AstroSkipper

My investigation of all other drivers, which might be problematic in that partition, has been finally finished. After all these already performed measures and all this hypothesis stuff , here's now a little and in my opinion quite interesting proof (what else could you expect from me ): Prerequisites: A native Windows XP Professional SP3 32-bit without any POSReady updates, Thorium SSE2 122.0.6261.171 WINXP x32, Total Commander and Process Hacker. Assertion: Driver remnants from previously uninstalled programmes can cause spontaneous, irregular crashes of the Thorium browser under Windows XP Professional 32-bit. Proof: The current status at the beginning of this proof: Thorium is still crashing spontaneously, sooner or later, without any recognisable pattern. After a thorough investigation of all installed drivers in this Windows XP partition listed in the prerequisites, I found three drivers in the c:\WINDOWS\system32\drivers\ folder in the form of SYS files, which are loaded correctly after system startup, though. These files are avgntflt.sys, avipbb.sys and avkmgr.sys whose manufacturer is Avira. They are automatically loaded at system startup and logged in the Event Viewer as properly loaded drivers. That's why I didn't notice them earlier as I only looked there for errors. However, there are no programmes of this manufacturer installed any longer in this partition. This was indeed the crucial hint. After checking and unpacking all installation files that could have installed these drivers in the past, I was able to identify the culprit. It was the crappy Wise Anti Malware 2.2.1, which was installed for a short time for testing purposes only. With the help of Process Hacker and Total Commander, I was able to stop the associated processes and completely remove all three driver files, but only after restarting the system. Immediately afterwards, I opened the Thorium browser, configured it completely and restarted the browser several times. No first launch crashes anymore! Various pages were called up. The browser did not crash even once, no matter what actions were performed. Now, Thorium is working absolute stably in this partition as it is in my POSReady partition. So, these three drivers or at least, one of them (probably the avgntflt.sys driver as it is a filter driver and was more difficult to remove), were definitely the cause of the permanent Thorium crashes. q.e.d. PS: So, forget about my formerly created hypotheses H0! They all have to be rejected with an empirically determined probability of type I error of nearly 0. There is no indication that Thorium requires POSReady updates under Windows XP 32-bit, at least at the moment.

Regarding my issue with sudden Thorium crashes, I performed the following measures so far: Checking the file system for errors. Checking the integrity of the partition structure. Purging the partition in terms of trash/temp files. Complete defragmentation of the partition. Offline defragmentation of pagefile and registry files. Clearing and reorganising of the prefetch files. Checking the registry for errors and fixing them. Checking for missing updates performed by Legacy Update. Adding a couple of Windows updates which weren't offered by WU/MU. Installation of the Microsoft Visual C++ 2015 - 2019 Redistributable Package as it is also installed in the POSReady partition. Verifying the integrity of the DirectX 9.0c installation. Verifying the integrity of all .NET Framework installations. Uninstalling or disabling programmes which are permanently loaded and don't exist in my POSReady installation. Checking all autostart programmes and services. Fixing of all errors from the event log. Configuration of all services corresponding to my POSReady installation. Updating of programmes, which are permanently loaded, to the same version as in the POSReady installation. Update of the root certificates. Fixing a network error and adjusting all network settings. Graphics card driver completely uninstalled, all leftovers cleaned and more recent driver version installed. Now, my Windows XP Pro SP3 installation without POSReady updates is completely error-free except the issue with sudden Thorium crashes . The good is this Windows XP installation is now even fresher and faster than ever. My next measure is the check of all other drivers in that partition which might be problematic.

I checked my Thorium processes regarding loaded modules, and on my Windows XP Pro SP3 32-bit, the vulkan-1.dll is not loaded.

You should better use GPU-Z to see all features of your graphics card.

