MiKl

New KernelEx.dll installed on two machines and had no problems so far !! ~ Regarding the flash.306.msi installation problem. I have compared the three PCs and the major changes between the one where installation still worked and the two were installation failed are: - KernelEx updates - disabling Kex on system dlls I will test installing .306 after re-enabling system dlls. Would the following be good/meaningful/helpful ? REGEDIT4[HKEY_LOCAL_MACHINE\Software\KernelEx\AppSettings\Flags]"C:\\WINDOWS\\KERNELEX\\*"=dword:00000001

The newly released stable VLC 2.2.2 runs beautifully on KernelEx-v11&v12 ! Did not tested other configs. But I have some issues with the latest flashplayer 20.0.0.306 ! I have 3 main PCs that are more or less in use. But I could install .306 only on the PC that is the at least maintained one. It still uses Jumper's KEX v5 and only basic Kext. So I thought that the simple solution would be to just install this config on the other two PC and install Flash. But, of course, that would have been tooooo simple.

kexbasen v12 is 'causing' this problem only on this particular system - don't know why. All other 98SE machines are fine. So I went back to v10.

Hi Jumper, unfortunately I have to admit that I need more infos on debug console. I am still a noob on most things. And I have similar problems with VLC on a P4/ICH4-system running v11&v12-configuration that Dave-H reported earlier. Nothing happens when I try to start VLC. After enableing KEX on kexbasen.dll the follwing error message appears. I maybe just have to change one comp-mode but I can't find the culprit. Thanks for your hard work !! Datum 02/05/2016 Uhrzeit 20:53VLC verursachte einen Stapelfehler in Modul KEXBASEN.DLL bei 017f:7d00a034.Register:EAX=00440000 CS=017f EIP=7d00a034 EFLGS=00010293EBX=928c5e68 SS=0187 ESP=00572000 EBP=0057201cECX=9289efd4 DS=0187 ESI=7d000000 FS=4b1fEDX=7d017004 ES=0187 EDI=00000000 GS=0000Bytes bei CS:EIP:ff 15 28 40 01 7d c9 c2 0c 00 55 8b ec 51 81 7dStapelwerte:00440000 00000000 00000010 00000000 00000000 00000000 00000000 00572040 7d00a03a 00440000 00000000 00000010 00000000 00000000 00000000 00000000

Awesome - thanks !!! (Quick first test: SeaMonkey 2.7.x. and 2.8 are unfortunately still freezing.)

I am not a programmer and have only limited knowledge but I think Certcli.dll 5.1.2600.5512 seem to be fine. All first level dependecies of kexbasen.dll are 'green' in DW. Certcli.dll itself reports a few things missing like: [netapi32.dll] DsGetDcNameW DsRoleFreeMemory DsRoleGetPrimaryDomainInformation I don't know how important these are so please let me know if I should find all and post them here.

I am using ME's Ole32.dll (4.71.3328). There is indeed the hour-glass symbol on Certcli.dll ! I have made a screenshot but I am not sure how to upload it or can I just email it to you ? I am not sure if I understood you correctly, but here we go: - above certcli.dll is wintrust.dll (5.131.1880.14, no hour glass, kex disabled). - above wintrust.dll is cryptui.dll (5.131.1878.14, hour glass, kex disabled) - above is kexbasen.dll

I have the same problem with VLC on the system where v11 works !! On the other one using kexbasen v10 VLC still runs fine. On both systems the recommended changes from post 232 had been done .... strange. After re-enabling ole32.dll for KernelEx there actually is a error message: Update: I was finally able to get v11 working on the above mentioned AMD/VIA machine but unfortunately all of the kernelex-apps refused to work except Sunbird because this one seem to load only kexcom.dll. But - I played around a little bit and when kex is deactivated on msvcrt.dll (renamed msvcr70.dll 7.00.9981.0) apps like Notepad++ 'crash' on start but load and run fine anyway ! The same happens on SumatraPDF, OpenOffice, etc. Only VLC stubbornly still refuses to run. Hope this observation helps a little ... @DougB: Please check the comp-mode of msvcrt.dll when back at v10. @Jumper: When checking kexbasen.dll with DependecyWalker it reports a missing dll named certcli.dll. Should this one be added ?

