Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Dariusz Piatkowski

Pages: [1] 2 3 ... 81
1
Programming / Re: WebSite / HTML editor - what do we have?
« on: February 05, 2023, 04:05:57 pm »
Dave, everyone...
Hi Dariusz, since everyone else is running it fine and it only fails for you when Firefox is open and you're using LIBPATHSTRICT, it seems the most likely is resource starvation, and memory makes sense.
...

Yup!

Messing around with VIRTUALADDRESSLIMIT (by making it larger) gets me a working Kompozer install here. Granted, this is only a half-win at best because whatever amount I increase that setting by is certainly taking away from the fs cache, and I'm sure not going to give up my giant JFS cache...LOL!

Anyways...I have my answer though.

So I'm curious: is no one else running this with LIBPATHSTRICT setting like I am?  :o

Heck, maybe the simplest solution yet - which I haven't actually tried given that the DLLs between Kompozer and our current FF would be so different - would be to drop all the provided DLLs and try to use the FF ones?

OK..another test, here we go...  8)

2
Programming / Re: WebSite / HTML editor - what do we have?
« on: February 04, 2023, 07:25:25 pm »
Dave,
All my libcxxx DLLs are sitting in the \usr\lib directory, no dups anywhere else. I did also try setting different load options for LIBC, but none of them resulted in any improvements.

Maybe VIRTUALADDRESSLIMIT is the key, worth a try I suppose...

Ivan,
The issue I'm seeing with Kompozer ONLY happens when FF is already loaded. Since I have Thunderbird deployed as well here, there are some pretty heavy resource demands, so I suspect I'm actually running into some kind of a conflict given that Kompozer is based on the Mozilla framework as well.

3
Programming / Re: WebSite / HTML editor - what do we have?
« on: February 04, 2023, 06:41:21 pm »
Dave,

As I said, "set BEGINLIBPATH=.\."
...
Should do it. If it tries to update, stop it. Also make sure the current directory is G:\code\tools\kompozer\

Hmm, interesting. Firing up the app I see it close immedately and get the following in the LOG:

Code: [Select]
Killed by SIGSEGV
pid=0x015f ppid=0x001f tid=0x0001 slot=0x010e pri=0x0200 mc=0x0001 ps=0x0010
G:\CODE\TOOLS\KOMPOZER\KOMPOZER.EXE
MOZILLA 1:000fd718
cs:eip=004a7a38:00aad718      ss:esp=0000:015423a0      ebp=00000053
 ds=0000      es=0000      fs=0006      gs=0000     efl=00d67e40
eax=00aad49c ebx=1ffc9d7c ecx=00aad4b0 edx=00aad4d4 edi=00000000 esi=00aaff7c
Creating G:\TMP\TRP\015F_01.TRP

...and the TRP file itself shows this (just the front end of it for brevity):

Code: [Select]
______________________________________________________________________

 Exception Report - created 2023/02/04 12:08:13
______________________________________________________________________

 LIBC: Killed by SIGSEGV

 Hostname:         NEUROBOX
 OS2/eCS Version:  2.45
 # of Processors:  6
 Physical Memory:  3199 mb
 Virt Addr Limit:  2048 mb
 Exceptq Version:  7.11.5-shl BETA8 (Jun  1 2020 18:37:02)

______________________________________________________________________

 Exception C0000005 - Access Violation
______________________________________________________________________

 Process:  G:\CODE\TOOLS\KOMPOZER\KOMPOZER.EXE (12/23/2011 09:59:15 6,412,142)
 PID:      15F (351)
 TID:      01 (1)
 Slot:     10E (270)
 Priority: 200

 Module:   LIBCN0
 Filename: G:\USR\LIB\LIBCN0.DLL (08/26/2021 09:13:22 1,266,106)
 Address:  005B:00D7C9E3 (0000:FFFFFFFF)
 Cause:    Attempted to write to 03503462
           (not a valid address)
 Code:     failing instruction can not be disassembled

