Jump to content
MSFN is made available via donations, subscriptions and advertising revenue. The use of ad-blocking software hurts the site. Please disable ad-blocking software or set an exception for MSFN. ×


  • Posts

  • Joined

  • Last visited

  • Donations

  • Country


Posts posted by TmEE

  1. I haven't been here in a while now... I use 98SE and ME but no longer on my main use machine, which runs XP primarly instead. It has been very difficult to do every day things involving internet on 9x machines. I have stopped looking at any developments on KernelEx and related front also...

  2. The problem is not coming from the service pack. I just made a new installation and added nothing but KernelEx 4.5.2 to it. Old versions of XMplay worked fine then, including the crash test I set up, no invalid page fault error or that "corrupt data" message.

    Then I upgraded vanilla KernelEx 4.5.2 to the release 17 and the old XMplay versions stop working and the crash test starts crashing so the problem is definitely introduced by updated KernelEx.

  3. I use bunch of Winamp plugins with XMplay due to lack of native plugins for those particular formats and there are some that require help from KernelEx.

    I also tried some older versions of XMplay and versions 3.4.1 to 3.5.1 crashed with an invalid page fault on startup unless extensions are disabled. With vanilla 4.5.2 they run fine.  Versions starting with 3.6 work fine up to the very latest one, I haven't tired any of the older versions to see how far back the particular problem goes.

  4. Allo !

    It seems that one XMPLAY plugin to allow RAR compressed files use is broken since at least Release 4 on my machine with Win ME + RP9.7.2 + the new service pack, same configuration works properly with vanilla KEX 4.5.2.

    I tested 4, 5, 7, 8, 9, 10, 11, 16 and 17 and with all of them the program dies with "corrupt data!" message after which it enters into endless invalid page fault error box throwing state, unkillable by any means so you got to do a hard reset to the machine in order for it to be shut down. Vanilla 4.5.2 has the program working fine without any sort of problems. I removed Kstub822 for testing purposes and it made no difference. All versions were tested by restoring Kex to 4.5.2 prior to updating the relevant files of each release. Disabling KEX extensions on the program would allow the plugin to work but since there are other plugins that do require KEX that isn't an option, also enabling ANY compatibility mode will cause the program to exhibit the problem.

    I ended up looking for any possible updated version of the plugin and there is one which does work without problems, but that still won't address that there seems to be some sort of problem in the updated KEX that might affect other programs. Here's an archive that contains the problem combination for debugging/analysis : http://www.tmeeco.eu/Fileden/XMPLAYcrash.RAR

    Nearly all other programs used on that machine seem to work without any surprises. (The one that doesn't won't work with Kex4.5.2 either so no discussion about it will happen here as per rules of this topic).

    Thänk you for all this great work ~

  5. Older versions are artificially blocked by version check. You can patch the version number reported and get things to work.



    When autologin is enabled Skype stores a certificate in the profile folder which is used for logon, and it lasts 30 days.

    To get one month of login ability you would want to install latest version that can log in (6.15 I think, I never determined the point where restriction is set), log in with autologon enabled and then install older version on top.

    The older version can logon now until cert runs out or you log out yourself.

    I have been using this on XP though, so I can run 6.0 which actually works, the latest ones have memory leaks and very occasional freezing making their use near impossible...


    Lot of cool info ~~~~


  6. Word of warning :


    The util replaces IPHLPAPI.DLL for something newer/different and it breaks a ton of programs that KernelEx can make work such as Foxit Roader.

    I don't have compatible hardware but I was interested in trying out the util. After uninstall it does not restore the original file. I had to extract it from windows install CD.

  7. I found TWO half fixes :

    1. Change link speed from Auto to 10half/10full/100half/100full in card driver settings page.

    2. Enable file sharing from network control panel.

    By half I mean I no longer have to browse to the shares or access unprotected share first, but I do have to go to the already found share from the network places icon. Shortcuts on desktop or elsewhere do not seem to work on first try... but sometimes they do...

