FF4 GA - Is It Stable For You?

classic Classic list List threaded Threaded
36 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Rich Walsh
On Mon, 25 Apr 2011 18:14:48 UTC, James Moe <[hidden email]> wrote:

> On 04/22/2011 01:38 PM, Steve Wendt wrote:
> >
> >> SYS1804: The system cannot find the file DIVE.
> >>    Where can I find a DIVE distribution?
> >
> > If it's not part of the base OS, it's probably part of MMOS2.  I'm very
> > surprised you don't have DIVE.  I'm almost positive it's included in the
> > FixPak.
>
>   I (re-)discovered why <c:/mmos2/dll> is not in the LIBPATH: the WPS
> hangs on startup; PMSHELL starts, WPS does not.
>   I did get Firefox to work, though, by using Rich Walsh's most
> excellent RUN! program. By adding the "e" option and an environment
> setting to add the path to ENDLIBPATH, Firefox started right up.

My _guess_ is that your 'mmpm2.ini' is hosed.  You may want to replace
yours with a virgin copy to see if that permits you to load MMOS2.

For a more immediate fix, just copy 'dive.dll' to some directory on
your LIBPATH.  It doesn't appear to rely on anything else in the
MMOS2 tree.



--
== == almost usable email address:  Rich AT E-vertise DOT Com == ==

_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

mozilla_test
In reply to this post by Rich Walsh
On 21.04.11 05.24, Rich Walsh wrote:
> My impression is that Firefox 4.0 GA is much less crash-prone than any
> of the betas.  So far, the only crashes reported have been in TCPIP32.DLL.
> While any crash is unfortunate, the number and variety of problems seem to
> be significantly reduced.  Is this the case or is it just wishful thinking?

hello rich,
i would like to have a more stable firefox and seamonkey. I'm using
mostly seamonkey and 80% of all crashes are in xul.dll, this will happen
after clicking on a link or if the browser tries to loads a page, one
other crash was in MOZALLOC.DLL and libc. I always have around 30 tabs open.

Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0b13pre) Gecko/20110327 SeaMonkey/2.1b3pre


--
cheers
mozilla_test
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Rich Walsh
On Tue, 26 Apr 2011 21:48:20 UTC, mozilla_test <[hidden email]> wrote:

> i would like to have a more stable firefox and seamonkey.
> [...]
> Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0b13pre) Gecko/20110327 SeaMonkey/2.1b3pre

In that case, you should delete the outdated versions you're using and
get the new ones which are far more stable:

http://releases.mozilla.org/pub/mozilla.org/firefox/releases/4.0/contrib/firefox-4.0.en-US.os2.zip

  Build ID: Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0) Gecko/20100101 Firefox/4.0


http://releases.mozilla.org/pub/mozilla.org/seamonkey/releases/2.1b3-real/contrib/seamonkey-2.1b3.en
-US.os2.zip

  Build ID: Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0pre) Gecko/20110411 Firefox/4.0pre SeaMonkey/2.1b3


--
== == almost usable email address:  Rich AT E-vertise DOT Com == ==

_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

David McKenna
In reply to this post by James Moe-3
James,

On Mon, 25 Apr 2011 11:14:48 -0700, James Moe wrote:

>On 04/22/2011 01:38 PM, Steve Wendt wrote:
>>
>>> SYS1804: The system cannot find the file DIVE.
>>>    Where can I find a DIVE distribution?
>>
>> If it's not part of the base OS, it's probably part of MMOS2.  I'm very
>> surprised you don't have DIVE.  I'm almost positive it's included in the
>> FixPak.
>
>  I (re-)discovered why <c:/mmos2/dll> is not in the LIBPATH: the WPS
>hangs on startup; PMSHELL starts, WPS does not.
>  I did get Firefox to work, though, by using Rich Walsh's most
>excellent RUN! program. By adding the "e" option and an environment
>setting to add the path to ENDLIBPATH, Firefox started right up.
>
>--
>James Moe
>jmm-list at sohnen-moe dot com

  FWIW - I don't know if this applies to your situation, but I recently had a
desktop archive set all of the files it archives to 'read-only'. This caused
the WPS to hang on startup. 3 of these files were in C:\mmos2. When I removed
the 'read-only' setting on them, the WPS would start again. Check *.ini files
in \MMOS2 for 'read-only'. The only one that should be is MMPM.INI (it will
be set that way at WPS start).

Regards,

Dave McKenna



_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Steve Wendt
In reply to this post by Rich Walsh
Rich Walsh wrote:

> So far, the only crashes reported have been in TCPIP32.DLL.