@ Jumper, about the list of files from post 232 - should I also disable KernelEx on LZ32.dll ?? Kext re-directs to this dll, isn't it ? Update: Posting now from my AMD Athlon/Via machine and I think I made some progress. With full v11 I still get the runonce error. But after clicking OK I now reach desktop and some of the kernelex-apps like Sunbird work, while others like SeaMonkey do not. Update 2: On my 'best' system (P4/ICH6R) v11 seem to work perfectly. SeaMonkey 2.7.x. and 2.8. are running much longer now (several minutes) before the freezings happen. SeaMonkey 2.9. still freezes almost immediately.

Seem to be not working for me. Tried reg-merge and/or by manually adding this entry.

O.K. can hardly wait. By the way, I think what Drugwash and I meant is that after you have reached a point where updating the old gecko does not make any sense any longer to then work on later versions. Of course, you can't compile with VC6 then but most of us are on KEX anyway !! Fixing the bookmark issue and maybe also that SM freezes in versions 2.7.x up to 2.9 would be just awesome.

Interesting ! Firefox 2.0.0.20 already had a sessionstore system so it should be possible to port it over, shouldn't it ?

Yes that would be nice especially now as NoScript also dropped support for older geckos but as I understand it it is rn's first goal to create an as up-to-date as possible browser for vanilla 9x's !! Once that is achieved (within reason) he can still take the next step with what KernelEx allows at that point. Right now it seems to be SeaMonkey 2.6.1.

Very cool project !! I especially like the idea that the profile is now located within the actual app folder. Otherwise when testing diff. versions of SeaMonkey you'd have to very careful that your regular profile is not changed that much that is 'unusable' afterwards. Installing the German language pack worked also great. But it is kinda weird to use a browser again that does not save opened tabs/sessions for the next use. Do you plan to port this feature ? And an add-on manager is also needed. 2016 starts great for us 'oldies'

Hmmm, the fix is not working here - tested on 2 diff. systems.

Hi Jumper, since almost all of 'us' are running more or less 'pimped' Win98SE installs or are using ME I think it would be a good idea to test your builds also in these environments - at least on a MDGX-updated SE. Thank you for your hard work !!!

After playing around with some of the files and drivers I found after the great help here the situation has somewhat changed although I am not sure if it was a progress. The ATK has now moved to the system section in the device manager. Unfortunately it is still listed as unknown and it is deactivated with this red cross in the icon. Manually enabling it does not work. After a reboot it is deactivated again.

Same here unfortunately !! **********************************************************************Datum 01/01/2016 Uhrzeit 19:21RUNONCE verursachte einen Fehler durch eine ungültige Seitein Modul KERNEL32.DLL bei 017f:bff741ca.Register:EAX=c22c1980 CS=017f EIP=bff741ca EFLGS=00010202EBX=928804fc SS=0187 ESP=0063fbd0 EBP=0063fbfcECX=7d00b90d DS=0187 ESI=7fe40000 FS=4407EDX=00000000 ES=0187 EDI=00000000 GS=0000Bytes bei CS:EIP:ff 4a 10 75 08 89 42 08 ff 42 04 90 c3 39 42 08Stapelwerte:7d00b90d 00000000 7fe40000 928804fc bff20000 bff20000 bff741f7 928804b9 00000000 bffcd4c0 928817d4 0063fdcc 7fe41606 7fe40000 bff7ddd6 7fe40000**********************************************************************Datum 01/01/2016 Uhrzeit 19:21EXPLORER verursachte einen Fehler durch eine ungültige Seitein Modul KERNEL32.DLL bei 017f:bff741ca.Register:EAX=c22b9840 CS=017f EIP=bff741ca EFLGS=00010202EBX=928806d8 SS=0187 ESP=0059fbd0 EBP=0059fbfcECX=7d00b90d DS=0187 ESI=7fe40000 FS=50bfEDX=00000000 ES=0187 EDI=00000000 GS=0000Bytes bei CS:EIP:ff 4a 10 75 08 89 42 08 ff 42 04 90 c3 39 42 08Stapelwerte:7d00b90d 00000000 7fe40000 928806d8 bff20000 bff20000 bff741f7 92880695 00000000 bffcd4c0 9287fab8 0059fdcc 7fe41606 7fe40000 bff7ddd6 7fe40000**********************************************************************Datum 01/01/2016 Uhrzeit 19:22RUNDLL32 verursachte einen Fehler durch eine ungültige Seitein Modul KERNEL32.DLL bei 017f:bff741ca.Register:EAX=c22c6000 CS=017f EIP=bff741ca EFLGS=00010202EBX=9288165c SS=0187 ESP=0063fbd0 EBP=0063fbfcECX=7d00b90d DS=0187 ESI=7fe40000 FS=0fcfEDX=00000000 ES=0187 EDI=00000000 GS=0000Bytes bei CS:EIP:ff 4a 10 75 08 89 42 08 ff 42 04 90 c3 39 42 08Stapelwerte:7d00b90d 00000000 7fe40000 9288165c bff20000 bff20000 bff741f7 92881619 00000000 bffcd4c0 92881494 0063fdcc 7fe41606 7fe40000 bff7ddd6 7fe40000

Thank you !! There is another weird thing. The "IO read data port for ISA Plug and Play enumerator" states that it uses no known configuration.

