. I don’t think the software is really that bad at all, but could need some polishing.

]]> Comment on kde bugzilla ui changes: feedback wanted by FiNeX http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29524 FiNeX Fri, 20 Nov 2009 16:45:35 +0000 http://mattr.info:8080/blog/?p=184#comment-29524 I'd prefer the comment reply text field after the comments. I’d prefer the comment reply text field after the comments.

]]> Comment on kde bugzilla ui changes: feedback wanted by Thomas Zander http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29523 Thomas Zander Fri, 20 Nov 2009 09:06:59 +0000 http://mattr.info:8080/blog/?p=184#comment-29523 Slightly off topic, I think one big problem is that the site is slow and thus searching becomes frustrating. For that reason I started a Qt client that uses bugs.kde.org APIs and stores the bugs locally to search much more intelligently. So for anyone doing triaging, having this app running will likely be a big help too; http://gitorious.org/bugsviewer Slightly off topic, I think one big problem is that the site is slow and thus searching becomes frustrating.
For that reason I started a Qt client that uses bugs.kde.org APIs and stores the bugs locally to search much more intelligently.

So for anyone doing triaging, having this app running will likely be a big help too; http://gitorious.org/bugsviewer

]]> Comment on kde bugzilla ui changes: feedback wanted by Boudewijn Rempt http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29522 Boudewijn Rempt Fri, 20 Nov 2009 08:18:23 +0000 http://mattr.info:8080/blog/?p=184#comment-29522 The new layout is so much better that I would love it if the real bugzilla switched to it today. The new layout is so much better that I would love it if the real bugzilla switched to it today.

]]> Comment on kde bugzilla ui changes: feedback wanted by Milian Wolff http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29521 Milian Wolff Thu, 19 Nov 2009 19:39:00 +0000 http://mattr.info:8080/blog/?p=184#comment-29521 Some feedback (hope it's OK here as well): 1) it's much easier / faster to report a bug on the plain bugzilla installation. I _hate_ the "wizard" on bugs.kde.org, it slows down. The plain one from bugzilla just gets down to the game, doesn't get into your way. Select a product and than file in the missing stuff. Much better! Maybe this is not "good" for the first-time bug-reporters, but it's lightyears faster for "advanced" bug reporters, or developers like me that use bugzilla also for bugs they find (and want to fix eventually). 2) Of course somewhere we need a check for duplicates, but imo this requires a better search form than currently available: I rarely (and others neither) find the same keywords for a given problem. One should _either_ insert keywords _or_ a backtrace. Searching the backtrace for duplicates would imo be much better and saves Dario some time. searching keywords should be optional but required for wishlists and enhancements. Some feedback (hope it’s OK here as well):

1) it’s much easier / faster to report a bug on the plain bugzilla installation. I _hate_ the “wizard” on bugs.kde.org, it slows down. The plain one from bugzilla just gets down to the game, doesn’t get into your way. Select a product and than file in the missing stuff. Much better!

Maybe this is not “good” for the first-time bug-reporters, but it’s lightyears faster for “advanced” bug reporters, or developers like me that use bugzilla also for bugs they find (and want to fix eventually).

2) Of course somewhere we need a check for duplicates, but imo this requires a better search form than currently available: I rarely (and others neither) find the same keywords for a given problem. One should _either_ insert keywords _or_ a backtrace. Searching the backtrace for duplicates would imo be much better and saves Dario some time. searching keywords should be optional but required for wishlists and enhancements.

]]> Comment on kde bugzilla ui changes: feedback wanted by Mark http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29520 Mark Thu, 19 Nov 2009 19:33:24 +0000 http://mattr.info:8080/blog/?p=184#comment-29520 - autoreminder to reporter, asking if bug still exists in new version - would be nice if bugs/wishes could be tied to the development schedule http://techbase.kde.org/Schedules , so that a user knows that a bug is targeted for KDE 4.4, 4.4.1, 4,5.... or as "unknown" This would make development more transparent. Could help to find people who adopt them or let users "lobby" for there bugs. Crashes should be automatically tied to the next release, so that developers have it easier to get an overview and they get a higher priority. - It should be possible to report another bug/ wish to the same component/programe after reporting a bug / wish. - it should be possible to ad an attachment to the initial report. - There should be the possibility to confirm a bug exits still in the actual version. At the moment it is only possible in the text, but I believe it may be more useful to have this automated, so that they can be searched for. Bugs which are confirmed are more likely to still exit and still bother users. - autoreminder to reporter, asking if bug still exists in new version
- would be nice if bugs/wishes could be tied to the development schedule http://techbase.kde.org/Schedules , so that a user knows that a bug is targeted for KDE 4.4, 4.4.1, 4,5…. or as “unknown”
This would make development more transparent. Could help to find people who adopt them or let users “lobby” for there bugs. Crashes should be automatically tied to the next release, so that developers have it easier to get an overview and they get a higher priority.
- It should be possible to report another bug/ wish to the same component/programe after reporting a bug / wish.
- it should be possible to ad an attachment to the initial report.
- There should be the possibility to confirm a bug exits still in the actual version. At the moment it is only possible in the text, but I believe it may be more useful to have this automated, so that they can be searched for. Bugs which are confirmed are more likely to still exit and still bother users.

