Author Topic: Firefox and Traps  (Read 57223 times)

Ben

  • Guest
Re: Firefox and Traps
« Reply #45 on: April 12, 2018, 01:26:43 am »
OK, Dave.

I'll utilize that information and adjust all of my server startup cmds tomorrow. While I do not think that it will solve the Firefox(s) issue, it will be a step in the right direction.

Maybe by the time I get that done I will have come up with something else; *ponders* if I could get BS_info to behave I could track down any duplicate Dlls.

I wonder if the two problems are related... *shrugs*


Pete

  • Hero Member
  • *****
  • Posts: 1294
  • Karma: +9/-0
    • View Profile
Re: Firefox and Traps
« Reply #46 on: April 12, 2018, 03:32:00 am »
Hi Ben

What problem(s) you are having with BS-Info? - and which version is involved?


Regards

Pete

Andi B.

  • Hero Member
  • *****
  • Posts: 817
  • Karma: +11/-2
    • View Profile
Re: Firefox and Traps
« Reply #47 on: April 12, 2018, 09:04:02 am »
Quote
One suggestion is to use cmd files to set up your environment for all your servers. Something like, and this is guess work about your programs,
set PATH=c:\weasel;%PATH%
set BEGINLIBPATH=c:\weasel;%BEGINLIBPATH%
set LIBPATHSTRICT=T
weasel.exe

Adjusting as needed for your various services. This should help with DLL hell

This may help for some situations. But I fear in Bens case it will make things worse. LIBPATHSTRICT and BEGINLIBPATH opens another can of worms. Limited resources (for todays needs) our OS can provide and handle is one of them. The list of services Ben wants to run is much over the limit LIBPATHSTRICT can work reliable IMHO. My advice is using LIBPATHSTRICT only for a single or two applications if really needed (running SM and FF in parallel for testing purposes). Not in any case for more than a few programs.

Ben has duplicate/wrong dlls. Probably spread all over his drives. He need to clean up this mess. One thing he can do is use f.i. DataSeeker (or another search tool which searches all drives) and search for every single dll FF needs with 'Search all drives' selected. And rename/delete all duplicates on all drives. Tools which only check the LIBPATH for dublicates (XWP or 'which' f.i.) do not help in Bens case as he has the .; in the LIBPATH and fear he does not want to delete it from there.

Lars

  • Hero Member
  • *****
  • Posts: 1272
  • Karma: +65/-0
    • View Profile
Re: Firefox and Traps
« Reply #48 on: April 12, 2018, 12:42:49 pm »
Hm,

I am beginning to believe that Andi is right about the problems with LIBPATHSTRICT.

there exists a kernel data structure called "PTDA" (per task data area") that effectively exists for each process started.
Within that structure there are 2 pointer fields that can take a pointer to the BEGINLIBPATH and ENDLIBPATH strings respectively.
This effectively means that each process can define its own BEGINLIBPATH and ENDLIBPATH. However, I can see no such field for the LIBPATHSTRICT setting (which can take on 2 values: TRUE and FALSE).

That leads to the conclusion that there must be a GLOBAL list somewhere in the kernel that remembers which process asked for LIBPATHSTRICT in which case, when the process execution leads to loading of DLLs, the process will, if LIBPATHSTRICT is TRUE, preferrably load from its BEGINLIBPATH setting before it loads from any location defined in LIBPATH even if a DLL of the same name has already been loaded (and is still loaded) from a different directory.

A GLOBAL list will likely have a limit ...

Lars

Andi B.

  • Hero Member
  • *****
  • Posts: 817
  • Karma: +11/-2
    • View Profile
Re: Firefox and Traps
« Reply #49 on: April 12, 2018, 01:19:31 pm »
Additionally to Lars statement about kernel internals it's simply a waste of memory (= scare resource given todays applications) when 'same' dlls are loaded more then once via LIBPATHSTRICT.

Practical experiments showed me that you can easily bring down OS/2 when playing with LIBPATHSTRICT and BEGINLIBPATH.

Ben

  • Guest
Re: Firefox and Traps
« Reply #50 on: April 12, 2018, 08:27:21 pm »
OK.

So, from all that I see clearly that this is a mega-mess with no solution... other than to comb through decades of DLLs keeping the latest copies and deleting the deprecated duplicates... hoping that all will be well when done.