I thought I had been missing out - I finally got one of these.  :-)
You probably aren't too interested in the trap log?
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Dave Yeo-3
Steve Wendt wrote:
> Rich Walsh wrote:
>
>> So far, the only crashes reported have been in TCPIP32.DLL.
>
> I thought I had been missing out - I finally got one of these. :-)
> You probably aren't too interested in the trap log?

I'm curious and have access to crash@e-vertise
Dave
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Dave Yeo-3
In reply to this post by Rich Walsh
Rich Walsh wrote:
> My impression is that Firefox 4.0 GA is much less crash-prone than any
> of the betas.  So far, the only crashes reported have been in TCPIP32.DLL.
> While any crash is unfortunate, the number and variety of problems seem to
> be significantly reduced.  Is this the case or is it just wishful thinking?

I finally got a crash (after the wal fixes etc). It was a SYS3171 with
SeaMonkey pre2.1 while composing a reply to a newsgroup posting.
04-29-2011  15:25:53  SYS3171  PID 0067  TID 0001  Slot 00c2
F:\MOZILLA\SEAMONKEY\SEAMONKEY.EXE
c0000005
1fcb709a
P1=00000002  P2=0002ff88  P3=XXXXXXXX  P4=XXXXXXXX
EAX=00000001  EBX=00000038  ECX=00000038  EDX=20bf8320
ESI=20bf8320  EDI=00000000
DS=0053  DSACC=f0f3  DSLIM=ffffffff
ES=0053  ESACC=f0f3  ESLIM=ffffffff
FS=150b  FSACC=00f3  FSLIM=00000030
GS=0000  GSACC=****  GSLIM=********
CS:EIP=005b:1fcb709a  CSACC=f0df  CSLIM=ffffffff
SS:ESP=0053:0002ff8c  SSACC=f0f3  SSLIM=ffffffff
EBP=00030024  FLG=00012246

PMMERGE.DLL 0004:0010709a

I wonder if the stack size should be bigger? I seem to get sys3171's
mostly when building something (a certain QT solitaire game is notorious
for this on my system).

Dave
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Steven Levine-3
On Sat, 30 Apr 2011 06:45:21 UTC, Dave Yeo <[hidden email]>
wrote:

> I finally got a crash (after the wal fixes etc). It was a SYS3171 with
> SeaMonkey pre2.1 while composing a reply to a newsgroup posting.
> 04-29-2011  15:25:53  SYS3171  PID 0067  TID 0001  Slot 00c2
> F:\MOZILLA\SEAMONKEY\SEAMONKEY.EXE
> c0000005
> 1fcb709a
> P1=00000002  P2=0002ff88  P3=XXXXXXXX  P4=XXXXXXXX
> EAX=00000001  EBX=00000038  ECX=00000038  EDX=20bf8320
> ESI=20bf8320  EDI=00000000
> DS=0053  DSACC=f0f3  DSLIM=ffffffff
> ES=0053  ESACC=f0f3  ESLIM=ffffffff
> FS=150b  FSACC=00f3  FSLIM=00000030
> GS=0000  GSACC=****  GSLIM=********
> CS:EIP=005b:1fcb709a  CSACC=f0df  CSLIM=ffffffff
> SS:ESP=0053:0002ff8c  SSACC=f0f3  SSLIM=ffffffff
> EBP=00030024  FLG=00012246
>
> PMMERGE.DLL 0004:0010709a
>
> I wonder if the stack size should be bigger? I seem to get sys3171's
> mostly when building something (a certain QT solitaire game is notorious
> for this on my system).

This certainly looks like it could be stack overflow, although some
sort of runaway recursion is also a possibility.

I suspect the trap is in WIN32DEFWINDOWPROC which is often the case
when a message handlers recurse unexpectedly.  You should verify this
with the pmmerge.sym for the pmmerge.dll you are running.

Steven

--
---------------------------------------------------------------------
Steven Levine <[hidden email]>
eCS/Warp/DIY etc. www.scoug.com www.ecomstation.com
---------------------------------------------------------------------
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Rich Walsh
In reply to this post by Dave Yeo-3
On Sat, 30 Apr 2011 06:45:21 UTC, Dave Yeo <[hidden email]> wrote:

> I finally got a crash (after the wal fixes etc). It was a SYS3171 with
> SeaMonkey pre2.1 while composing a reply to a newsgroup posting.
>
> 04-29-2011  15:25:53  SYS3171  PID 0067  TID 0001  Slot 00c2
> F:\MOZILLA\SEAMONKEY\SEAMONKEY.EXE
> c0000005
> 1fcb709a
> P1=00000002  P2=0002ff88  P3=XXXXXXXX  P4=XXXXXXXX
> [snip]
> SS:ESP=0053:0002ff8c  SSACC=f0f3  SSLIM=ffffffff
>
> PMMERGE.DLL 0004:0010709a
>
> I wonder if the stack size should be bigger?

