All Activity

My Pentium D runs @ 3Ghz and that might explain why performance on my computer is better. I'm happy with it.

Yes, I think so. _video.py.diff.zip I use the `git diff` command to make diff files. I presume you have git for Windows installed? The last version to support Windows Vista and Server 2008 was v2.37.1 https://git-scm.com/docs/git-diff

That link worked OK , so thanks , but I'm afraid new "_video.py.diff" ISN'T what I meant in my previous post ; isn't how the (whole) diff should look, but it just contains the segment/lines of your original "_video.py.diff" (for which the link now 404s) that I would modify, to make the code appear as it was before upstream commit 895e762 ; do you now get it? In your second "_video.py.diff", you left out the segment for "testing" the iOS HLS streams (which 895e762 also removed) and this part is indispensable; Windows lacks a native diff app, so I usually employ a Text Editor to modify source; for your sake, I installed a diff.exe Windows port and modifying the latest upstream version of file "_video.py", this is the patch I came up with: --- _video.py.a 2025-09-23 20:56:26.965800000 +0300 +++ _video.py.b 2025-09-23 21:16:39.283800000 +0300 @@ -257,7 +257,7 @@ '401': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'vcodec': 'av01.0.12M.08'}, } _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'srt', 'vtt') - _DEFAULT_CLIENTS = ('tv_simply', 'tv', 'web') + _DEFAULT_CLIENTS = ('tv_simply', 'tv', 'ios', 'web') _DEFAULT_AUTHED_CLIENTS = ('tv', 'web_safari', 'web') # Premium does not require POT (except for subtitles) _DEFAULT_PREMIUM_CLIENTS = ('tv', 'web_creator', 'web') @@ -3284,9 +3284,11 @@ f'{video_id}: {client_name} client {proto} formats require a GVS PO Token which was not provided. ' 'They will be skipped as they may yield HTTP Error 403. ' f'You can manually pass a GVS PO Token for this client with --extractor-args "youtube:po_token={client_name}.gvs+XXX". ' - f'For more information, refer to {PO_TOKEN_GUIDE_URL}') + f'For more information, refer to {PO_TOKEN_GUIDE_URL} . ' + 'To enable these broken formats anyway, pass --extractor-args "youtube:formats=missing_pot"') # Only raise a warning for non-default clients, to not confuse users. + # iOS HLS formats still work without PO Token, so we don't need to warn about them. if client_name in (*self._DEFAULT_CLIENTS, *self._DEFAULT_AUTHED_CLIENTS): self.write_debug(msg, only_once=True) else: @@ -3579,6 +3581,10 @@ if missing_pot: f['format_note'] = join_nonempty(f.get('format_note'), 'MISSING POT', delim=' ') f['source_preference'] -= 20 + # XXX: Check if IOS HLS formats are affected by PO token enforcement; temporary + # See https://github.com/yt-dlp/yt-dlp/issues/13511 + if proto == 'hls' and client_name == 'ios': + f['__needs_testing'] = True itags[itag].add(key) @@ -4389,6 +4395,7 @@ if upload_date and live_status not in ('is_live', 'post_live', 'is_upcoming'): # Newly uploaded videos' HLS formats are potentially problematic and need to be checked + # XXX: This is redundant for as long as we are already checking all IOS HLS formats upload_datetime = datetime_from_str(upload_date).replace(tzinfo=dt.timezone.utc) if upload_datetime >= datetime_from_str('today-2days'): for fmt in info['formats']: I hope it's now clear to you ... Kind regards.

Hmm... Must have made a mistake. Try this.

True, but even with just a single 4 gb 1333 MHz stick, the same slowdown behavior happens. Win98 works in strange ways from what I've seen - components that work perfectly in other setups, may suddenly start giving problems when used together. e.g. this RAM related slowdown doesn't happen if I use the same RAM, GPU, etc. in a G41MT-ES2L board

I have license for 1 PC, is possible upgrade this license to 3 PC, or is necessery buy another license for 2 pc?