]]> Comment on kde bugzilla ui changes: feedback wanted by Anne Wilson http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29519 Anne Wilson Thu, 19 Nov 2009 19:20:59 +0000 http://mattr.info:8080/blog/?p=184#comment-29519 I tried the test in rekonq. It's cleaner, and certainly the package list makes things easier, but I did find a couple of problems. I am using Fedora packages, which it said were unlisted binaries. The Component and Version sections were just black patches. I have a screenshot of that part of the window if you'd like to see it. I tried the test in rekonq. It’s cleaner, and certainly the package list makes things easier, but I did find a couple of problems. I am using Fedora packages, which it said were unlisted binaries. The Component and Version sections were just black patches. I have a screenshot of that part of the window if you’d like to see it.

]]> Comment on kde bugzilla ui changes: feedback wanted by Martin Gräßlin http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29518 Martin Gräßlin Thu, 19 Nov 2009 19:02:45 +0000 http://mattr.info:8080/blog/?p=184#comment-29518 So I just played a little bit with the test installation. What I really like compared to bugs.kde.org is that the "mark as duplicate" is at the top of the page. It's so often that I see bugbot reporting a new bug on IRC and I think "dup" and just want to enter the number. But that might be special to kwin as we have a high duplicate bug rate. So for others it might be better to have it below the comment section. Another issue which is perhaps better in the test installation: when I reassign a bug to e.g. plasma I have to change the component then scroll down to the assignee section and click on edit and reset assignee to default. In the test installation it's in one section so I don't need to scroll and I think that lowers the risk that it is forgotten. (Happens quite often that component is changed to kwin, but the assignee is not changed so not email on the mailinglist and no notification) So I just played a little bit with the test installation. What I really like compared to bugs.kde.org is that the “mark as duplicate” is at the top of the page. It’s so often that I see bugbot reporting a new bug on IRC and I think “dup” and just want to enter the number. But that might be special to kwin as we have a high duplicate bug rate. So for others it might be better to have it below the comment section.

Another issue which is perhaps better in the test installation: when I reassign a bug to e.g. plasma I have to change the component then scroll down to the assignee section and click on edit and reset assignee to default. In the test installation it’s in one section so I don’t need to scroll and I think that lowers the risk that it is forgotten. (Happens quite often that component is changed to kwin, but the assignee is not changed so not email on the mailinglist and no notification)

]]> Comment on kde bugzilla ui changes: feedback wanted by FiNeX http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29517 FiNeX Thu, 19 Nov 2009 18:52:19 +0000 http://mattr.info:8080/blog/?p=184#comment-29517 I still think that bugzilla is not the main problem. We need more devs and triagers at first... Anyway, if we have to found something about bugzilla, I could say that a system to reduce duplicates could be helpful... but I really don't know if other tools could help or how to modify bugzilla... I'll give a try to the demo :-) bye! I still think that bugzilla is not the main problem. We need more devs and triagers at first… Anyway, if we have to found something about bugzilla, I could say that a system to reduce duplicates could be helpful… but I really don’t know if other tools could help or how to modify bugzilla…

I’ll give a try to the demo :-)

bye!

]]> Comment on kde bugzilla ui changes: feedback wanted by kiemlo http://mattr.info:8080/blog/2009/11/18/kde-bugzilla-ui-changes-feedback-wanted/comment-page-1/#comment-29516 kiemlo Thu, 19 Nov 2009 18:50:41 +0000 http://mattr.info:8080/blog/?p=184#comment-29516 I think one of the reasons people are not happy with KDE's Bugzilla is the fact that nobody seem to care about fixing the bugs reported there. I think one of the reasons people are not happy with KDE’s Bugzilla is the fact that nobody seem to care about fixing the bugs reported there.

]]>