Welcome to NeoOffice developer notes and announcements
NeoOffice
Developer notes and announcements
 
 

This website is an archive and is no longer active
NeoOffice announcements have moved to the NeoOffice News website


Support
· Forums
· NeoOffice Support
· NeoWiki


Announcements
· Twitter @NeoOffice


Downloads
· Download NeoOffice


  
NeoOffice :: View topic - Patch-8 AppleEvent Debugging libvcl (Bug 209)
Patch-8 AppleEvent Debugging libvcl (Bug 209)
 
   NeoOffice Forum Index -> NeoOffice Testing
View previous topic :: View next topic  
Author Message
OPENSTEP
The One
The One


Joined: May 25, 2003
Posts: 4752
Location: Santa Barbara, CA

PostPosted: Wed Dec 08, 2004 11:56 pm    Post subject: Patch-8 AppleEvent Debugging libvcl (Bug 209)

OK, I've got a patch-8 level VCL library built for trying to track down the AppleEvents that are leading to the various crashes people are seeing with Bug 209 (random crashes in "thread 9" with an ae* function in the trace). The actual bug report this realates to is:

http://bugzilla.neooffice.org/bug.php?op=show&bugid=209

If you're experiencing these crashes, please download the following library:

ftp://ftp.neooffice.org/pub/AEDeubg_NeoJ_Patch8/libvcl645mxp.dylib.gz

MD5Sum : e32c2b855c0a458487115f0233986f01

(MD5 is of the compressed library)

Once you download and uncompress the library, copy it over /Applications/NeoOfficeJ.app/Contents/MacOS/libvcl645mxp.dylib

When you experience the NeoJ crash, open the Console application and look through the console.log (you may need to reload it) for the last bit of text that looks like:

Code:

AE2000 (7438): Sending an event:
------oo start of event oo------
{ 1 } 'aevt':  aevt/quit (ppc ){
          return id: 487456769 (0x1d0e0001)
     transaction id: 0 (0x0)
  interaction level: 112 (0x70)
     reply required: 0 (0x0)
             remote: 0 (0x0)
  target:
    { 2 } 'psn ':  8 bytes {
      { 0x0, 0x2 } (NeoOffice/J)
    }
  optional attributes:
    < empty record >
  event data:
    { 1 } 'aevt':  - 0 items {
    }
}

------oo  end of event  oo------


The output above, for example, is printing out the AppleEvent sent to the virtual machine in response to the application menu Quit item (NeoOffice/J > Quit NeoOffice/J).

The idea behind this "debugging" library is that hopefully someone can get me the printout of the AppleEvent that is causing all of these crashes.

I can't get the crash to reproduce on my machine here and I also am not seeing any AppleEvent being sent as I try to use my activate/deactivate test case, so it's most likely something different. I'll have to install this on my work machine to try to reproduce the crash test case. FWIW, the same test case didn't crash it today. I plan on running with this debug library continuously until I can get that crash to happen again at which point, hopefully, I'll be able to catch that rogue AppleEvent red-handed.

If you feel up to the task, *please* help me to track it down!

ed
Back to top
wade1234
Guest





PostPosted: Thu Dec 09, 2004 12:30 am    Post subject:

There is a problem trying to access that file Ed!

I used ftp client, safari, firefox....get errors with all trying to get it.

You can email me
felho at lycos dot com
Back to top
Waldo
Oracle


Joined: Dec 03, 2004
Posts: 239

PostPosted: Thu Dec 09, 2004 8:14 am    Post subject:

wade1234 wrote:
There is a problem trying to access that file Ed!

I used ftp client, safari, firefox....get errors with all trying to get it.

You can email me
felho at lycos dot com


Yep-- "You don't have permission to access that server"

not even the /pub directory itself..


W
Back to top
OPENSTEP
The One
The One


Joined: May 25, 2003
Posts: 4752
Location: Santa Barbara, CA

PostPosted: Thu Dec 09, 2004 9:21 pm    Post subject:

Yes, there does appear to be something wrong with the FTP server. I will try to fix that asap.

ed
Back to top
OPENSTEP
The One
The One


Joined: May 25, 2003
Posts: 4752
Location: Santa Barbara, CA

PostPosted: Thu Dec 09, 2004 9:37 pm    Post subject:

OK, I found the problematic configuration on the FTP server. Please try again and you should be able to download now.

Note that anonymous users are restricted to only two simultaneous users at a time; if you don't have the tester's user login and password, message me and I'll pass it along. Again, try not to share as the anonymous limit is to keep more bandwidth for trinity Wink

ed
Back to top
wade1234
Guest





PostPosted: Fri Dec 10, 2004 12:29 am    Post subject:

Here is what I get with my crash from the console with your new lib installed:(I can get it to crash everytime)
Hope this helps let me know how it goes!!

