AstroSkipper Posted April 25 Share Posted April 25 7 minutes ago, NotHereToPlayGames said: I wouldn't rule it out, if you have a way to test without POSReady. We (plural) [work-related] had several several SEVERAL issues with POSReady. Our IT ended up "banning" POSReady on all factory computers running XP. Ok. As usual the proof is in the pudding. I have two partitions with Windows XP installed. One with POSReady 2009 and one without. I will test that and report here. 3 Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 25 Share Posted April 25 (edited) 5 hours ago, UCyborg said: Some bugs could look like that in practice. One program I work with at my workplace had an issue in the past where a specific function caused the crash, we noticed at some point during diagnosis the customer is using single-core CPU. It crashed on our test systems as well when we ran the program with affinity set to 1 core. Thanks for your assessment! I tend to agree with that. My Windows XP computer is indeed equipped with a 32-bit single-core CPU. I think using Mypal 68 in single-process mode is more suitable as I already did from the very first. In this mode, it uses also much less RAM. Edited April 25 by AstroSkipper 2 Link to comment Share on other sites More sharing options...
NotHereToPlayGames Posted April 25 Share Posted April 25 On 4/23/2024 at 12:08 PM, AstroSkipper said: @feodor2 I have installed your latest version Mypal 68.14.0b and tried it in multi-process mode. 1 hour ago, AstroSkipper said: My Windows XP computer is indeed equipped with a 32-bit single-core CPU. I'm confused. If your computer is single-core, then why were you testing in multi-process mode? Guess I "assumed" that multi-proces was ONLY for multi-core. But it is an "assumption". Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 25 Share Posted April 25 (edited) 21 minutes ago, NotHereToPlayGames said: I'm confused. If your computer is single-core, then why were you testing in multi-process mode? Guess I "assumed" that multi-proces was ONLY for multi-core. But it is an "assumption". Of course, this is all a little confusing. But the multi-process mode of Mypal 68 has nothing directly to do with a multi-core CPU. In multi-process mode, Mypal 68 opens every single tab, all extensions and other things into new browser processes, just as Chrome (including 360Chrome) has been doing for a long time. The advantage of Mypal 68 is that the single-process mode still exists and can be switched on if desired. Each of these modes has advantages, but also disadvantages, as you surely know. Edited April 25 by AstroSkipper Update of content 3 Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 25 Share Posted April 25 (edited) 6 hours ago, AstroSkipper said: 6 hours ago, NotHereToPlayGames said: I wouldn't rule it out, if you have a way to test without POSReady. We (plural) [work-related] had several several SEVERAL issues with POSReady. Our IT ended up "banning" POSReady on all factory computers running XP. Ok. As usual the proof is in the pudding. I have two partitions with Windows XP installed. One with POSReady 2009 and one without. I will test that and report here. 7 hours ago, genieautravail said: @AstroSkipper If I remember well, your computer is full of Windows Embedded POSReady 2009 patches. Perhaps that the issue comes from that. Your OS isn't anymore a real XP and I don't think that @feodor2 has installed all these patches on his computers. Comparing yours results will be very complicated. Regards I have just tested uBlock Origin 1.57.2 in Mypal 68.14.0b multi-process mode on my Windows XP partition without any POSReady 2009 updates. Same issue. When starting to load the filter list Adguard - Ads, all is over. Endless loading and uBlock Origin is then completely broken. Thus, as I already assumed, this issue is definitely not related to POSReady 2009 patches/updates. Edited April 25 by AstroSkipper Update of content 3 Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 25 Share Posted April 25 (edited) @feodor2 I have deeply tested uBlock Origin 1.57.2 and older versions of uBO in Mypal 68.14.0b, Mypal 68.13.7b and Mypal 68.13.9b. All in multi-process mode on my computer with a Pentium 4 32-bit single-core CPU, even in different Windows XP installations. I got the same issues, the same problems as already described here. Conclusion at the moment: uBlock Origin in Mypal 68 with default multi-process mode is not working properly on my computer with a single-core CPU. Updating of filter lists is broken which results in a complete breakage of uBlock Origin. As long as this cannot be confirmed or fixed by you, I'll revert all installations back to single-process mode which is more stable on my computer with a single-core CPU than your recommended and by default preconfigured multi-process mode. At least as for uBlock Origin. In single-process mode, uBlock Origin is working as it should. Sorry for these bad news! Nevertheless, I really appreciate your efforts in developing Mypal 68. PS: If anyone were or weren't able to confirm my observations on a computer equipped with a single-core CPU, I would really appreciate that. But there is also the possibility that I am the only one, for whatever reason. So, the question is not "To be or not to be" but am I really the only one? Edited April 25 by AstroSkipper Update of content 2 Link to comment Share on other sites More sharing options...