______________________________________________________________________

 Registers
______________________________________________________________________

 EAX : 00AAD600   EBX  : 00AAD718   ECX : 01638F14   EDX  : 01522EC4
 ESI : 00AAD718   EDI  : 004A7A38
 ESP : 00AAD604   EBP  : 00AAD620   EIP : 00D7C9E3   EFLG : 00010206
 CS  : 005B       CSLIM: FFFFFFFF   SS  : 0053       SSLIM: FFFFFFFF

 EAX : read/write memory on this thread's stack
 EBX : read/write memory on this thread's stack
 ECX : read/write memory allocated by LIBCN0
 EDX : read/write memory allocated by LIBCN0
 ESI : read/write memory on this thread's stack
 EDI : read/exec  memory at 0001:00497A38 in MOZILLA

______________________________________________________________________

 Stack Info for Thread 01
______________________________________________________________________

   Size       Base        ESP         Max         Top
 00100000   00AB0000 -> 00AAD604 -> 00AA8000 -> 009B0000

______________________________________________________________________

 Call Stack
______________________________________________________________________

   EBP     Address    Module     Obj:Offset    Nearest Public Symbol
 --------  ---------  --------  -------------  -----------------------
 Trap  ->  00D7C9E3   *Unknown*

 00AAD620  0052938A   MOZILLA   0001:0051938A
....


I've marked tke Kompozer DLLs as loadhigh (code only), tried other options but those resulted in errors as well.

Now the key issue here is that this result is DIFFERENT when FF isn't already running. In that case Kompozer opens up fine, stays running and only tosses an error when I actually close the app.

I'm doing some DLL load tracing to figure out if Kompozer is pulling another DLL that might not exist in it's own directory and may cause a conflict. I found distorm.dll so far that sits in \usr\lib and that I do not recall what it's used for.

"yum provides " returns nothing either...

EDIT => distorm.dll is provided by EXCEPTQ package, I was using our "\" instead of their "/"...

4
Wow...I never knew we had 'conspiracy theorists' and 'against all I'm unfamiliar with' thinkers on our board! ...how the heck have you survived all these years???!!!

LOL, maybe I should have know that though?  :-\ ...but why the "sales pitch" accusations???

Regardless whether JTA is a real person or not, I think the posts provided clearly identified the thought process behind pushing OS/2 into a VM, and how that in turn would remove the 'hardware suitability' from the equation we all try to solve, all in a relatively self-contained package.

Anyways, it is good to have options while the standard blockers are being removed slowly, one piece at a time. 

The browser problem is the one thing, I fear, that has the real "stones" to hurt us badly. Superb efforts are ongoing to extend our viability, but I still struggle with why this hasn't become a prioritized deliverable for our platform - and by that I mean the apparent lack of official backing by the only commercial player in our space, that being AN.

5
Programming / Re: WebSite / HTML editor - what do we have?
« on: February 02, 2023, 05:03:33 am »
Dave!

Thanks for the uber quic response to this!

Kompozer runs fine here using run!, Perhaps your cmd also needs set BEGINLIBPATH=.\. Also try redirecting output to see any other errors.
There's also SeaMonkey and its composer, which Komposer was based on. Hasn't had many updates in a long time but does what it does well.

Well, here is what I actually have in my CMD file:

Code: [Select]
SET LIBPATHSTRICT=T
rem the following may be needed if another Mozilla based app is in LIBPATH
SET BEGINLIBPATH=G:\code\tools\kompozer
REM SET MOZILLA_HOME=<profile_path>
G:\code\tools\kompozer\kompozer.exe

I've rarely utilized BEGINLIBPATH, so how should I use the ".\." that you mentioned in your post Dave?

I will try the runtime re-direct next to see what's actually going on...I suspect I am running into some DLL conflict?

6
Programming / WebSite / HTML editor - what do we have?
« on: February 02, 2023, 04:46:28 am »
...as the title asks: what exactly do we have on our platform?

Dave Yeo ported Kompozer sometime back in 2018 I think...but tonight I just tried it and while it starts up it then completely dies. It is Mozilla based, the full set of DLLs comes with it, and well it looks like it's not quite compatible with our FF 45.9x release because I had apparently tried this before and ended up setting up a CMD that uses "SET LIBPATHSTRICT=T".

