0.3.17 testing

[BrentNewland]

How about having another bug drive before release? Just like before, ask the community to test every bug that's still open to verify the issue. Perhaps try to get at least one developer comment on each remaining bug, explaining what it could possibly be related to, how important it is, etc.

open JIRA bugs

I bet a lot of those 1,553 bugs are no longer valid.

[Z98]

For what I sincerely hope is the last RC. For this one, there are two fixes that I would like checked. First, if people using vbox could see if when they click on desktop icons, the icons are actually highlighted blue and selected. Second, also in vbox, when moving around windows, there isn't a cascading effect that leaves afterimages of the window. Last, this RC contains a fix to font rendering that some people were complaining about showing up in winamp and Java's installer.

http://sourceforge.net/projects/reactos ... p/download

[Webunny]

There's no sound or drivers related commit in there. Since you have 34 revisions remaining without any suspect, the best idea would be to go in power-of-2 steps. You're almost there, should know in 4 attempts.

Well, I guess I'll continue this evening...


edit: 64217 to 64234

...say, how comes 64223, 64224 and 64225 aren't on there?

...I begin to suspect 64222 will work, and 64226 won't...

Aaaand...nope.

Range: 64217 to 64222

[hbelusca]

...say, how comes 64223, 64224 and 64225 aren't on there?

Because those were not commits to trunk (check SVN).

[Webunny]

...say, how comes 64223, 64224 and 64225 aren't on there?

Because those were not commits to trunk (check SVN).

Well, 64220 works, while 64222 didn't work...any ideas?

Ok, 64221 works too, so it's confirmed that it's definitely 64222 which borked it.

[Webunny]

Ok, going for the putty-log (serial cable), since I'm real hardware (regression)testing (specs, including soundcard, can be seen here: http://www.reactos.org/wiki/PC_ROS_Rigs). The first log is just starting up 64221, after clean install, no attempt to install the sounddrivers:

http://pastebin.com/Bhf7eZM4

This is after installing the drivers (with an exe file). However, something weird happend; when restarting, it normally starts up with the 'musictone' and the window of the application (wich installs the sound). Now I just...hanged. Restarted a few times, even in normal mode, but then I got a BSOD. I never got anything like that in all my prior regressiontesting, so I presume it has something to do with the fact I'm debugging with serial, now? Strange:

http://pastebin.com/AUqQwXyW

Well, dunno what happened, but it always bsod'ed, so I reinstalled everything as last time, and after the instalment I restarted, and now the application did work (window of the app pops up, sound/musictone is hearable). I note it says 'hack' there at the end. Maybe that was a hack that made it work, and that has been removed, making it unworkable now? (edit: according to source 'V', this has nothing to do with it, though):

http://pastebin.com/dE8Tmf0h

---------------------------------------------------


This one is of 64222, first without installing anything (of the audiodrivers):

http://pastebin.com/9r5DDVw8


This is after installing the drivers:

http://pastebin.com/DmWf3W2Y

This is when restarting ROS after the installation (exe file) of the drivers, similar to with 64221:

http://pastebin.com/D190kkEi

Note that this time, no window appears, no music is hearable, etc. This, in contrast to the 64221 version.

--------------------------------------------------------

Last edit: I got some (well, a lot of) help in pinpointing the problem and it turns out it's not as much an error with the sounddriver, as with the mixer. The result is still no sound when you play a vid or music, but it's still another thing to look for.


So 64222 didn't break my audio driver, it broke (the instalment of) my mixer application. And it likely broke my mixer cause it didn't find the registered coms (64222 changed the registry hives). I was told that it should have been corrected later on (the hive-thingy), but since I checked up to 65000 and RC4 (where it still failed), maybe it wasn't fixed correctly. But the VM-experts (aka, not me ) then know what to look for, namely some registry thing concerning com files, at 64222.

[Black_Fox]

First, if people using vbox could see if when they click on desktop icons, the icons are actually highlighted blue and selected. Second, also in vbox, when moving around windows, there isn't a cascading effect that leaves afterimages of the window.