UCyborg Posted April 25 Share Posted April 25 No idea, but I didn't manage to reproduce on my machine, I tried forcing Mypal 68.14.0b on single-core, kept multi-process enabled, couldn't break uBO 1.57.2. 1 Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 25 Share Posted April 25 (edited) 1 hour ago, UCyborg said: No idea, but I didn't manage to reproduce on my machine, I tried forcing Mypal 68.14.0b on single-core, kept multi-process enabled, couldn't break uBO 1.57.2. Thanks for testing! Even if you try to force Mypal 68.14.0b being executed on only one core of your 64-bit multi-core CPU, it is probably not the same as a real 32-bit single-core CPU. Or what do you think? Edited April 25 by AstroSkipper 2 Link to comment Share on other sites More sharing options...
feodor2 Posted April 26 Share Posted April 26 (edited) @genieautravail If 5 users will gether, okey next version will have SSE. What doesn't suit your already existing version, because new version is more bloated and likely has new errors. @AstroSkipper What about other things or this is happens only with Adguard, soon I shall make testing on P4 single core, @AstroSkipper@NotHereToPlayGames Multiprocess is not about cpu cores. Windows OS multiprocess itself since 1993 I think, and Intel 386 the first CPU which supports multiprocess. Also Firefox is not like Chrome here, they try be it latter at 100's, I think making every piece on the separate process is overkill. In general mp is good for performance, and stability too. If you have ar least 2gb ram prefer stay mp. Furthermore POSready updates installed everywhere on my side, never heard about issues with this. Edited April 26 by feodor2 2 Link to comment Share on other sites More sharing options...
UCyborg Posted April 26 Share Posted April 26 10 hours ago, AstroSkipper said: Or what do you think? I think we don't really know the real troubling factor and just speculating so far. If you force affinity to one core, then OS will schedule that program on that one core, so theoretically, things wouldn't be able to truly run in parallel in that program. Regarding 32-bit and 64-bit, they're just 2 modes and the OS must setup things for 32-bit programs, so I guess 32-bit context is as real as on 32-bit CPUs. That one core on multi-core CPU will obviously be different than the core of other single-core CPUs, but I guess you could say the same for different single-core CPUs. Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 26 Share Posted April 26 (edited) 44 minutes ago, UCyborg said: I think we don't really know the real troubling factor and just speculating so far. If you force affinity to one core, then OS will schedule that program on that one core, so theoretically, things wouldn't be able to truly run in parallel in that program. Regarding 32-bit and 64-bit, they're just 2 modes and the OS must setup things for 32-bit programs, so I guess 32-bit context is as real as on 32-bit CPUs. That one core on multi-core CPU will obviously be different than the core of other single-core CPUs, but I guess you could say the same for different single-core CPUs. Of course, we are only speculating here. Unfortunately, I have not yet been able to discover any error message or anything that would be helpful in investigating the cause of this issue. I wouldn't even have noticed this error, as I was actually only running Mypal 68 in single-process mode. I tried the multi-process mode just to see if it works well on my old computer and noticed that Codeberg.org runs stable in this mode. As already described, Codeberg.org stupidly leads to crashes in single-process mode.And it was @feodor2 last year who told me in his issues on GitHub that Mypal 68 produces less errors in multi-process mode than in single-process mode and would be therefore the more stable mode. So, I wanted to give this mode a try. Edited April 26 by AstroSkipper Update of content 2 Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 26 Share Posted April 26 (edited) 2 hours ago, feodor2 said: @AstroSkipper What about other things or this is happens only with Adguard Thanks for replying! Unfortunately, I don't understand what you wanted to say or ask. I never used Adguard with multi-process mode in Mypal 68 but only uBlock Origin. 2 hours ago, feodor2 said: soon I shall make testing on P4 single core, That would be great. Edited April 26 by AstroSkipper 2 Link to comment Share on other sites More sharing options...
grey_rat Posted April 26 Share Posted April 26 AstroSkipper, try to switch extensions.webextensions.remote - false 2 Link to comment Share on other sites More sharing options...
AstroSkipper Posted April 26 Share Posted April 26 (edited) 1 hour ago, grey_rat said: AstroSkipper, try to switch extensions.webextensions.remote - false Thank you very much! I already tried different preferences. Also switching both preferences, extensions.webextensions.remote and extensions.webextensions.protocol.remote, together to the value false. But this didn't work. In contrast, as suggested by you, setting only the preference extensions.webextensions.remote to false seems to do the trick. Of course, I have to observe this for a while to be sure all is good now but at the moment, it is definitely working. So, many thanks again! BTW, does switching this preference to false have any disadvantages or negative side-effects? Edited April 26 by AstroSkipper Update of content 2 Link to comment Share on other sites More sharing options...
UCyborg Posted April 26 Share Posted April 26 Multi-process won't work as designed then. https://www.ghacks.net/2017/07/11/firefox-nightly-out-of-process-web-extensions/ And potential compatibility issues with extensions, eg. https://github.com/hackademix/noscript/issues/148 Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now