Is there anything else I can look at?

I have my VSE here, but that's all HTML code and I'd rather be lazy and do some WYSIWYG instead (I've got a couple of very simple index.html pages to build).

7
Applications / Re: Porting question
« on: February 01, 2023, 04:09:53 am »
Ugh...  I remember again why I'm not a programmer.....   >:(

Well, let's not get discouraged far too easily and/or quickly...there are resources out there.

Take a look at https://archive.org/search?query=os2, pick a book that will ease you into the OS/2 DEV game...say something like:

1) "OS/2 Warp Programming For Dummies"
- don't let the title turn you off either, we all start somewhere

2) "OS/2 Presentation Manager Programming" by Charls Petzold
- this will really do a nice job of walking you through every piece of PM API: building controls, windows, putting content into the window, manipulating it, etc.

3) "OS/2 Warp Presentation Manager for Power Programmers"
- in the meantime this little gem will step into the more advanced features, stuff you might not readily touch each and every time, it spends a good amount of time on each of the building blocks of PM programming: menus, notebooks, even stuff like inter-process communciations, etc.

4) "The Art of OS/2 Warp Programming"
- this is the most up-to-date book AFAIK, the original release covered OS/2 2.1, this one is updated to cover V4 and does a very good job of reviewing everything

Once you understand the OS/2 framework itself you are going to be in a much better position attempting to do some porting, where inherently you'll need to be able to map that "other" framework to our OS/2 model.

Hopefully, if you are lucky, your starting point in the porting effort may be pretty close, and therefore require very little "elbow grease".

8
Applications / Re: ANPM - YUM error. [Errno 256] No more mirrors to try
« on: January 28, 2023, 11:20:11 pm »
Martin,

It looks like netlabs repo is off line, you could always try the AOS mirror => "https://www.2rosenthals.com/rpm.netlabs.org/release/$releasever/$basearch/"


9
Hardware / Re: Some successful installs with ArcaOS 5.0
« on: January 28, 2023, 11:14:53 pm »
It might depend on the controller. I used to have an Lenovo Thinkpad x121e. I needed to switch the SATA Controller to compatible mode as in ahci mode no drive was found using ArcaOS 5.04 (i do not have this Laptop anymore so i couldn't test with more recent versions.)

On my motherboard, which is a AMD 890FX and SB850 chipset, I couldn't get the AHCI driver to work reliably. It would find the SSD, no problem there, it would access it just fine...however periodically the machine would literally freeze...no response, everything stopped, this would last for about 5-10 secs, after which time it would then proceed to run once again.

I tried to narrow this down to either my JFS cache settings, or driver parameters, heck, even played with the motherboard ports I was connecting to...never established the root cause, so I gave up on trying to get AHCI working here.

As to the speed comment...I'm not sure that it's any faster actually. The testing I was able to do showed exactly the same results, however, I will say that the system 'response' seems to be improved, and that's precisely what AHCI can deliver, after all it re-arranges the drive commands in such a way that the hardware itself delivers the information faster. This is normally very pronounced when dealing with spinning disks since the amount of head movement is minimized, and that in turn allows for a more efficient retrieval of data from the disk. SSDs on the other hand don't seem to benefit from this as much, at least that was my experience while I was playing around with it.

Bottom line: you need to test your hardware, and if you are running AOS and encountering problems do report this to AN by logging an Issue, it's the only way to improve the software by making the authors aware of the issue.

10
Hardware / Re: Some successful installs with ArcaOS 5.0
« on: January 28, 2023, 11:06:02 pm »
...
ArcaOS 5.07 is not the holy Grail, we have OS/2, eComstation and Arcaos Users here. Any Hardware Information is welcome!
...
For myself I am using a pimped up OS/2 4.52 Version on Hardware, Arcaos 5.07 wont even Start the Installation. But that does not mean, that my OS/2 is the best around.

