ScummVM IRC Log - 08 Jan 2026

[Back to Index]

[00:04] <grogbot> [discord] <dwa> Hi @bosca_bo
[00:04] <grogbot> [discord] <dwa> I'm back to my big-endian machines here. I've posted my findings in that Sword1 bug report:
[00:04] <grogbot> [discord] <dwa> https://bugs.scummvm.org/ticket/16393#comment:1
[00:04] <grogbot> [discord] <dwa> Basically, it looks like the endianness problem is only triggered with _some_ file formats we support, among the various releases
[00:04] <grogbot> [discord] <dwa> the one where I could reproduce the problem are the GOG "original release (bonus)" (with Vorbis speech/audio, that seems important) and the PSX demo we host
[00:04] <grogbot> [discord] <dwa> then, the problem appears on that reference big-endian VM we have
[00:04] <grogbot> [discord] <dwa> WARNING: put the volume at a very low-level, because the buggy audio can be ear-damaging
[00:07] <grogbot> [discord] <dwa> I've spent a couple of hours on that, and came to a change similar to what you tested and then reverted
[00:07] <grogbot> [discord] <dwa> it'd unbreak that GOG release, but then it'd also break some other ones (see the Trac ticket)
[00:07] <grogbot> [discord] <dwa> I wonder if it could be related to those `readBuffer((int16 *)buf, ...` calls, especially that cast.
[00:08] <grogbot> [discord] <dwa> I've spent a couple of hours on that, and came to a change similar to what you tested and then reverted
[00:08] <grogbot> [discord] <dwa> it'd unbreak that GOG release, but then it'd also break some other ones (see the Trac ticket)
[00:08] <grogbot> [discord] <dwa> I wonder if it could be related to those `readBuffer((int16 *)buf, ...` calls, especially that cast.
[00:08] <grogbot> [discord] <dwa> I have a GDB session attached to it, if it makes easier to compare things (edited)
[00:37] <-- sev left irc: Quit: This computer has gone to sleep
[01:57] <-- unilynx left irc: Quit: My MacBook has gone to sleep. ZZZzzz&
[04:01] <-- stanrifkin left irc: Quit: Leaving
[05:11] --> Strengths joined #scummvm.
[06:33] <Scummette> [scummvm] dependabot[bot] pushed 1 new commits to dependabot/pip/doc/docportal/urllib3-2.6.3: https://is.gd/GGg7tB
[06:33] <Scummette> scummvm/dependabot/pip/doc/docportal/urllib3-2.6.3 e121c24 dependabot[bot]: Bump urllib3 from 2.6.0 to 2.6.3 in /doc/docportal
[06:33] <Scummette> [scummvm] dependabot[bot] opened pull request #7133: Bump urllib3 from 2.6.0 to 2.6.3 in /doc/docportal (master...dependabot/pip/doc/docportal/urllib3-2.6.3) https://is.gd/IfHzrP
[07:17] --> DominusExult joined #scummvm.
[07:20] <-- Dominus left irc: Ping timeout: 240 seconds
[07:20] Nick change: DominusExult -> Dominus
[08:01] <-- Smedles left irc: Ping timeout: 246 seconds
[08:01] --> Smedles joined #scummvm.
[08:04] --> ldevulder joined #scummvm.
[08:31] --> borosky joined #scummvm.
[09:22] <grogbot> [discord] <bhumikasharma0909> hii everybody here Bhumika Sharma looking forward to contribute with honestly!!
[09:26] <grogbot> [discord] <gu3> if you want to discussion contributions for the gsoc, check the #
[09:35] <grogbot> [discord] <bhumikasharma0909> where is repository
[09:36] <grogbot> [discord] <bhumikasharma0909> where do i get
[09:40] <grogbot> [discord] <bmsbotje> https://wiki.scummvm.org/index.php?title=Developer_Central
[09:40] <grogbot> [discord] <bmsbotje> https://wiki.scummvm.org/index.php?title=Developer_Central has all you need to get started. (edited)
[10:08] <grogbot> [discord] <.eriktorbjorn> Not really a screenshot, but I think it's interesting that Sierra must have licensed an existing book to use with their game Gold Rush. Because the was released in 1988, but the book was apparently from 1980.