YIKES!

While I have already spent more than a few hours doing that, at this point I must seriously consider other possibilities such as, is there another usable browser out there that works with OS/2 and doesn't have similar requirements to Firefox, (old or new), and Seamonkey etc?

But, is that even viable at all given the new trend in OS/2 with regards to installing new software?

With all the services that I am running as suggested, I do not have a lot of free, usable memory, (though I have 4 gigs of RAM it is a struggle to keep enough available for apps and tools), so I cannot afford to be loading double versions of DLLs.

In the meantime, I will continue to hunt for duplicates.


Ben

  • Guest
Re: Firefox and Traps
« Reply #51 on: April 12, 2018, 08:30:17 pm »
Hi Ben

What problem(s) you are having with BS-Info? - and which version is involved?

Pete, on the server, it starts fine and displays the first tab with the clock, but after about five seconds it hangs, maxes out one core, (I have four), and has to be terminated.

It works fine on my other machine.

I downloaded the latest version from the link that you posted. 2.07?

Pete

  • Hero Member
  • *****
  • Posts: 1294
  • Karma: +9/-0
    • View Profile
Re: Firefox and Traps
« Reply #52 on: April 13, 2018, 01:09:00 am »
Hi Ben

That sounds like the problem I saw with an earlier version when using os2ahci.add v2.x

Is there anything displayed on the status bar that may give us a clue?

You could also add the Parameter -debug to the BS-Info object Properties which should create a log called BS_Info.err in the BS-Info directory which may help identify the problem.


Regards

Pete

Ben

  • Guest
Re: Firefox and Traps
« Reply #53 on: April 13, 2018, 03:10:22 am »
Pete, the status bar says "cups...".

I attached the debug file.
« Last Edit: April 13, 2018, 11:43:01 am by Ben »

Pete

  • Hero Member
  • *****
  • Posts: 1294
  • Karma: +9/-0
    • View Profile
Re: Firefox and Traps
« Reply #54 on: April 13, 2018, 05:21:21 am »
Hi Ben

The debug log is called BS_Info.err

Re-checking the version I have installed I see that I am using v2.1 beta which Rolf sent me by email. It has a reworked cups search which may resolve the problem you are seeing - but does take a second or several to appear onscreen.

The v2.1 beta does not yet seem to have made it to a release level so I hope Rolf does not mind me posting the beta location www.bausys.ch\files\bs_info_2_1_beta.wpi

I am sure Rolf would appreciate any feedback though.


Regards

Pete




Ben

  • Guest
Re: Firefox and Traps
« Reply #55 on: April 13, 2018, 11:42:34 am »
OK, thx, Pete.

If such an error  file was generated it was not put into the BS Info folder.

I will give the new one a try.

Time passes...

I installed 2.1b and tried it with the same results; it hangs on cups detection.

Pete

  • Hero Member
  • *****
  • Posts: 1294
  • Karma: +9/-0
    • View Profile
Re: Firefox and Traps
« Reply #56 on: April 13, 2018, 02:12:56 pm »
Hi Ben

Strange...

Do you have the Parameter set in the program object properties? It should be

   -debug


Here that setting definitely causes BS_Info.err to be generated in the BS_Info directory. I just ran BS_Info and generated a log so have no idea what could be going wrong on your system. I can only suggest contacting Rolf and reporting problems.


Regards

Pete





Ben

  • Guest
Re: Firefox and Traps
« Reply #57 on: April 13, 2018, 02:31:44 pm »
Will do.

Thanks, Pete.

Rolf

  • Guest
Re: Firefox and Traps
« Reply #58 on: April 13, 2018, 06:51:57 pm »
Hello Ben
Hello Pete

I uploaded the latest beta version (April 2018) of BS_Info.

www.bausys.ch\files\bs_info_2_1_beta.wpi


Rolf

Ben

  • Guest
Re: Firefox and Traps
« Reply #59 on: April 13, 2018, 07:43:07 pm »
Thanks Rolf.

I downloaded it and gave it a try; same problem hanging as before on cups with no error code created.

 However... when I looked at the title/version it reported November of last year, (2017), not April of 2018.

Was there an error in the link?