Hm, I saw a related commit being added into the branch, but it doesn't help with left-clicking on windows, still don't react. OTOH, the windows don't have any cascading effect.
EDIT: the left-clicking works for me in trunk 65108, probably because of the revert in 65060.

Last, this RC contains a fix to font rendering that some people were complaining about showing up in winamp and Java's installer.

Some people being basically me, I am now satisfied with the results

[Webunny]

Ok, going for the putty-log (serial cable), since I'm real hardware (regression)testing (specs, including soundcard, can be seen here: http://www.reactos.org/wiki/PC_ROS_Rigs). The first log is just starting up 64221, after clean install, no attempt to install the sounddrivers:

http://pastebin.com/Bhf7eZM4

This is after installing the drivers (with an exe file). However, something weird happend; when restarting, it normally starts up with the 'musictone' and the window of the application (wich installs the sound). Now I just...hanged. Restarted a few times, even in normal mode, but then I got a BSOD. I never got anything like that in all my prior regressiontesting, so I presume it has something to do with the fact I'm debugging with serial, now? Strange:

http://pastebin.com/AUqQwXyW

Well, dunno what happened, but it always bsod'ed, so I reinstalled everything as last time, and after the instalment I restarted, and now the application did work (window of the app pops up, sound/musictone is hearable). I note it says 'hack' there at the end. Maybe that was a hack that made it work, and that has been removed, making it unworkable now?:

http://pastebin.com/dE8Tmf0h

---------------------------------------------------


This one is of 64222, first without installing anything (of the audiodrivers):

http://pastebin.com/9r5DDVw8


This is after installing the drivers:

http://pastebin.com/DmWf3W2Y

This is when restarting ROS after the installation (exe file) of the drivers, similar to with 64221:

http://pastebin.com/D190kkEi

Note that this time, no window appears, no music is hearable, etc. This, in contrast to the 64221 version.

--------------------------------------------------------

Last edit: well, I got some help in pinpointing the problem and it turns out it's not as much an error with the sounddriver, as with the mixer. The result is still no sound when you play a vid or music, but it's still another thing to look for.


So 64222 didn't break my audio driver, it broke (the instalment of) my mixer application. And it likely broke my mixer cause it didn't find the registered coms (64222 changed the registry hives). I was told that it should have been corrected later on (the hive-thingy), but since I checked up to 65000 and RC4 (where it still failed), maybe it wasn't fixed correctly. But the VM-experts (aka, not me ) then know what to look for, namely some registry thing concerning com files.

Made a JIRA report of it: https://jira.reactos.org/browse/CORE-8726

PS. I hope this regression can be rectified for the 0.3.17 release?

[MadWolf]

For what I sincerely hope is the last RC. For this one, there are two fixes that I would like checked. First, if people using vbox could see if when they click on desktop icons, the icons are actually highlighted blue and selected.

it is not fixed a temp fix is to go into the settings -> system -> pointing Device -> ps/2 Mouse you get this bug if you have this setting set to USB Tablet
edit
i think the bug is in the USB human interface device driver

[Webunny]

So, anyway...should I give more info or pastebin another log of something, for the problem I described? If so, then please tell me if anything else is required. I would like to see it solved, since it's a major bummer, but I don't see any reaction to it on JIRA, so I don't know if something else is needed or not, from my part.

This makes me think about the other thread where the 'tags' of JIRA are discussed; this would be an example where it would be useful to know whether nothing is done with it because it lacks sufficient info, or because it's been overlooked in the mass of bugreports of lately, or it's of really low priority, or what...

[Z98]

1) Any logs associated with a report should be directly attached to a ticket. Relying on external links is an excellent way to lose information if there is ever a need to look back at old reports.

2) Jerome's reaction to the information currently available has been to wonder if the system wasn't already trashed before the driver was installed. He said he would take a look at it later in detail.

[Webunny]

1) Any logs associated with a report should be directly attached to a ticket. Relying on external links is an excellent way to lose information if there is ever a need to look back at old reports.

