Jump to content

schwups

Member
  • Posts

    1,203
  • Joined

  • Last visited

  • Days Won

    3
  • Donations

    0.00 USD 
  • Country

    Germany

Everything posted by schwups

  1. Odd path - not standard - Local Settings is a XP folder I think it should be on ME/98: C:\WINDOWS\Application Data\Sun\Java
  2. I know from my own experience that Java installer errors can drive you to despair. I had a variety of Java errors, but I can't remember this one. - 1.5.0_06-b05.xml? Did you install version 1.5.0_06 before? Maybe a leftover or something is pending of that installation. - WindowsInstaller 2 (2.0.2600.2) required. - Yes, Java 6 requires KernelEx. https://kernelex.sourceforge.net/wiki/Sun_Java_Runtime_Environment This helps in some cases, but probably it doesn't include a solution for your case: https://web.archive.org/web/20111229065524/http://www.javafixer.org https://www.oracle.com/java/technologies/javase/jdk-v5-microsoft-windows.html
  3. Yes, Oracle removed the JavaRuntime offline installer jre-1_5_0_22-windows-i586-p.exe, but why you think the file on oldversion.com is an online installer? It is offline. You can also look on archive.org, for example https://web.archive.org/web/20210127174644/https://www.oldapps.com/java.php Did you try to install it without internet connection. The setup could finish after a short while. You can try 1.6.0.7, too. It is very common that old Java leftovers cause problems.
  4. @ Goodmaneuver Yes, 26pre1 offers indeed advantages in testing and troubleshooting. The apphelp.dll of ReactOS doesn't have my problem dependency BaseUpdateAppcompatCache. I added two definitions to Kstub.ini only. [Ntdll.dll] RTLFindCharInUnicodeString=n4 RTLPcToFileHeader=n2 So I think can also make pre1 usable for the ordinary use by including renaming the corresponding affected files.
  5. Very nice, great work, but IMO compiling KernelEx doesn't really fit here. This is not the topic here and a discussion would be off topic. It should be moved to https://msfn.org/board/topic/162223-kernelex-on-building-debugging-and-related-matters/.
  6. No, I ruled that out. I tried different versions, but the message "AppHelp.dll is linked to missing export Kernel32.dll: BaseUpdateAppcompatCache" still appeared on OS start. (using KernelEx.dllpre1). Version 26pre2 is OK. XP file (MSGina) was another main culprit. I added this file a while ago, but I don't know anymore which program it was for. Until then the file was unremarkable and forgotten. After renaming Msgina some messages were gone (using KernelEx.dllpre1). Ok, you are right. I used version 19.
  7. No, I'm not happy with new KernelEx.dll. It isn't usable in this pattern for me. Basically, I think it's a very good feature for testing, but not for normal, ordinary usage. It gives always a prompt for me, if an API is missing. Previously, non-essential APIs were ignored by the program. They are now becoming a problem by normal use. First a program freezes, then the OS freezes. In the majority of cases, however, then Ctrl Alt Del works. It already begins on OS start with message: "Kex22.KernelEx 2022 v4.22.26.1 AppHelp.dll is linked to missing export Kernel32.dll: BaseUpdateAppcompatCache. Notice: Select [OK], runApiLog.exe, then rerun your program." I've BaseUpdateAppcompatCache still disabled in the kstub.ini, because I wasn't able to find a useful def so far. If enabled ME already hangs on start (using KernelEx.dll25). With the new file on prompt I can click OK and then it goes on, whats positive on the other hand. Now e.g. "missing export msvcrt.dll: memmove_s" leads to freezings. So far I enabled it for testing only. In my opinion this new feature should be switchable. An ongoing KernelEx.dll version change for testing is too cumbersome. Sheet.dll: Not really tested, but level of desired API support didn't stick on first tries. It remains on the default setting. Maybe I must check something before? Kstub.ini: Ordinals are generally undocumented. My question regarding the example: Is f1e usually a good def value? Even if it sounds stupid after my comments in this post, thanks for all your work!!
  8. Tools>Options>General>uncheck Thunderbird Start Page - Maybe not perfect, but should avoid these errors.