I don't think so.  I looked at most of the Exceptq reports we've gotten
and none of them showed more than about 200k of stack usage for TID 1.
I think Steve is right that this is a case of runaway recursion.

If you'd like to see what the typical maximum stack usage is for SM,
add the following function to src\suite\app\nsSuiteApp.cpp along with
the stuff marked with a '+'.  The figure will show up on the console
after you close SM.  E.g.  Max stack usage:  0x3200  (200K)


+#define INCL_DOS
+#include <os2.h>

void    GetMaxStack(void)
{
  APIRET    rc;
  ULONG     ulSize;
  ULONG     ulState;
  PTIB      ptib;
  PPIB      ppib;
  char *    pStackPtr;

  DosGetInfoBlocks(&ptib, &ppib);

  /* Find accessible stack range */
  for (pStackPtr = ((char*)ptib->tib_pstacklimit - 0x1000);
       pStackPtr >= (char*)ptib->tib_pstack;
       pStackPtr -= 0x1000) {
    ulSize = 0x1000;
    rc = DosQueryMemState(pStackPtr, &ulSize, &ulState);
    if (rc || ~ulState & PAG_PRESENT) {
      break;
    }
  }
  pStackPtr += 0x1000;

  ulSize = (ULONG)ptib->tib_pstacklimit - (ULONG)pStackPtr;
  printf("Max stack usage:  0x%X  (%uK)\n",
         (unsigned)ulSize, (unsigned)(ulSize/1024));
  return;
}

int main(int argc, char* argv[])
{
  ScopedLogging log;
  [...]
  int result = XRE_main(argc, argv, appData);
  XRE_FreeAppData(appData);
+ GetMaxStack();
  return result;
}


--
== == almost usable email address:  Rich AT E-vertise DOT Com == ==

_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Dave Yeo-3
In reply to this post by Steven Levine-3
Steven Levine wrote:

>> PMMERGE.DLL 0004:0010709a
>> >
>> >  I wonder if the stack size should be bigger? I seem to get sys3171's
>> >  mostly when building something (a certain QT solitaire game is notorious
>> >  for this on my system).
> This certainly looks like it could be stack overflow, although some
> sort of runaway recursion is also a possibility.
>
> I suspect the trap is in WIN32DEFWINDOWPROC which is often the case
> when a message handlers recurse unexpectedly.  You should verify this
> with the pmmerge.sym for the pmmerge.dll you are running.

I notice that the PMMERGE address is in a multiple SYS3171's in
popup.log though not all, 0004:00106447 is also common (not SeaMonkey).
How do I find the function in the sym file?
Dave
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Dave Yeo-3
In reply to this post by Rich Walsh
Rich Walsh wrote:
>> I wonder if the stack size should be bigger?
> I don't think so.  I looked at most of the Exceptq reports we've gotten
> and none of them showed more than about 200k of stack usage for TID 1.
> I think Steve is right that this is a case of runaway recursion.
>
> If you'd like to see what the typical maximum stack usage is for SM,
> add the following function to src\suite\app\nsSuiteApp.cpp along with
> the stuff marked with a '+'.  The figure will show up on the console
> after you close SM.  E.g.  Max stack usage:  0x3200  (200K)

Unluckily today's build of SeaMonkey hangs on exit so all I get is the
stack usage of the profile manager, 44 K.
Dave
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Steven Levine-3
In reply to this post by Dave Yeo-3
On Sun, 1 May 2011 04:40:45 UTC, Dave Yeo <[hidden email]>
wrote:

Hi,

> I notice that the PMMERGE address is in a multiple SYS3171's in
> popup.log though not all, 0004:00106447 is also common (not SeaMonkey).
> How do I find the function in the sym file?

Use sym.exe from the exceptq package to dump the .sym file in a human
readable form.  For .xqs files, use mapxqs with the -d option.

Steven

--
---------------------------------------------------------------------
Steven Levine <[hidden email]>
eCS/Warp/DIY etc. www.scoug.com www.ecomstation.com
---------------------------------------------------------------------
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Rich Walsh
On Sun, 1 May 2011 14:06:50 UTC, "Steven Levine" <[hidden email]> wrote:
> On Sun, 1 May 2011 04:40:45 UTC, Dave Yeo <[hidden email]> wrote:
>
> > I notice that the PMMERGE address is in a multiple SYS3171's in
> > popup.log though not all, 0004:00106447 is also common (not SeaMonkey).
> > How do I find the function in the sym file?
>
> Use sym.exe from the exceptq package to dump the .sym file in a human
> readable form.  For .xqs files, use mapxqs with the -d option.

