The appdb has a "browse by ratings" page, e.g. you can see all apps that have at least one Gold-rated version at http://appdb.winehq.org/browse_by_rating.php?sRating=Gold
This can serve as a kind of poor man's application regression finder; look at each app on that page, and look for apps whose latest version is rated lower than gold.
For instance, WinZip is listed on the Gold page. WinZip 9 is indeed rated gold, but winzip 10 is only silver. Likesize, Sametime Connect has an old version rated gold, and a new one rated garbage.
If nothing else, it would be great for someone to look at all such 'regressions', test them with current wine, and update the appdb and bugzilla with what they find. Who knows, maybe we can clear a few of these up without too much effort. - Dan
On 9/27/06, Dan Kegel dank@kegel.com wrote:
The appdb has a "browse by ratings" page, e.g. you can see all apps that have at least one Gold-rated version at http://appdb.winehq.org/browse_by_rating.php?sRating=Gold
This can serve as a kind of poor man's application regression finder; look at each app on that page, and look for apps whose latest version is rated lower than gold.
For instance, WinZip is listed on the Gold page. WinZip 9 is indeed rated gold, but winzip 10 is only silver. Likesize, Sametime Connect has an old version rated gold, and a new one rated garbage.
If nothing else, it would be great for someone to look at all such 'regressions', test them with current wine, and update the appdb and bugzilla with what they find. Who knows, maybe we can clear a few of these up without too much effort.
- Dan
When test results were added Tony and myself talked about all kinds of interesting ways we could farm information from those results. Regression detection via test results is one of the interesting ways we can use the data in the appdb. Basically we could look at the test results across differing versions of wine and look for cases where the rating decreased while the wine version increased.
Another interesting one that I thought would be useful is to look for discrepancies between the results of the same application and version and same wine version across different linux distributions. We could look at this on a large scale to see if some distributions had more problems than others for the same versions of wine.
These are both pretty easy to implement. If anyone is interested in doing so and would like some assistance feel free to email me.
Chris
On 9/27/06, Chris Morgan chmorgan@gmail.com wrote:
When test results were added Tony and myself talked about all kinds of interesting ways we could farm information from those results. [suggestion for two particular kinds of searches]
I'd like to see a more general search function rather than specialized searches. An advanced search that lets you search by rating, by unfixed bugilla bug count, and by most recently tested wine version, would enable quite a few kinds of regression searches.
Also, I'd like to see us add a checkbox which is the equivalent of the 'download' keyword. That would let people who want to work on problems find ones that don't require purchasing software. - Dan
On 9/27/06, Dan Kegel dank@kegel.com wrote:
On 9/27/06, Chris Morgan chmorgan@gmail.com wrote:
When test results were added Tony and myself talked about all kinds of interesting ways we could farm information from those results. [suggestion for two particular kinds of searches]
I'd like to see a more general search function rather than specialized searches. An advanced search that lets you search by rating, by unfixed bugilla bug count, and by most recently tested wine version, would enable quite a few kinds of regression searches.
Also, I'd like to see us add a checkbox which is the equivalent of the 'download' keyword. That would let people who want to work on problems find ones that don't require purchasing software.
- Dan
Having both useful specialized reports AND a more advanced general search would be useful. EA Durbin mentioned an interest in improving the search in the ways you've mentioned. Maybe he can comment on what he was thinking.
Chris
On 9/27/06, Dan Kegel dank@kegel.com wrote:
Also, I'd like to see us add a checkbox which is the equivalent of the 'download' keyword. That would let people who want to work on problems find ones that don't require purchasing software.
- Dan
One thing that might make this a little more helpful (or at least a little more clear) would be if we split the download keyword into separate keywords for trial version downloads, and freeware version downloads, and possibly open source downloads. That way we know without having to navigate the software publishers' site whether it is freely downloadable, open source, or trialware. That should make it a little easier to help separate bugs that appear in the full version but not in the trial version, from ones that appear in the trial but not the full.
On 9/27/06, Tom Spear speeddymon@gmail.com wrote:
On 9/27/06, Dan Kegel dank@kegel.com wrote:
Also, I'd like to see us add a checkbox which is the equivalent of the 'download' keyword. That would let people who want to work on problems find ones that don't require purchasing software.
One thing that might make this a little more helpful (or at least a little more clear) would be if we split the download keyword into separate keywords for trial version downloads, and freeware version downloads, and possibly open source downloads. That way we know without having to navigate the software publishers' site whether it is freely downloadable, open source, or trialware. That should make it a little easier to help separate bugs that appear in the full version but not in the trial version, from ones that appear in the trial but not the full.
I haven't run into too many of those. I'm afraid it might be too much information...
On 9/27/06, Dan Kegel dank@kegel.com wrote:
On 9/27/06, Tom Spear speeddymon@gmail.com wrote:
On 9/27/06, Dan Kegel dank@kegel.com wrote:
Also, I'd like to see us add a checkbox which is the equivalent of the 'download' keyword. That would let people who want to work on problems find ones that don't require purchasing software.
One thing that might make this a little more helpful (or at least a
little
more clear) would be if we split the download keyword into separate
keywords
for trial version downloads, and freeware version downloads, and
possibly
open source downloads. That way we know without having to navigate the software publishers' site whether it is freely downloadable, open
source, or
trialware. That should make it a little easier to help separate bugs
that
appear in the full version but not in the trial version, from ones that appear in the trial but not the full.
I haven't run into too many of those. I'm afraid it might be too much information...