This Forums live because we have all flavours of OS/2 here, ArcaOs is just one of them, the most modern and the one I would suggest for a new User.

I couldn't have said it any better Sigurd.

The 'ArcaOS or Bust' mantra is a perfect example of what I would call "group think". Don't get me wrong, if you are NEW to OS/2 platform AOS is probably THE way to get your feet wet...heck, not even "probably", I would in fact say it is THE way to do so!

However...plenty of us here are running on old hardware (as well as some new stuff) and using the older versions of OS/2...most likely Warp4 derived/based stuff though.

Therefore, I would also encourage the sharing of your experience...others may, and most probably will, benefit from it, any by chance you may in fact get some hints and pointers on how to address whatever issue you may be running into!

11
Applications / Re: Creating an OS/2 image from HDD
« on: January 23, 2023, 05:20:01 am »
Hi Dave,

...Which reminds me Dariusz, you posted a bunch of chkdsk flags earlier, could you repost them?

here is what I have:

Code: [Select]
Check FS content
 
  Usage: chkdsk drive: [/A] [/B] [/C] [/D] [/F[:n]] [/O] [/P] [/V]
 
    /A      Enable autocheck mode (undocumented)
    /B      Perform LVM Bad Block List utility processing only
    /C      Process ifdirty after logredo
    /D      Enable debug mode (undocumented)
            all log messages written to stdout in english
    /F[:n]  Full check
            0 => read only (default)
            1 => prompt for permission for each repair
            2 => preen (default if n omitted)
            3 => Assume permission given for all repairs.
                 If exiting with clean file system, invoke LVM Bad Block
                 List utility processing on the way out.
            Note: :1 and :2 are converted to :3
    /O      Omit logredo (undocumented)
    /P      Enable pmchkdsk mode (undocumented)
    /V      Verbose messaging (ignored)
 
  Both / and - prefix supported
  Options are case-insenitive
 
  Called as
 
    chkdsk drive: /F /A [/C]
 
  for autocheck

12
Applications / Re: Creating an OS/2 image from HDD
« on: January 23, 2023, 12:26:02 am »
Hi ivan,

Have you considered creating two partitions on your new enlarged SSD leaving your boot partition at 200 Gb (which is excessive for a boot partition, mine is 2 GB with over 1 GB still free) and putting tools, programs, etc on the remaining partitioned space?

What I currently have is the legacy of having installed my original OS/2 Warp3 way back when in the early 1990s I think...LOL...yeah, it's been that long!

Several upgrades later (such as Warp4 for example), with all of my apps having been installed on the same drive as the boot drive I simply do not have the desire to split into the arguably much better approach you suggest.

I figured that if AOS is ever released in UPGRADE mode I'll take that plunge, until then I simply manage what I have.

Therefore in my case I could replicate the 200G drive in as-is layout, and just have the extra 300G left for other uses, maybe move some storage from the boot drive to that one?

The challenge/problem with splitting the SSD into such partitions is that I'm not certain it actually gains anything in terms of performance. It is still the same hardware and having the extra empty space on a SSD is actually a good thing (think overprovisioning).

As is, I will have my current 200G drive which will become that storage space and it being a separate hardware device allocates the full channel to just those disk IOps.

13
Applications / Re: Creating an OS/2 image from HDD
« on: January 22, 2023, 07:26:34 pm »
This may be a silly question you guys, but can the DFSee created image (is it DISK or PARTITION?) be restored onto a differently sized TARGET?

What I mean is the following:

1) I have a 200G SOURCE (current OS/2 partition on a SDD)
2) I have a 500G TARGET (future OS/2 partition on a NEW SDD)

Can I used DFSee image of the SOURCE to "move" my whole system to the ned SDD?

FYI - I have been looking at doing either the old fashioned XCOPY thing, or RSYNC, or DFSee...was aware that some such DFSee functionality existed, but I haven't investigated any of it. Perhaps DFSee could actually do a real-time copy???

Thanks!