  9. Thunderbird 2.0.0.24: WinSSLwrap supports TLS 1.2! https://msfn.org/board/topic/174247-windows-9x-ssl-protocol-winsslwrap-117-rev-08/ It works with OutlookExpress6 here and it should work with Thunderbird. https://msfn.org/board/topic/183829-email-on-windows-98-2022/?do=findComment&comment=1224261 Firefox 9/10 and Thunderbird 9/10 are compiled with VC2005 (msvc80). So it is very likely that the replacement by RoyTam's nss files usually works. I use the RT's nss files for v10 with Firefox 9 and I can't remember that there are any problems. Maybe there is a numbering, but the program doesn't check it. However, I have not or hardly tested this for TB. Yes, I had the the same issue. I haven't done any further research on this lately. The issue doesn't occur with Thunderbird 42 and higher. Usually I use TB 52. Probably you must enable TLS 1.2 on v24. Firefox: "security.tls.version.max" to 3 (TLS 1.2 enabled). TLS 1.2 is enabled since version 27 by default. TB 27 with Start Page: tls13.akamai.io (TLS checker) - It is necessary to update the cert file nssckbi.dll (NSS Builtin Trusted Root CAs) by the one of TB 52.9 or higher, otherwise the page doesn't open. In the meantime, I can also start and run TB 36- 41 and 53. These versions also doesn't run in the background after close. But I have some other issues with these versions. I'm not sure. I haven't bothered with this version lately either. Try Kexmode Basesn on msvcr100.
  10. Are you looking for this: https://web.archive.org/web/20150604222527/http://www.mdgx.com/files/?loc=/&arange=nd User: ME280800 Gdi: ME918547 98SE: Q918547
  11. Oh yes, today I remembered loblo's method for Opera. Why shouldn't it work for Chrome to get TLS 1.2/1.3 support: Using Burp Suite Community Edition as an HTTPS proxy for Opera running on Azul OpenJDK 8. https://msfn.org/board/topic/178283-browsing-the-web-on-98me-in-2019-and-beyond/?do=findComment&comment=1184454 and following posts
  12. BWC's Chrometool isn't compatible with Win ME / KernelEx. I think KernelEx would have to be considerably rebuilt to get compatibility. But I'm happy to be taught better. Download links for SP's: IMO service packs will not really help to run Chrome.
  13. There is no workaround yet, but who knows what's in a year. I have hardly dealt with Chrome in the last weeks or months. And what about the cryptographic support TLS 1.2/1.3? Google chrome doesn't support it itself. There is a solution for XP (POSReady 2009 updates...), but I don't think there's a simple solution to transfer that to ME. TLS 1.2 is standard today and browsing is almost no longer possible without it. I believe we're not far away to run IE7 or 8. Of course, those IE browsers are very old and obsolete and don't support TLS 1.2, too. I see valid hope for Mypal 68 in the near future. Currently the best and most modern browser I have is Palemoon (31) on the mypal 29.3 "engine" (xul, omni.ja,...). The advantage I've is that this configuration is significantly more stable than Mypal. Errors in Xul usually do not occur.
  14. Add-on "Get Archive" is a really fine thing to browse through the web archive (archive.org) for me as ME user. Version 2.5.2 - works with Firefox 3.6 - 56.* I can also use version 3.2 on Firefox 53 (compatible with Firefox 53 and later).
  15. Issue Light 48/49: Classic Add-ons Archive (2.0.3) doesn't work. Classic Add-ons Archive site becomes blank on search extensions.
  16. Neither KernelEx 4.5.2 nor core update 25 support RtlSetBits yet. Read and learn how to use Kext (DoItYourself KernelEx Extensions) here. Kstub.ini: [ntdll.dll] RtlSetBits= Such a simple entry in the Kstub.ini file may or may not work. Use the ImportPatcher to find missing dependencies.
  17. It can also be newer file versions. There is nothing, even no numbering, that prevents even a current version (NSS Builtin Trusted Root CAs) - like 91.13 from being used,
  18. Light author cstkingkey: "By default, Light only trust a few of Root CAs. It may cause problems. If you run into cert problems when visiting https sites and it works with firefox, you can dowload or find nssckbi.dll shipped in firefox and put it into light's folder." https://sourceforge.net/p/lightfirefox/wiki/certs/ Common message: Not trusted certificate because the issuer certificate has expired... etcetera I tested it with nssckbi.dll from Firefox 52.9.
  19. Light Browser. Not up to date, but slimmed down. As the name says, it is significantly smaller than the corresponding Firefox version. It can start faster and use less memory. It is no longer being developed since 2016. https://sourceforge.net/projects/lightfirefox/ https://sourceforge.net/projects/lightfirefox/files/ Portable Mode: https://sourceforge.net/p/lightfirefox/wiki/portable mode/ Version 48: It starts with UI and the graphical interface is fairly stable. That means you can run it without Hardware Acceleration (Compositing > Basic). I guess the Light Browser is better suited for underperforming or older Hardware. You should still set AzureContentBackend to skia. Supplement: Light - Firefox fork (versions 29 - 49). I continued my tests. Only versions 47 - 49 are suitable to run without Hardware Acceleration - lower versions aren't. Light 47 - 49: Set "gfx.content.azure.backends" to skia. Lower versions don't support skia. The support has been removed here. The addon ClassicThemeRestorer 1.5.3 can be used to solve the tab glitch. Version 49 requires Kext, but not the reg file "APISets Stub dll's/Knowndll's" (v49 compiled with VS2010). I think it's also interesting, that some versions between 36 and 41 run (with the known issues), namely 36r2, 40 and 41. (37, 38, 39 do not)
×
×
  • Create New...