Mind if I ask how so? Many games bugs cant be properly diagnosed because the bug occurs somewhere that the trial version doesnt include.
On 9/27/06, Tom Spear speeddymon@gmail.com wrote:
On 9/27/06, Dan Kegel dank@kegel.com wrote:
On 9/27/06, Tom Spear speeddymon@gmail.com wrote:
On 9/27/06, Dan Kegel dank@kegel.com wrote:
Also, I'd like to see us add a checkbox which is the equivalent of the 'download' keyword. That would let people who want to work on problems find ones that don't require purchasing software.
One thing that might make this a little more helpful (or at least a
little
more clear) would be if we split the download keyword into separate
keywords
for trial version downloads, and freeware version downloads, and
possibly
open source downloads. That way we know without having to navigate the software publishers' site whether it is freely downloadable, open
source, or
trialware. That should make it a little easier to help separate bugs
that
appear in the full version but not in the trial version, from ones that appear in the trial but not the full.
I haven't run into too many of those. I'm afraid it might be too much information...
Mind if I ask how so? Many games bugs cant be properly diagnosed because the bug occurs somewhere that the trial version doesnt include.
We've run into this before with the Warhammer 40k installer. When we find that the downloadable trial version works, but the cd installer doesn't, we take off the download keyword. Adding more than one type of download keyword will add too much noise with absolutely no benefit.
On 9/27/06, James Hawkins truiken@gmail.com wrote:
We've run into this before with the Warhammer 40k installer. When we find that the downloadable trial version works, but the cd installer doesn't, we take off the download keyword. Adding more than one type of download keyword will add too much noise with absolutely no benefit.
-- James Hawkins
Arguably, I can understand that a cd installer wouldnt fall under the download category, nor would a bug in the retail version of the game, so removing the download keyword would do fine, but what if there are 2 different free versions of a program (ie. zonealarm has a free version and a trial pro version, same for ad-aware). If you have a trial a freeware keyword, then it would be easier to tell which version is being referenced, and therefore easier (and faster) to diagnose and fix the bug.
On 9/27/06, Tom Spear speeddymon@gmail.com wrote:
On 9/27/06, James Hawkins truiken@gmail.com wrote:
We've run into this before with the Warhammer 40k installer. When we find that the downloadable trial version works, but the cd installer doesn't, we take off the download keyword. Adding more than one type of download keyword will add too much noise with absolutely no benefit.
-- James Hawkins
Arguably, I can understand that a cd installer wouldnt fall under the download category, nor would a bug in the retail version of the game, so removing the download keyword would do fine, but what if there are 2 different free versions of a program (ie. zonealarm has a free version and a trial pro version, same for ad-aware). If you have a trial a freeware keyword, then it would be easier to tell which version is being referenced, and therefore easier (and faster) to diagnose and fix the bug.
The answer is in your reply. The difference between the two is in the versions. One is a free version, the other is the trial of a pro version. There's a difference, and from our point of view, they should be considered as two different applications, so a bug report would just say which version it is, e.g., "ZoneAlarm free version fails to install" and "ZoneAlarm Pro Trial fails to install".
On Wed, 27 Sep 2006, Tom Spear wrote: [...]
I haven't run into too many of those. I'm afraid it might be too much information...
Mind if I ask how so? Many games bugs cant be properly diagnosed because the bug occurs somewhere that the trial version doesnt include.
If the bug cannot be reproduced in the trial version, then it should not have the 'download' keyword since it cannot be reproduced simply dy downloading the software. Same thing if the full version is downloadable but requires a paying subscription of some sort in order to reproduce the bug.
The 'download' keyword really stands for 'the free download will let you reproduce the bug'.
On 9/27/06, Chris Morgan chmorgan@gmail.com wrote:
On 9/27/06, Dan Kegel dank@kegel.com wrote:
The appdb has a "browse by ratings" page, e.g. you can see all apps that have at least one Gold-rated version at http://appdb.winehq.org/browse_by_rating.php?sRating=Gold
This can serve as a kind of poor man's application regression finder; look at each app on that page, and look for apps whose latest version is rated lower than gold.
For instance, WinZip is listed on the Gold page. WinZip 9 is indeed rated gold, but winzip 10 is only silver. Likesize, Sametime Connect has an old version rated gold, and a new one rated garbage.
If nothing else, it would be great for someone to look at all such 'regressions', test them with current wine, and update the appdb and bugzilla with what they find. Who knows, maybe we can clear a few of these up without too much effort.
- Dan
When test results were added Tony and myself talked about all kinds of interesting ways we could farm information from those results. Regression detection via test results is one of the interesting ways we can use the data in the appdb. Basically we could look at the test results across differing versions of wine and look for cases where the rating decreased while the wine version increased.
Another interesting one that I thought would be useful is to look for discrepancies between the results of the same application and version and same wine version across different linux distributions. We could look at this on a large scale to see if some distributions had more problems than others for the same versions of wine.
These are both pretty easy to implement. If anyone is interested in doing so and would like some assistance feel free to email me.
Chris
Also, on that note, just a helpful site pointer. There are several different chat and im clients that work on wine, as long as the correct version is installed, while other versions dont work for squat. oldversion.com has the majority of officially released versions of the most popular ones, including but not limited to: AIM, ICQ, MSN, Trillian, and Yahoo Messenger.
Hope that helps.
-
Chris Morgan -
Dan Kegel -
Francois Gouget -
James Hawkins -
Tom Spear