Today, I have installed Legacy Update 1.9 on my Windows XP Professional SP3 installation where POSReady updates have never been applied. It's just a pure Windows XP 32-bit. Due to the lack of POSReady updates, I decided to test Legacy Update in this XP installation. It was installed in seconds and calling up this service worked without any problems. It even offered to install root certificates and POSReady updates, which I didn't want and had to untick during the install process, though. The Windows Update search was performed in the Internet Explorer without any settings or actions by the user, and only few minutes later, it was done. Great service! I can definitely recommend it, especially to those who haven't use the POSReady hack. BTW, the result was the same I got in 2014 as all WU/MU updates have been installed at that time. I just wanted to be sure that no updates are missing.

The Vulkan feature is not an issue for me as my graphics card does not support this. It is simply disabled in Thorium.

I have installed all Microsoft Visual C++ Redistributable Packages up to the last XP-compatible version 2019. You can't really determine which packages are outdated or not as older programmes need very often older VC++ versions. And I don't think these package are responsible for the Thorium crashes as they all are also installed in my Windows XP Pro SP3 + POSReady partition. And there, all is fine.

TBH, so far I have been able to find and even fix every error in my Windows XP partitions. My personal statistics therefore clearly indicate that I will also eliminate this issue. The current, empirically obtained data from other users does not, of course, confirm my null hypothesis H0. As I said before, I am not willing to install POSReady updates in my Windows XP Pro SP3-only partition. So I need to take alternative options and investigate them in depth.

@NotHereToPlayGames You should not fall into the belief that @AstroSkipper is plagued by bias and narrow-mindedness. As a scientist and mathematician, it is a matter of course to open the mind in all directions. Hypotheses are made and either accepted or rejected. Assertions, on the other hand, have to be proven. Since I formulated the relation between Thorium crashes and the non-existence of POSReady updates in a Windows XP installation as a hypothesis from the outset, there is of course the possibility of accepting a hypothesis that is simply wrong, which is called an type II error in statistics. I made this clear from the very beginning. My investigations go in all directions. From the beginning, an alternative theory was that some loaded modules, programmes or drivers could be responsible for the Thorium crashes. I am investigating my problem in all possible directions as I always do.

The main reason for the values in that stats is the fact that Windows computers have been extremely displaced by smartphones and tablets in the last 20 years. And if a normal consumer uses a notebook and, now rather rarely, a desktop computer, it will probably be a Windows 10 or 11 computer. And there, the normal user is more likely to install what they know from their mobile devices. Chrome or Chromium. The younger generation now only seems to know Windows computers from the stories of their elders. But all that doesn't make Chrome and Chromium browsers any better.

But too much time-consuming, considering that we are talking about 365 updates, as I mentioned earlier, although I am also interested in the cause of these sudden closes. After a deep research in the internet, I found out that many users suffer from this Chrome issue and have switched to Firefox as there seems to be no real solution. If this error also occurs on my Windows XP Pro SP3 + POSReady partition, I will immediately give up Thorium.

Thorium is Chrome based. Search for the string " Chrome immediately closes when starting" or similar search strings, and you will see I am not the only one. My issue is a common error in Chrome. You should actually know that as a Chrome/Chromium fan.

Both partitions were created by me in 2004. Since then, there has been no reinstallation due to a perfectly functioning imaging system. I correct all errors myself. The registry of both have been modified, cleaned and fixed by myself. The main difference between them is that I had to stop the updating of Windows XP with the last update in 2014 in one of them, and in the other, I applied the POSReady hack and performed updating until 2019. PS: I never had an infection, which had to be removed by an antimalware programme, in both Windows XP installations. And don't forget there is an undetermined probability of type II error for accepting the null hypothesis H0.

I tend to assume that my computer is such certain hardware. The reason for my assumption is that everything works in my Windows XP Pro SP3 without POSReady partition. I have never had such problems. Firefox, New Moon and also Mypal 68 work perfectly there. This partition is configured in exactly the same way as the one with Windows XP Pro SP3 + POSReady updates. Actually, my issue can only be due to a missing patch or update. And the only updates missing there are the POSReady ones.