14
Storage / Re: JFS - chkdsk LOG - where?
« on: January 14, 2023, 11:14:05 pm »
Hi Dave!

Yup, booted to my MAINT partition and manually ran a CHKDSK against all the JFS drives. Only one reported errors, and that was the backup of the MAINT partition, so low risk.

Given that I do nightly backups I was able to run a quick compare between the drive listings afterwards and narrow this down to about 10 different files (all over the place though, so not just a single directory).

Anyways, that allowed me to fix the issues and get my machine back up to a fully functioning status.

But man...not to have visibility to what JFS CHKDSK does is a scary feeling...why isn't it producing a CHKDSK.LOG like HPFS did? Strange since running 'chkdks drive: /p' will even show the MSG statements for each check operation and that's primarily the sort of stuff that HPFS used to log into the LOG file.

15
Storage / JFS - chkdsk LOG - where?
« on: January 14, 2023, 06:58:33 pm »
Hi Everyone,

Umm...something strange happened to me this morning: I was editting my RSYNC LOG file so I could archive the 2022 stuff and start new for 2023. Needless to say, the file is about 25M in size...so plain vanilla editor just takes a little extra time.

I had done this with the good ol' trusty E.exe before, but I figured I would give AE a try (v1.9.x). Editing went fine, no faster than E.exe, however once edited EA hung trying to save it back to the disk. I waited, and waited, following which I tried to kil EA, no go though. Rolled out the full gamut of kill options, still no go.

Ultimately I selected the Reboot option from CAD-Popup as EA was literally blocking WPS (nothing else worked), but even than hung. So as a final attempt to recover I did CTL-ALT-NUMLCK x2, which put me in the dump screen, and then I reset.

Now, in the past, that desparate attempt to recover has always worked. It also has NEVER caused me a JFS issue, however this time...man...upon reboot the JFS CHKDSK went to work full-time...pile of stuff scrolled by (too fast of course to see any of it) and machine came back up. INIs were corrupted, no problem, I have regular backups, so a couple of reboots later I was back up and running.

However, not liking that longish list of CHKDSK results I went fishing to learn more about it.

As it happens I have the following in my STARTUP.CMD:

Code: [Select]
...
chklgjfs g: >> g:\tmp\log\chklgjfs-g.log
...

which is done for ALL the drives.

In my case, this produced the following:

Code: [Select]
____________________________________________________________________
chkdsk service log selected:  MOST RECENT

XCHKLOG P superblock is valid.


**********************************************************


CHKDSK  processing started: 1/14/2310.56.1

CHKDSK parameter detected:  AutoCheck mode

CHKDSK parameter detected:  IfDirty

CHKDSK  The current hard disk drive is:  g:

CHKDSK  DosOpen(...OPEN_SHARE_DENYREADWRITE|OPEN_ACCESS_READWRITE...) returned rc = 0

CHKDSK  DosDevIOCtl(...DSK_GETDEVICEPARAMS...) returned rc = 0

CHKDSK  Primary superblock is valid.

CHKDSK  The type of file system for the disk is JFS.

CHKDSK  The boot sector has been refreshed.

CHKDSK  Filesystem is clean.

CHKDSK  All observed inconsistencies have been repaired.

CHKDSK  Filesystem has been marked clean.

CHKDSK **** Filesystem was modified. ****

CHKDSK   LVM GetBadBlockInformation returned rc = 22

CHKDSK  processing terminated:  1/14/2310.56.1   with return code: 0.

____________________________________________________________________

The problem here is that this doesn't tell me anything about WHAT actually was detected and what JFS did to attempt to correct it.

Sure enough...as I attempted to fire up a few of my regular apps like PMMail for example, I came to discover that those directories have missing files, etc.

So clearly there is a JFS problem here, and as opposed to HPFS where the HPFS CHKDSK log viewer would at least clearly show the objects that were impacted, it would appear that JFS has no such functionality.

How can that be though? I must be missing something...

So how are you guys getting visibility into what JFS CHKDSK log has found and/or corrected?

BTW: Nothing in \lost+found directory, so no actual directories or files were recovered...

Pages: [1] 2 3 ... 81