=> ...

The EX58-UD4 supports triple channel memory. So it depends on where you put the second stick of memory and also depends on the speed of that second stick, if the second stick is slower than the first the BIOS will automatically set the memory speed to the speed of the slowest stick.

I used to use UltraEdit but now I use VS Code. Of the two IDEs I use, it is the only one with an set of LLMs built into it as well which helps often but not always. https://en.wikipedia.org/wiki/Visual_Studio_Code

Updated yt-dlp_x86 for Vista: link.

I do not own any Pentium D. Also, that seems a bit vague, "Pentium D" reveals a benchmark score of anywhere from a terrible 374 to an impressive 3240. That's only talking about the launch time for the GUI to load, that doesn't really reveal "good performance". For me, running all three v132s on the same exact hardware, Supermium is THE SLOWEST PERFORMANCE compared to "real" Chrome and "real" Chromium.

Unfortunately, none of these solutions helped. I think it may be a quirk with my particular hardware setup (X58 chipset more specifically) Thanks a lot mercury. This didn't resolve the issue (maybe some issue with Intel X58 or Gigabyte board that I have (EX58-UD4))

Bad news: Deno is only available as 64bit and for Windows 10 and higher...

Oh my bad, let me update that thread with the results for that particular issue

Yes. You asked for the help, got recommendation and demonstrated zero reaction after that. I assume you should to know that this is not smart behaviour.

Fixed with latest beta, thanks so much And with the latest beta, when you download a file and then go to another web page while it's downloading, the download progress bar in the taskbar is also present on the website. does not appear if you stay on the download page I didn't notice if this was the case with previous versions.

@VistaLover Here are new diff's. Is this what you mean?

@Tihiy Firstly, thank you for fixing the other bug I reported last year. Much appreciated. I have been having another issue now with a multi-monitor setup for the last few months, but I'm honestly unsure if its just a Windows 11 24H2 bug or a StartAllBack bug. It has to do with the placement of the Notification/Calendar Windows 11 fly-out. It's being incorrectly placed on the wrong monitor (the right monitor). For context, the screenshots below are taken on a triple monitor PC setup in the following layout (1920x1080, 2560x1440, 1920x1080). The main screen is the central screen (2560x1440). Issue present shortly after fresh boot: This is after killing explorer.exe and restarting it. Now in the correct position: If I swap to Windows 10 fly-outs, this issue doesn't happen as far as I'm aware.