2) Jerome's reaction to the information currently available has been to wonder if the system wasn't already trashed before the driver was installed. He said he would take a look at it later in detail.

1)Well, maybe, but unless the forum of ROS is going to disappear any time soon, it will last its day of solving. (hopefully)

2)Don't quite understand that. What about the "So 64222 didn't break my audio driver, it broke (the instalment of) my mixer application. And it likely broke my mixer cause it didn't find the registered coms (64222 changed the registry hives). I was told that it should have been corrected later on (the hive-thingy), but since I checked up to 65000 and RC4 (where it still failed), maybe it wasn't fixed correctly. But the VM-experts (aka, not me ) then know what to look for, namely some registry thing concerning com files, at 64222." That was actually based on what people in IRC said to me (especially 'V'), so I doubt it's all that nonsensical.

But the point is; you know I have been regression-testing, right? I've been burning and (clean)installing a dozen times through CD-RW's. All of which failed with the sound after 64222, and all of which worked with and before 64221. So, I'm not getting what you mean by 'the system was trashed before the driver'. If you're talking about it 'hanging' on the second try: that was weird, true, but as said, that wasn't classical behaviour for the regressiontesting, and as far as I'm aware, had nothing directly top do with the sound-problem. The point what it really is about, is the difference between 64221 and 64222, where the sound/mixer works with the former, but not with the latter. (and it works with all the builds before 64221 and not with all the builds after 64222; a classical regression at that point, thus). So it can't be that it was 'trashed already' before that point in regard to this problem, since it worked.

The actual driver itself seemed to have installed, when I looked at 'properties for sound and audio devices'. 'The system' (aka, if you mean ROS) started and acts normally for the rest: only the sound is gone (from 64222 onwards). If you mean the sound-instalment application: well, the driver got on there, allright. If you mean 'my rig' with 'the system', see above. The fact still remains it worked with 64221, and didn't with 64222.

[Z98]

1) You pasted the logs in pastebin, and then only included the link to the forum. You're requiring people who look at reports to take two additional hops to get to the relevant data, which is located on an external site over which we have no control. Let me put it another way, any logs or screenshots that are not directly attached to a ticket might as well not exist as far as the jira workflow is concerned. If you create inconveniences for the developers that try to follow up on reports, their reaction will be to ignore said reports.

2) I said as much to Jerome, hence why he said he'd look into the issue later. Take it as an indication of him not having any idea why this bug is occurring. And the unlikelihood that this would get resolved before I freeze the 0.3.17 branch.

[Webunny]

1) You pasted the logs in pastebin, and then only included the link to the forum. You're requiring people who look at reports to take two additional hops to get to the relevant data, which is located on an external site over which we have no control. Let me put it another way, any logs or screenshots that are not directly attached to a ticket might as well not exist as far as the jira workflow is concerned. If you create inconveniences for the developers that try to follow up on reports, their reaction will be to ignore said reports.

2) I said as much to Jerome, hence why he said he'd look into the issue later. Take it as an indication of him not having any idea why this bug is occurring. And the unlikelihood that this would get resolved before I freeze the 0.3.17 branch.

1)Are you proposing I copy/paste the logfiles directly into JIRA? That way, it'll only take zero hops to look at it. That said, 6 logs might be a bit too much (I only did that per suggestion of 'V'). What do you suggest which log I take? The last one (of 64222)?

2)Yeah, since it worked before, it's difficult to see how it would have been 'already trashed before'. It is RHW testing, though, so some anomalies could pop up that isn't seen in VM (at least, I didn't hear anyone else complain about the sound). But it definitely worked with 64221. No idea? So the registry thing concerning com files, at 64222, which changed the registry hives, is no good as an indication? Bummer. I liked having sound. That said, I don't always install it on every build, otherwise I would have noted it sooner, so it's not really dramatic/critical neither.

[Z98]

Jira tickets allow for attachments. It's how everyone else who provides logs adds them to tickets.