{ 1 } 'aevt': aevt/odoc (****){
return id: 62456044 (0x3b900ec)
transaction id: 0 (0x0)
interaction level: 112 (0x70)
reply required: 0 (0x0)
remote: 0 (0x0)
target:
<no target specified>
optional attributes:
< empty record >
event data:
{ 1 } 'aevt': - 1 items {
key '----' -
{ 1 } 'list': - 1 elements {
{ 1 } 'alis': 244 bytes {
FSSpec{ -100, 194062, 12_02_04_Ops_Staff.sxi }
}
}
}
}

AE2000 (5670): Received an event:
------oo start of event oo------
{ 1 } 'aevt': aevt/rapp (ppc ){
return id: 62325413 (0x3b702a5)
transaction id: 0 (0x0)
interaction level: 112 (0x70)
reply required: 0 (0x0)
remote: 0 (0x0)
target:
{ 1 } 'psn ': 8 bytes {
{ 0x0, 0x60001 } (Dock)
}
optional attributes:
{ 1 } 'reco': - 1 items {
key 'optk' -
{ 1 } 'list': - 1 elements {
{ 1 } 'keyw': 4 bytes {
'frnt'
}
}
}

event data:
{ 1 } 'aevt': - 1 items {
key 'frnt' -
{ 1 } 'bool': 1 bytes {
false
}
}
}

------oo end of event oo------
AE2000 (5670): Sending an event:
------oo start of event oo------
{ 1 } 'aevt': aevt/quit (ppc ){
return id: 371589120 (0x16260000)
transaction id: 0 (0x0)
interaction level: 112 (0x70)
reply required: 0 (0x0)
remote: 0 (0x0)
target:
{ 2 } 'psn ': 8 bytes {
{ 0x0, 0x2 } (NeoOffice/J)
}
optional attributes:
< empty record >
event data:
{ 1 } 'aevt': - 0 items {
}
}
Back to top
Waldo
Oracle


Joined: Dec 03, 2004
Posts: 239

PostPosted: Fri Dec 10, 2004 1:04 pm    Post subject:

What I did:

Loaded Neo/J. Went to Tools->Options clicked around, went back adn forth from Neo/J to other apps for a while. couldn't get it to crash.

next I closed Options and typed garbage in Writer. Went back to Tools->Options.. clikced around. Dang it, it won't crash.

So I said maybe I should restart. went file->quit.

pinwheeled. Still going. CPU at 100%. No crash after about 5 minutes, but i hear it working the HD.

Here's the latest bit from the console.

------oo start of event oo------
{ 1 } 'aevt': aevt/quit (ppc ){
return id: 209977852 (0xc8401fc)
transaction id: 0 (0x0)
interaction level: 112 (0x70)
reply required: 0 (0x0)
remote: 0 (0x0)
target:
{ 1 } 'psn ': 8 bytes {
{ 0x0, 0x60001 } (Dock)
}
optional attributes:
< empty record >
event data:
{ 1 } 'aevt': - 0 items {
}
}

------oo end of event oo------

Hope this helps.

W
Back to top
OPENSTEP
The One
The One


Joined: May 25, 2003
Posts: 4752
Location: Santa Barbara, CA

PostPosted: Sun Dec 12, 2004 5:42 pm    Post subject:

Thank you very much for those apple event logs. The quit one I did expect...but at the same time it's disturbing that the VM is crashing trying to dispatch that event to the Quit handler. I did see the Quit one myself, so probably should be able to get it to fire over again. Both of your traces are really seem to be pointing me to something in the VM being unable to dispatch the Quit handler to the Java handler. First thought that I have is that perhaps I can just map it in Carbon without going through the VM.

The 'rapp' event that you found in your log is probably the one that is causing some of the other "nothing" crashes. I was unaware of it myself...apparently it's a "reopen application" event that is sent by the Finder. That may be the source of some of the other ones that have been happening at random times aside from Quit times.

Thank you very much! I'm a bit busy tonight (packing for travelling tomorrow), so may not be able to get a test library up until Wednesday.

ed
Back to top
Waldo
Oracle


Joined: Dec 03, 2004
Posts: 239

PostPosted: Sun Dec 12, 2004 5:52 pm    Post subject:

ed--

Cool. I'm glad you found it helpful. Sounds like the reopen application event is the culprit for that "did nothing" crash.

Looking forward to the new test library...

