All Activity

This is not my question there but thanks in any case.

Why would I upvote something that doesn't work with yahoo and paypal? Never heard before! This topic caster helped ne a lot. I want to upvote him, sorry you're not happy with it. I'll give you a like for agreeing with an obvious statement though. If you're on github, if you like this browser, help us to reconvince the creator.

Very sharp observation. 👍️ I have no idea how that might have happened... The build environment is still the same Python 3.11 from @cmalex. What is that file supposed to do. Can it do any harm?

I cannot reply in your Russian forum but you are mistaken regarding IFSMGR.VXD 2227, it was Q301540.EXE and not 302540. How can you ignore it!

Another thing I wanted to mention to you @nicolaasjan is that in both the latest "WinXP" packages: https://github.com/nicolaasjan/yt-dlp/releases/download/2025.11.25.144622/yt-dlp_x86_winXP.exe https://github.com/nicolaasjan/yt-dlp/releases/download/2025.11.25.144622/yt-dlp_x86_winXP.zip a new DLL appears to have infiltrated, "fileextd.dll", which wasn't there in all of your previous "WinXP" releases, e.g. in https://github.com/nicolaasjan/yt-dlp/releases/download/2025.11.12.051143/yt-dlp_x86_winXP.exe https://github.com/nicolaasjan/yt-dlp/releases/download/2025.11.12.051143/yt-dlp_x86_winXP.zip Was that an inadvertent inclusion?

Ditto that...

The problem is, the same happens with multiple types of SSD, on multiple HP z210 CMT workstations. The users reported significantly longer boot times when compared to other computers. To be more precise, users reported computer malfunctions. When compared to standard HDD, computers with SSD took much more time to go through the blank (black) screen phase. Since other computers were faster in the same boot phase, the assumptiom was made the workstation with SSD were faulty, so workers started to power off the workastation with SSDs to fix the problem. After investigation, it came to light, the workstations with SSD took about a minute of black screen time, when workstations with HDD did the same in less than 10 seconds. After informing the workers about the extended wait time, the power cycling and computer failure reports stopped. All workstations have the same specifications, except hard drives, and monitors. The system partitions were cloned, so they do have the same drivers. After cloning system it is visible the HDD boot is longer (3m 22s). But with each sequential boot the time goes shorter. On the other hand SSD is faster first (1m 42s) but the time does not improve significantly, after that. All workstations are not using Microsoft Account. In any case, both SSD and HDD workstations would take the same time with the MSA and OneDrive, probably. There is no domain logging, as well. It all came to light with SSD and HDD mixed enviroment. In a different location all workstations are fitted with SSD, so there is no time difference between them. So workers assumed long boot time is normal behavior, probably. I will have to check how they are performing on boot, when I have an opportunity to be there.

I still build on XP the old fashioned way: python -m pip install -U "setuptools>=71.0.2,<81" pip wheel pip install -U -r requirements.txt python -m pip install pyinstaller==5.13.2 python devscripts/update-version.py -c "nicolaasjan/yt-dlp" -r "nicolaasjan/yt-dlp" "2025.11.25.144622" python -m bundle.pyinstaller --onedir Where `requirements.txt` is: mutagen pycryptodomex brotli; implementation_name=='cpython' certifi requests>=2.32.2,<3 urllib3>=2.0.2,<3 websockets>=13.0 yt-dlp-ejs I have tested building with PyInstaller 6.16.0 from 3dyd (waiting for 6.17.0 now...) It can be done and it works, but it also requires a reinstall, because of the totally different contents of the zip file (link)... (and then also the ugly presence of `psapi.dll` in the root directory to make it work) TL;DR; as long as it works the old fashioned way, I'm hesitant to build with the "modern" PyInstaller on XP. 🤔 It's annoying that yt-dlp always wants the latest and "greatest". 😠

When using dark magic sharp corners and a custom msstyles after restarting the computer the highlight over the start button is rounded.

that all are working on xp ? it would make sence to use a dll file for ech that pieces

Makes sense; if Z was the last to have been loaded in the tab when that tab was closed, why do you expect otherwise? After all, the browser feature is called "Recently Closed Tabs", not "Recently Visited Sites" ; however, a closed-tab's history isn't being deleted (it's saved in session data); re-open the tab with site Z on it, then use the back button once to revisit site Y (or twice for site X, etc.) ...

I use Supermium now exclusively, but had this issue in "new" Opera before. I see now one way it differs from history. If I had visited X,Y,Z in the same tab, I can only see Z on the list. There is no way to recall X or Y without opening history.

