Closed jens-maus closed 8 years ago
Originally by AmigaSurfer@t-online.de on 2014-09-04 22:32:34 +0200
Attachment added: winuaelog YAM.debug 2.10dev2014-09-27 Enforcer.zip
(7.8 KiB)
Enforcer-Log
Originally by AmigaSurfer@t-online.de on 2014-09-04 22:33:16 +0200
Attachment added: winuaelog YAM.debug 2.10dev2014-09-27 MuForce MuGuardian.zip
(10.4 KiB)
MuForce MuGuardianAngel Log
Originally on 2014-09-05 08:35:01 +0200
I cannot reproduce this issue. For me that nightly build version starts up without any problem on my WinUAE setup.
Did you try the previous nightly builds between 2014-08-09 and 2014-08-27? They can be downloaded from http://nightly.yam.ch
Originally by AmigaSurfer@t-online.de on 2014-09-08 23:25:28 +0200
I tried YAM.debug from version 2014-08-27 to 2014-08-09 and then 2014-09-06 without WBStartup, but with sashimi, wipeout and winuaeenforcer. The crashes now appears after I accept the beta-requester. 2014-08-27 hangs whole WinUAE while the progress bar shows "Warte auf Benutzeranmeldung" but the buttons for the two accounts are not visible yet. All other versions did crash WinUAE while the progress bar shows "Initialisiere Bibliotheken und Klassen". I had to kill WinUAE with the task manager. 2014-08-09 did not crash. Latest 2014-09-06 did crash. In all cases I quit WinUAE after copying the corresponding YAM.debug and start WinUAE again, to get a clean system with no other program running before.
YAM.debug 2.10dev-2014-08-27 B-Trap fa72 YAM.debug 2.10dev-2014-08-26 B-Trap f9c2 YAM.debug 2.10dev-2014-08-20 B-Trap fc2a YAM.debug 2.10dev-2014-08-19 B-Trap f93a YAM.debug 2.10dev-2014-08-18 B-Trap fe0a YAM.debug 2.10dev-2014-08-17 Illegal instruction 0fbe YAM.debug 2.10dev-2014-08-16 Illegal instruction c87e YAM.debug 2.10dev-2014-08-09 no Crash
YAM.debug 2.10dev-2014-09-06 Exception 3
Then I tried the latest version 2014-09-06 with MuFurce and MuGuardianAngel. I used WinUAE64 for this, to get 2 GByte of FastRAM. After a while there were only 410 MByte FastRAM left. I waited about 15 minutes, but the amount of free RAM didn't change. Except if I close a window at the workbench, then a little bit RAM was freed. Then I quit WinUAE. Maybe YAM.debug was in an endless loop? The sashimi.log is really big (15 MByte)!
Originally by AmigaSurfer@t-online.de on 2014-09-08 23:28:40 +0200
Attachment added: winuaelog YAM.debug 2.10dev-2014-09-08 ohne WBStartup Enforcer Exception 3.zip
(7.8 KiB)
Originally by AmigaSurfer@t-online.de on 2014-09-08 23:29:55 +0200
Attachment added: winuaelog YAM.debug 2.10dev-2014-09-08 ohne WBStartup MuForce.zip
(308.4 KiB)
Originally by AmigaSurfer@t-online.de on 2014-09-08 23:33:36 +0200
I added both logs (enforcer and MuForce) with latest nightly build 2014-09-06. There is an error in the file names. It is 2014-09-06, not 2014-09-08.
Originally by AmigaSurfer@t-online.de on 2014-09-09 00:01:25 +0200
Attachment added: winuaelog YAM.debug 2.10dev-2014-09-06 ohne WBStartup MuFurce ohne Wipeout.zip
(7.7 KiB)
Originally by AmigaSurfer@t-online.de on 2014-09-09 00:06:04 +0200
I remembered now, that you told me some time ago to not use wipeout together with MuForce and MuGuardian. I made a new log. Version 2014-09-06 seems to freeze the mouse pointer now while the progress bar shows "Warte auf Benutzeranmeldung...". I waited some time. Then I opened the WinUAE config window and quit WinUAE64 with F11. This was still working. The new log is attached. There seems to be no error in the log. :(
Originally on 2014-09-09 06:21:18 +0200
Since the nightly build of 2014-08-09 is the last working one I suspect my change done on 2014-08-15 to be the one causing the trouble. It is just a simple one which allows to enter the user password immediately instead of having to activate the string object first manually.
Are you still using MUI 3.8 or MUI 3.9 already? If the former is the case, then this is the proof that MUI 3.8 does not yet support asynchronous methods delayed by a certain amount of time. If you are using MUI 3.9 already then we will have to investigate further.
You should have mentioned that bit of information about the multiple users a bit earlier. Your steps for reproduction are far too scarce in this case. Usually people have just a single user configured in YAM. The information about multiple users is really important here. Please provide such information from the beginning next time.
Originally on 2014-09-09 06:26:11 +0200
Regarding the logs all I can say that there is something really fishy going on on your system again. See this line:
00:D|00:00:00.000000| NewReadArgs.c:441:CS_Buffer[1669537988]: '
'
The number in the squared brackets is the length of either the command line or all ToolTypes of YAM's icon. Usually this is a quite small number, most of the time far less than 1000. In your case it encodes a line consisting of 1.6GB!! This is something that cannot be true.
Originally on 2014-09-09 07:47:58 +0200
In (70483b9):
Originally by AmigaSurfer@t-online.de on 2014-09-10 20:35:39 +0200
Thank you. The new nightly build does not crash at startup. And yes, I'm still using MUI 3.8.
Originally by AmigaSurfer@t-online.de on 2014-09-04 22:32:00 +0200
Summary
Steps to reproduce
Expected results
Requester with message about YAM 2.10 is beta.
Actual results
WinUAE crash
Regression
YAM-dev-20140809 does not crash.
Notes
YAM.debug crash with WInUAEenforcer, too. Whole WinUAE is frozen. The last line in sashimi.log is: 00:D|00:00:00.000000| YAM.c:234:InitLib: library xpkmaster.library v5.2 successfully opened. The last line in winuae.log is: 17-568 [1482 225-072]: B-Trap f058 at 10279e0a (98129E0A)
YAM.debug did not crash with MuForce and MuGuardianAngel. The beta-Requster appears. But then the message, that YAM is out of memory. I have already increased the memory from 512 MByte to 1024 MByte. The sashimi.log shows some errors.
Both logs are attached.