Initial tests with a patched "2025.09.17.84413" yt-dlp version (the zip import "yt-dlp" file was patched locally) look promising ; testing was done with a "bare bones" setup, with no optional Python modules and no FFmpeg : yt-dlp --extractor-args "youtube:player_client=ios" -vF "O5TS2CbBhKo" [debug] Command-line config: ['--extractor-args', 'youtube:player_client=ios', '-vF', 'O5TS2CbBhKo'] [debug] Encodings: locale cp1253, fs utf-8, pref cp1253, out utf-8 (No VT), error utf-8 (No VT), screen utf-8 (No VT) [debug] yt-dlp version nicolaasjan/yt-dlp@2025.09.17.084413 [e123a48f1] (zip) [debug] Python 3.11.13 (CPython x86 32bit) - Windows-Vista-6.0.6003-SP2 (OpenSSL 3.0.16 11 Feb 2025) [debug] exe versions: none [debug] Optional libraries: sqlite3-3.45.1 [debug] Proxy map: {} [debug] Request Handlers: urllib [debug] Plugin directories: none [debug] Loaded 1833 extractors [debug] [youtube] [pot] PO Token Providers: none [debug] [youtube] [pot] PO Token Cache Providers: memory [debug] [youtube] [pot] PO Token Cache Spec Providers: webpo [youtube] Extracting URL: O5TS2CbBhKo [youtube] O5TS2CbBhKo: Downloading webpage [youtube] O5TS2CbBhKo: Downloading ios player API JSON WARNING: [youtube] O5TS2CbBhKo: ios client https formats require a GVS PO Token which was not provided. They will be skipped as they may yield HTTP Error 403. You can manually pass a GVS PO Token for this client with --extractor-args "youtube:po_token=iosgvs+XXX". For more information, refer to https://github.com/yt-dlp/yt-dlp/wiki/PO-Token-Guide . To enable these broken formats anyway, pass --extractor-args "youtube:formats=missing_pot" [youtube] O5TS2CbBhKo: Downloading m3u8 information [debug] Sort order given by extractor: quality, res, fps, hdr:12, source, vcodec, channels, acodec, lang, proto [debug] Formats sorted by: hasvid, ie_pref, quality, res, fps, hdr:12(7), source, vcodec, channels, acodec, lang, proto, size, br, asr, vext, aext, hasaud, id [info] Available formats for O5TS2CbBhKo: ID EXT RESOLUTION FPS | FILESIZE TBR PROTO | VCODEC VBR ACODEC MORE INFO --------------------------------------------------------------------------------------------------------------- sb3 mhtml 48x27 0 | mhtml | images storyboard sb2 mhtml 80x45 1 | mhtml | images storyboard sb1 mhtml 160x90 1 | mhtml | images storyboard sb0 mhtml 320x180 1 | mhtml | images storyboard 233 mp4 audio only | m3u8 | audio only unknown Untested, Default, low, IOS 234 mp4 audio only | m3u8 | audio only unknown Untested, Default, high, IOS 602 mp4 256x144 12 | ~ 2.55MiB 89k m3u8 | vp09.00.10.08 89k video only Untested, IOS 269 mp4 256x144 24 | ~ 5.55MiB 193k m3u8 | avc1.4D400C 193k video only Untested, IOS 603 mp4 256x144 24 | ~ 5.06MiB 176k m3u8 | vp09.00.11.08 176k video only Untested, IOS 229 mp4 426x240 24 | ~ 10.22MiB 356k m3u8 | avc1.4D4015 356k video only Untested, IOS 604 mp4 426x240 24 | ~ 8.23MiB 286k m3u8 | vp09.00.20.08 286k video only Untested, IOS 230 mp4 640x360 24 | ~ 23.51MiB 818k m3u8 | avc1.4D401E 818k video only Untested, IOS 605 mp4 640x360 24 | ~ 14.13MiB 492k m3u8 | vp09.00.21.08 492k video only Untested, IOS 231 mp4 854x480 24 | ~ 41.82MiB 1456k m3u8 | avc1.4D401E 1456k video only Untested, IOS 606 mp4 854x480 24 | ~ 22.32MiB 777k m3u8 | vp09.00.30.08 777k video only Untested, IOS 232 mp4 1280x720 24 | ~ 72.92MiB 2538k m3u8 | avc1.64001F 2538k video only Untested, IOS 609 mp4 1280x720 24 | ~ 33.23MiB 1157k m3u8 | vp09.00.31.08 1157k video only Untested, IOS 270 mp4 1920x1080 24 | ~138.27MiB 4813k m3u8 | avc1.640028 4813k video only Untested, IOS 614 mp4 1920x1080 24 | ~ 50.15MiB 1746k m3u8 | vp09.00.40.08 1746k video only Untested, IOS and actual download: yt-dlp --extractor-args "youtube:player_client=ios" -f 234 "O5TS2CbBhKo" [youtube] Extracting URL: O5TS2CbBhKo [youtube] O5TS2CbBhKo: Downloading webpage [youtube] O5TS2CbBhKo: Downloading ios player API JSON WARNING: [youtube] O5TS2CbBhKo: ios client https formats require a GVS PO Token which was not provided. They will be skipped as they may yield HTTP Error 403. You can manually pass a GVS PO Token for this client with --extractor-args "youtube:po_token=iosgvs+XXX". For more information, refer to https://github.com/yt-dlp/yt-dlp/wiki/PO-Token-Guide . To enable these broken formats anyway, pass --extractor-args "youtube:formats=missing_pot" [youtube] O5TS2CbBhKo: Downloading m3u8 information [info] Testing format 234 [info] O5TS2CbBhKo: Downloading 1 format(s): 234 [hlsnative] Downloading m3u8 manifest [hlsnative] Total fragments: 45 [download] Destination: Love and Money - Halleluiah Man (HQ Audio) [O5TS2CbBhKo].mp4 [download] 100% of 3.75MiB in 00:00:12 at 309.95KiB/s WARNING: O5TS2CbBhKo: Possible MPEG-TS in MP4 container or malformed AAC timestamps. Install ffmpeg to fix this automatically Best regards.