W

OPENSTEP wrote:
Thank you very much for those apple event logs. The quit one I did expect...but at the same time it's disturbing that the VM is crashing trying to dispatch that event to the Quit handler. I did see the Quit one myself, so probably should be able to get it to fire over again. Both of your traces are really seem to be pointing me to something in the VM being unable to dispatch the Quit handler to the Java handler. First thought that I have is that perhaps I can just map it in Carbon without going through the VM.

The 'rapp' event that you found in your log is probably the one that is causing some of the other "nothing" crashes. I was unaware of it myself...apparently it's a "reopen application" event that is sent by the Finder. That may be the source of some of the other ones that have been happening at random times aside from Quit times.

Thank you very much! I'm a bit busy tonight (packing for travelling tomorrow), so may not be able to get a test library up until Wednesday.

ed
Back to top
pluby
The Architect
The Architect


Joined: Jun 16, 2003
Posts: 11949

PostPosted: Sun Dec 12, 2004 10:30 pm    Post subject:

I have uploaded a new patch here for you to try:

http://www.planamesa.com/test/testpatch1.tar.gz

Can you see if this patch fixes this bug? To install the patch, execute the following commands in a terminal:

cd /Applications/NeoOfficeJ.app
sudo tar zxvf "/path/to/downloaded/file"

Patrick
Back to top
Waldo
Oracle


Joined: Dec 03, 2004
Posts: 239

PostPosted: Mon Dec 13, 2004 1:43 am    Post subject:

Tried it and messed about. It didn't crash.... but I've only played with it for about 5 or 10 minutes. Let me try and get a crash a bit more and I'll let you know Wink

W


pluby wrote:
I have uploaded a new patch here for you to try:

http://www.planamesa.com/test/testpatch1.tar.gz

Can you see if this patch fixes this bug? To install the patch, execute the following commands in a terminal:

cd /Applications/NeoOfficeJ.app
sudo tar zxvf "/path/to/downloaded/file"

Patrick
Back to top
OPENSTEP
The One
The One


Joined: May 25, 2003
Posts: 4752
Location: Santa Barbara, CA

PostPosted: Mon Dec 13, 2004 9:20 am    Post subject:

Hey Patrick! The hard drive with my build on it is packed in a suitcase right now, but I'll take a look at the code when I get back east. I hadn't been having any luck with these on my main machine, but perhaps my portable with less memory will have a bit more luck.

I do have the AppleEvent tracing stuff, just some putenvs in main(), under a #ifdef DEBUG flag. Do you think it might be useful to check that in?

ed
Back to top
wade1234
Guest





PostPosted: Mon Dec 13, 2004 9:22 am    Post subject:

What does it fix? The quit crash or the do nothing crash?
Back to top
pluby
The Architect
The Architect


Joined: Jun 16, 2003
Posts: 11949

PostPosted: Mon Dec 13, 2004 10:15 am    Post subject:

wade1234 wrote:
What does it fix? The quit crash or the do nothing crash?


The "testpatch1.tar.gz" file should fix the quit crash. Basically, the fix is that I ignore any Apple Events once Neo/J starts shutting down.

Patrick
Back to top
OPENSTEP
The One
The One


Joined: May 25, 2003
Posts: 4752
Location: Santa Barbara, CA

PostPosted: Mon Dec 13, 2004 10:25 pm    Post subject:

I'm going to look more into the "do nothing" crash that I saw since I'm now on a machine with less memory again Smile I'll grab Patrick's latest fixes and see if I can get that to reproduce.

The series of steps I was using was to find a machine with low physical memory, load up a lot of memory hungry programs, open a few large documents, put the app in the background, use the other large apps to swap as much of NeoJ as much as possible onto disk, activate a NeoJ window, then switch applications prior to the menus becoming active. It's convoluted and doesn't always work and definitely doesn't work on my cube. The only difference that I can think of between the two machines is the memory (same updates, etc.). For the activations crashes I'm seeing, that 'rapp' event in the trace above is the most likely candidate as they weren't involving opening docs or the like.

If I can get it to reproduce again, I'm thinking of just installing an empty handler for that event type to ignore it and seeing if that will avoid the dyld signals.

ed
Back to top
Display posts from previous:   
   NeoOffice Forum Index -> NeoOffice Testing All times are GMT - 7 Hours
Goto page 1, 2  Next
Page 1 of 2

 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You cannot download files in this forum

Powered by phpBB © 2001, 2005 phpBB Group

All logos and trademarks in this site are property of their respective owner. The comments are property of their posters, all the rest © Planamesa Inc.
NeoOffice is a registered trademark of Planamesa Inc. and may not be used without permission.
PHP-Nuke Copyright © 2005 by Francisco Burzi. This is free software, and you may redistribute it under the GPL. PHP-Nuke comes with absolutely no warranty, for details, see the license.