@nicolaasjan: Sunday is Nov 30th, 2025, thus I believe you'll be having a "hot potato" in your hands : https://github.com/pyinstaller/pyinstaller/issues/9149#issuecomment-3036991021 [build] Bump PyInstaller minimum version requirement to 6.17.0 https://github.com/yt-dlp/yt-dlp/commit/280165026886a1f1614ab527c34c66d71faa5d69 For the Vista+(aka Win7) builds, your PyInstaller fork v6.17 came just in time : https://github.com/nicolaasjan/pyinstaller-builds/releases/tag/6.17.0 ... but what about those WinXP "onedir" packages which, IIANM, are still using PyInstaller-5.13.2 due to this ? (you can always revert 2801650 prior to the onedir compilation for XP, but how practical would that be in the long run? ...)

roytam1 why are you ignoring my question? Link

RegCool x32 v3.0.0.1 Portable XP https://kurtzimmermann.com/downrc_en.html A brief attempt to bring RegCool v3.0.0.1 to Windows XP. Warning, The Updater.exe has not been touched, and since RegCool.exe was little tested, make a backup copy of the Registry before making modifications. Windows XP Fix Download: https://www.upload.ee/files/18849687/RegCool_x32_v3.0.0.1_Portable_XP.zip.html http://www.kurtzimmermann.com/updates/regcoolversion.ini https://kurtzimmermann.com/files/RegCoolXP.zip

Firefox 115.30ESR officially, or there is the firefox-for-windows-7 fork on github, that allows for 147 nightly on 7 and 8.x. As of now, the latest stable Steam version has been ported to at least work on 8 and 8.1 without extended kernel, but a few more DLLs need to be patched for 7

I had already commented that you cannot get past the captcha: "We could not the security challenge" , despite passing validation (green tick).

Nope. I run v140 (OT to Supermium but you did specifically call out "all Chrome variants", so it seems this time that OT = On Topic) and the Recently Closed Tabs is dead-on accurate. I've never had to use the Hamburger for the purpose of recent history.

The 'recently closed' list has always seemed to be reasonably accurate for me, although I have seen occasional anomalies where something which should be there isn't.

Why is it that the Recently Closed Tabs feature under the down-arrow and the address bar don't show me the last addresses I visited? This is probably so in all Chrome variants and Google is employing some "wisdom". If I open Hamburger -> History -> History, the last entries are not the same as shown under Recently Closed Tabs. The way I see it, this recent exists to save me from using the Hamburger menu. Also, if I have visited example.com/a example.com/b example.com/c in this order, A may persistently show in the address bar auto-complete even though I haven't opened it for some time.

LordGarfio said: "By the way, your browser may show the following URL https://license.malcat.fr/?, because the two mentioned above cannot do it and all everything is correct in the URL https://malcat.fr/" My NM28 Version: 28.10.7a1 (32-bit)(2025-10-16) that works with PayPal has the following URL: https://www.paypal.com/signin SHA1 Fingerprint 12:2E:F7:4B:EE:2D:15:15:4F:4E:DE:63:48:9F:DB:6C:A1:B4:A4:6D ---------------------------------------------------------------------------------------- My Serpent 52.9.0 (2025-10-16)(32-bit) that does NOT work with PayPal because the PayPal captcha does NOT load (If PayPal worked had the captcha loaded is a different question.) has the following URL: https://www.paypal.com/signin?locale.x=de_DE SHA1 Fingerprint 12:2E:F7:4B:EE:2D:15:15:4F:4E:DE:63:48:9F:DB:6C:A1:B4:A4:6D

this site uses await, which is filed in https://repo.palemoon.org/MoonchildProductions/UXP/issues/2229

LordGarfio said: "By the way, your browser may show the following URL https://license.malcat.fr/?, because the two mentioned above cannot do it and all everything is correct in the URL https://malcat.fr/" My NM28 Version: 28.10.7a1 (32-bit)(2025-10-16) that works with PayPal has the following URL: https://www.paypal.com/signin SHA1 Fingerprint 12:2E:F7:4B:EE:2D:15:15:4F:4E:DE:63:48:9F:DB:6C:A1:B4:A4:6D

Agreed!

You wouldn't have that "problem" if you didn't spend 'all your time' to go through *ALL* posts by your 'favorite people' and post "likes" to content in JANUARY OF 2022 for NO REASON other than it was one of your "favorite people". "Kids these days..." edit: when you're done with all the "likes" for your 'favorite people', I'd kind of suggest you (and the other's that "do this") at least kind of EVEN THE SCORE by doing the SAME for folks like roytam, feodor, tihiy, et cetera.