Finally had the time to assemble a new system. MOBO is a ASUS P5GD1 Pro with NB: i915P and SB: 82801FR ICH6R. Went fine so far and I am posting from this PC right now but I have some issues in device manager: - 1 Unknown device 2 resource conflicts - Motherboard resources - Intel 82802 Firmware hub device The latter is the most confusing for me because of the southbridge mismatch or is this normal ?? Unfortunately it is not possible to change the resources for these devices. Anyone have maybe an idea what to do ? I already tried the PNP OS switch in the BIOS. All the best for 2016 !!!!!!!

With full version v9 and SeaMonkey 2.4.1. (un-hexed xul.dll) : I could succesfully 'print to PDF' a multi-page Email !! Of websites only page 1 is printed - all other pages are blank. Update: Have tested now SeaMonkey 2.6.1. with v9 and original xul.dll. - Printing emails is still fine. - Printing of certain websites works great while others fail with only priting page 1 or only fragments. Maybe a Java and/or Javasript issue ? @ Jumper, should I continue to post my print results here or in the dedicated thread ?

While testing SeaMonkey 2.3.3. I was hit by a BSOD. Datum 12/04/2015 Uhrzeit 20:58SEAMONKEY verursachte einen Fehler durch eine ungültige Seitein Modul KRNL386.EXE bei 0001:00002b23.Register:EAX=8f231f60 CS=014f EIP=00002b23 EFLGS=00000606EBX=00004258 SS=598f ESP=00005ab6 EBP=00005b1cECX=00002a90 DS=0177 ESI=8f23c11c FS=0167EDX=0000425f ES=0110 EDI=0000aa3c GS=0097Bytes bei CS:EIP:f3 67 66 a5 1f 52 66 50 e8 d0 23 65 80 4f 05 80Stapelwerte:000000ff ae200000 5b1c0000 5ad80000 0000e99f aac00000 aa400000 a6800000 00000000 01402cd3 22510000 00000840 00000840 0000aaa0 0000a680 00000523And maybe related: Datum 12/03/2015 Uhrzeit 23:09EXPLORER verursachte einen Fehler durch eine ungültige Seitein Modul KERNEL32.DLL bei 016f:bff88396.Register:EAX=c003098c CS=016f EIP=bff88396 EFLGS=00010206EBX=01ecffec SS=0177 ESP=01e8ff90 EBP=01e90108ECX=00000000 DS=0177 ESI=00000000 FS=3d87EDX=bff76855 ES=0177 EDI=bff79060 GS=0000Bytes bei CS:EIP:53 56 57 8b 75 10 8b 38 33 db 85 f6 75 2d 8d b5Stapelwerte:No 'Stapelwerte' were saved. I am reverting back to kernelex.dll v9 now. Hope to hear from Jumper soon. Thanks !! Added info: I am sure that Jumper is aware but with the v9/v10 configuration OpenOffice 3.2.1. does not start because of FindActCtxSectionStringA & FindActCtxSectionStringW now implemented in kexbases.

Hi Jumper, when you have the time could you look into browser printing running v10 ? With SeaMonkey 2.0.14 I can still print without any problems but SeaMonkey 2.1 and higher fail to print anything useful. If I patch xul.dll with [dll replacements] comdlg32.dll=comdlgEx.dll SM prints only blank pages. Both physical and to PDF. With original xul.dll SM prints: - from emails only header and subject but not the actual content - from websites only blank pages or the website's background if I activate that in the printing dialogue. Edit: Depending on hardware environment and SM build the above is also sometimes the other way around.

The Flash-crash problem seem to be gone with this version. Awesome !!

Weird, today it worked ! For 2.9.1. I needed to add to kstub.ini [ntdll.dll]: RtlAssert RtlConvertSidToUnicodeString RtlCreateEnvironment RtlDestroyEnvironment and SeaMonkey 2.9.1 loaded but always froze after a few seconds. Added info: SM 2.7.2 and 2.8. are a little bit more stable than 2.9.1. but also freeze sooner or later.