FYI... sym.exe can be found in 'legacy.zip' in the Exceptq developer's
package (exceptq71-dev.zip).



--
== == almost usable email address:  Rich AT E-vertise DOT Com == ==

_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Rich Walsh
In reply to this post by Dave Yeo-3
On Sun, 1 May 2011 06:35:49 UTC, Dave Yeo <[hidden email]> wrote:
> Rich Walsh wrote:

> >> I wonder if the stack size should be bigger?
> >
> > I don't think so.  I looked at most of the Exceptq reports we've gotten
> > and none of them showed more than about 200k of stack usage for TID 1.
> > I think Steve is right that this is a case of runaway recursion.
> >
> > If you'd like to see what the typical maximum stack usage is for SM,
> > add the following function to src\suite\app\nsSuiteApp.cpp along with
> > the stuff marked with a '+'.  The figure will show up on the console
> > after you close SM.  E.g.  Max stack usage:  0x3200  (200K)
>
> Unluckily today's build of SeaMonkey hangs on exit so all I get is the
> stack usage of the profile manager, 44 K.

To test that code, I modified the source I used for the my last SM beta,
then built just seamonkey.exe and used it to replace the exe in Walter's
"official" release.  nsSuiteApp.cpp hasn't (hadn't?) changed in a long
time, so the new exe is probably compatible with any relatively recent
version.

BTW...  I left out a zero in the sample output - it should have been
  Max stack usage:  0x32000  (200K)


--
== == almost usable email address:  Rich AT E-vertise DOT Com == ==

_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

Dave Yeo-3
Rich Walsh wrote:

> On Sun, 1 May 2011 06:35:49 UTC, Dave Yeo<[hidden email]>  wrote:
>> Rich Walsh wrote:
>
>>>> I wonder if the stack size should be bigger?
>>>
>>> I don't think so.  I looked at most of the Exceptq reports we've gotten
>>> and none of them showed more than about 200k of stack usage for TID 1.
>>> I think Steve is right that this is a case of runaway recursion.
>>>
>>> If you'd like to see what the typical maximum stack usage is for SM,
>>> add the following function to src\suite\app\nsSuiteApp.cpp along with
>>> the stuff marked with a '+'.  The figure will show up on the console
>>> after you close SM.  E.g.  Max stack usage:  0x3200  (200K)
>>
>> Unluckily today's build of SeaMonkey hangs on exit so all I get is the
>> stack usage of the profile manager, 44 K.
>
> To test that code, I modified the source I used for the my last SM beta,
> then built just seamonkey.exe and used it to replace the exe in Walter's
> "official" release.  nsSuiteApp.cpp hasn't (hadn't?) changed in a long
> time, so the new exe is probably compatible with any relatively recent
> version.
>
> BTW...  I left out a zero in the sample output - it should have been
>    Max stack usage:  0x32000  (200K)
>
>

Reverting to my last build and using the new SeaMonkey.exe worked. The
hang may be flashgot related as I now see in the log, "FLashGotDM is not
defined"
After only a few minutes of running the stack is 0xE000 (56K).
I'll let it run awhile and recheck
Dave
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
Reply | Threaded
Open this post in threaded view
|

Re: FF4 GA - Is It Stable For You?

mozilla_test
In reply to this post by Rich Walsh
On 27.04.11 02.09, Rich Walsh wrote:

hello rich,

>    Build ID: Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0) Gecko/20100101 Firefox/4.0
>
>
> http://releases.mozilla.org/pub/mozilla.org/seamonkey/releases/2.1b3-real/contrib/seamonkey-2.1b3.en
> -US.os2.zip
>
>    Build ID: Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0pre) Gecko/20110411 Firefox/4.0pre SeaMonkey/2.1b3

exactly these 2 builds showed the same behavior on another machine. Most
of the files in the seamonkey package have an old time stamp from 2010
in case its a newer build.
On the main machine i have now Mozilla/5.0 (OS/2; Warp 4.5; rv:2.0pre)
Gecko/20110411 SeaMonkey/2.1b3 and its still running without crashes
even with heavy usage. the things i noticed are still that the download
dir doesnt exist and thus downloads wont work, other is related to webm
that most of the time the sound init will fail, moving the slider back
to start will then enable sound (on ytube).

--
cheers
mozilla_test
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2
12