Yep! That's why I modified my null hypothesis H0 some days ago. I tend to assume that my computer is such certain hardware.

After all my measures, and I performed many of them, I have to confirm my last null hypothesis H0. Thorium behaves unstably on my Windows XP Pro SP3 without POSReady. Sometimes, the browser opens and lasts for a few seconds or minutes, very often the browser closes during startup. There is no recognisable pattern. I also tried different combinations of command line flags. No success. The internal site chrome://metrics-internals/ doesn't provide any information on my old machine. The good thing about the story is that my Windows XP Pro SP3 without POSReady installation is now absolutely error-free and updated.

I tried it on my Windows XP Pro SP3 + POSReady installation months ago. It didn't do anything interesting, so I removed it a short time later. I should mention that RAM consumption of Thorium is not the problem on my computer. And just for clarification, this flag actually can't be a solution for my issue, as Thorium works fine without it on my Windows XP Pro SP3 + POSReady installation. But I'll give it a try anyway. Just to be sure.

Yep! Same here. I have never installed Thorium using the install batch file or the REG files provided in the downloaded archive file. BTW, these should have to be adjusted to the respective system if you want to use them. Furthermore, the REG files contain syntax errors. For me, Thorium have been running error-free on my Windows XP Pro SP3 + POSReady for months by only extracting the archive and using a batch file or loader with command line flags for starting. But on my Windows XP Pro SP3 without POSReady, it permanently crashes and behaves unstably.

Thanks for your idea! That's why I originally assumed POSReady updates were mandatory for running Thorium. Maybe, there is really an POSReady update which is urgently needed to run Thorium on my hardware. But I should mention that there are 365 updates in my archive applied after the POSReady hack. And without this hack, you can't install those updates, either. So, it does not make sense to investigate them. And I already have a perfectly working Windows XP Professional installation with all POSReady updates. All my very deep investigations lead me to modify my null hypothesis: Here is my modified null hypothesis H0: "Thorium crashes due to the lack of one or more POSReady updates under certain hardware conditions". Personally, I tend to accept H0 due to my experiences and deep investigations. Unfortunately, the probability of type II error for this decision cannot be determined again due to the too small sample size, either.

To avoid messing up my main partition, I actually wanted to test Thorium in my second partition. But there, as already reported, the browser does not work stable. It crashes all the time. But in one of its bright moments, I was at least able to open a link from a PDF file in Thorium by directly clicking onto it. The REG files from Thorium were not the solution. They contain syntax errors. I made Thorium the default browser with the help of the Default Browser 1.8 tool that @dmiranda had already recommended here. For me, this tool worked as it should.

You seem to be using the older release Thorium_122.0.6261.168_WINXP_x32. But I am usimg the most recent release Thorium_SSE2_122.0.6261.171_WINXP_x32. Did you try this one on your computer where is installed Windows XP 32-bit without POSReady updates?

@NotHereToPlayGames Which version of Thorium do you run on your Windows XP 32-bit without POSReady updates?

Let H0 be the null hypothesis "Thorium crashes due to lack of POSReady updates". H0 must be rejected due to @NotHereToPlayGames's and @Saxon's report that Thorium runs fine in their Windows XP installations without POSReady updates. Unfortunately, the probability of error for this decision cannot be determined due to the too small sample size.

We already had this disscusion here: https://msfn.org/board/topic/184051-my-browser-builds-part-4/?do=findComment&comment=1251008 And there is nothing more to add to this statement.

Just for clarification. I am not a self-proclaimed "mathematician" but just a mathematician. I have not announced anything here, but only mentioned it in passing. No more, no less. As part of my search for the cause of the spontaneous Thorium crashes, I formed a lot of hypotheses and had to unfortunately reject them. And I have investigated many things as part of my hypotheses. But so far without success.