    This stuff is not working. Just maybe once out of 10 reboots...

    I managed to gather a bunch of versions of the driver, one even newer than the one on broadcom site ! I will be going through them one by one and hope one of them actually works.

    EDIT: Oldest driver I found works consistently across multiple reboots. I shall try the newer ones and report back the point where things seem to break !

    EDIT2: The oldest I found is the only one that actually works. Version works, next one I found which is 7.35.00 does not, nor do, (one on Broadcom site) and

  8. 1. Modern hardware, but that is what PCI slots are for.

    2. As long as IPv4 exists, that one will not be a problem.

    3. No. Latest Flash runs fine.

    4. SSL has never been an issue. What is in my country is that most stuff is done with the ID card and the software for it is XP and higher. I do have an XP machine around so that's not a problem really to me.

    5. Haven't had any problems around that.

    6. Rather opposite. The newer the browsers get the worse they seem to become as far as speed goes. Put the OS bloat on top and there is no way modern OS+browser wins.

    My 9x machines all run circles around the XP+ machines, even with same hardware. One of the more tangible reasons I keep using 9x, other really big one is the hardware I use which just isn't supported by newer stuff, but is indisposable in my "production" environment.

  9. That did not work. Or... TPC-PIII1400-03

    ...Or... TMEE.TPC-PIII1400-03

    ...is not the proper syntax. Computer named TPC-PIII1400-03 in workgroup TMEE that sits at static IP of

    Using any other NIC I have (PCMCIA and internal WiFi) will work correctly on first try using any shortcut or going through network places. Same install on different hardware also works correctly.

    ME itself seems to have no trouble locating the shares when I connect to the LAN, but accessing will be the problem...

    I have also ruled out the problem being with installation. I have been to some 10 installs on the hardware trying different things and the LAN issue presisted with the integrated LAN which is Broadcom Netlink 5788.

  10. There's not a whole lot else to think of. Hardware is most certainly fine (no issues with 98SE, XP, Vista and Win8).

    A PCMCIA NIC I got works just fine too, but it is 100M and onboard is 1000M, and I have got very used to the higher speeds...

    Is the driver a 98 driver or a Me driver, for trying to use 98 drivers on a Me system could be a pain in the butt.

    The driver is 98/SE/ME.

    Most just 95 or 98 drivers work fine, though I experienced a problem with one RAID card driver, that made shutdowns hang. I will have to see if there's other driver versions for that.

  11. Seems the problem is related to the laptop itself in some way, when it has ME on it.

    I set up ME on my other computer and all was fine and dandy, no weird issues.

    98 on the laptop, with latest driver from broadcom works fine, no weirdnesses... except hang on shutdown when IP is not released (but installing microsoft IPSec thing fixes that).

    ME on that laptop, with same driver as 98, the weirdnesses described. No hang on shutdown in any case.

    Only thing left to do is to try older drivers and hope one of them behaves better.

  12. And I poked him again, and this is what he told :

    18:28:30: Its not the section structure
    18:29:23: something changes the state of the pages to COPY WRITE
    18:30:34: the first two pages of .data section get changed from READ_WRITE to WRITE_COPY
    18:30:53: I do not know what api call does this, something todo with shared mapping
    18:31:07: however win9x does not support this page mode
    18:32:01: so there is no boundry between the remaining READ_WRITE pages of .data and the pages of .text which where temporarly made READ_WRITE
    18:35:09: I will try hooking MapViewOfFile and see if that is culpret

    18:44:06: actually the problem is nothing to do with the section structure
    18:44:29: it seems that the WRITE_COPY section is set by windows

    18:46:18: its a memory optimisation, setting the .data section to copy on write in case the program decides not to change it
    18:47:23: it seems to only apply to complete pages, that would explain why the last partial page is normal READ_WRITE
    18:47:50: if this is the case then the gcc startup code actually has a bug that would effect modern windows
    18:48:10: if the program has less than 4kb of data then it would crash I would think

    I'll edit the post when more info is available.

  • Create New...