The EAX (in fact sound hardware acceleration) related issue is independent of the nvidia driver - same issue happened even if I have an ATI X850. No longer though, since I removed IRQ steering. And about RLoew's patches, yeah - as far as I remember I have PATCHMEM and TBPLUS installed - though I did check the EAX related issue before installing TBPLUS and it was still there, so I guess TBPLUS isn't at fault

It's good performance after the first run. First run sets up a bunch of stuff on your computer (files/folders) and that slows it down a little bit. After that it's alright. And yes, YouTube works like you'd expect. There is a performance setting in settings. If you have ever used Chrome it's pretty much the same layout and should be familiar to you.

I think NOTHEARTOPLAYGAMES has answered your portable question. The version I downloaded is the 132 etc, etc version. I just picked the 32 bit version for XP but didn't notice if there was a portable version. . I suggest you try it. The first time I installed it I chose the "ungoogled" version. Problem was that eliminates Google as being one of the search engines you can make as the default. Since I regularly use Google for searching, I uninstalled it and re-installed it without checking the "ungoogled" mode. There is no "ungoogled" version of Supermium. It is a choice you make during the install.

Greetings ; let me just say that I'm not in a position to make any demands on you; I hope my previous post didn't come off as that (a demand) ; I just picked up from the observation @Reino made previously and, as a courtesy to him, I suggested a way for him to continue fetching iOS (i.e. HLS) YT streams without the need to hunt for and extract POT strings from a browser (or some other way) ; I have no personal interest on this, other than the stated dislike of available (download) options having been removed ... With that out of the way , my Python-fu is certainly worse than yours ; I've now seen that file "_video.py" has been further modified by upstream since 895e762, as can be seen here, so now it isn't just a matter of partially reverting that commit, as I had (foolishly) suggested before... Upstream have changed priorities of yt_clients in 3bd9154, so now iOS would have to be added as third? Let's hope this doesn't break things... In your downloaded "_video.py.diff" patch, I would personally stick closer to how things were on 895e762 , so, instead of: f'For more information, refer to {PO_TOKEN_GUIDE_URL}') + 'To enable these broken formats anyway, pass --extractor-args "youtube:formats=missing_pot"') I would go with: f'You can manually pass a GVS PO Token for this client with --extractor-args "youtube:po_token={client_name}.gvs+XXX". ' - f'For more information, refer to {PO_TOKEN_GUIDE_URL}') + f'For more information, refer to {PO_TOKEN_GUIDE_URL} . ' + 'To enable these broken formats anyway, pass --extractor-args "youtube:formats=missing_pot"') # Only raise a warning for non-default clients, to not confuse users. + # iOS HLS formats still work without PO Token, so we don't need to warn about them. if client_name in (*self._DEFAULT_CLIENTS, *self._DEFAULT_AUTHED_CLIENTS): However, as the English say, "the proof is in the pudding", so the patched code should be (compiled and) tested first for inadvertent "malfunctions" ... Thanks already ...

Probably an RF issue. Try different: Wifi channels / power levels shielding / reflectors USB ports / hubs / cables monitor refresh rates cpu and system clock speeds

Thanks for the update. Was this also before the Nvidia driver install, or after? And are any of RLoew's patches still installed?