[10:08] <grogbot> [discord] <.eriktorbjorn> Though even if you manage to match up the pages, you can't always use the original book to get through the game's copy protection. Several of the questions are about things Sierra added to their version.
[10:08] <grogbot> [discord] <.eriktorbjorn> https://cdn.discordapp.com/attachments/581224061091446795/1458764216860282910/image.png?ex=6960d38b&is=695f820b&hm=2fc5b78afe5e578b41a8ca5ae66c7d6a2a67ee6656294d87999786c596883564&
[10:10] <Scummette> [scummvm] dreammaster pushed 4 new commits to master: https://is.gd/nCdmMh
[10:10] <Scummette> scummvm/master a623c6e dreammaster: ULTIMA: ULTIMA0: Match dungeon wall colors to original
[10:10] <Scummette> scummvm/master e1e9cd0 dreammaster: ULTIMA: ULTIMA0: Monsters touching hidden door tiles reveals them
[10:10] <Scummette> scummvm/master 52a182d dreammaster: ULTIMA: ULTIMA0: Added monster console comand
[10:14] <grogbot> [discord] <.eriktorbjorn> Probably harder to find loose copies of the Sierra version than of the original. I got lucky. 🙂
[10:22] --> sev joined #scummvm.
[10:22] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[12:10] <grogbot> [discord] <strangerke> It's one of the rare Sierra games I didn't appreciate at all (with Codename Iceman), so I didn't try to find it at all
[12:10] <grogbot> [discord] <strangerke> I didn't remember it's so rare
[12:11] <grogbot> [discord] <strangerke> Is it easier to find Manhunter 1 & 2 ?
[12:34] <-- sev left irc: Quit: This computer has gone to sleep
[12:47] <grogbot> [discord] <.eriktorbjorn> I bought it because I remembered playing it back in the eighties so I felt I should buy it this time, when the original authors were nice enough to re-release it. Though I still haven't played it since then.
[12:47] <grogbot> [discord] <.eriktorbjorn> I assume the Manhunter games are harder to find, since they were never re-released on CD or digitally as far as I know? (I don't have a working floppy drive.)
[12:50] <grogbot> [discord] <.eriktorbjorn> I tried sending an email to Evryware to ask if the games were available anywhere - they do list them on https://www.evryware.com/chronology.html as part of their history - but never received a reply. Maybe they don't actually exist any more, and it's just an abandoned web page?
[12:54] <grogbot> [discord] <.eriktorbjorn> I only remember a few things about playing Gold Rush. Leaving home, choosing one of the routes, lots of instant deaths, look for clues, grind for gold, finding your brother... Probably rather repetitive in parts, but I think it's fun to compare it to the early AGI games. Lots of stuff going on in at least the early scenes.
[12:55] <grogbot> [discord] <.eriktorbjorn> That's why I'd like to be able to buy the Manhunter games, even though what I've seen of them looks stuffed with the kind of instant deaths, arcade bits, and moon logic that I don't like about the early Sierra games.
[12:55] <grogbot> [discord] <.eriktorbjorn> And it's why I think it was such a missed opportunity when they didn't include the AGI version of King's Quest 4 on at least the Roberta Williams Anthology.
[12:59] <grogbot> [discord] <strangerke> yes Manhunter is very bizarre
[12:59] <grogbot> [discord] <strangerke> but I have some good memories of them
[13:05] --> sev joined #scummvm.
[13:05] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[13:09] <grogbot> [discord] <strangerke> https://www.befr.ebay.be/itm/256886925698?_skw=manhunter+sierra&itmmeta=01KEEVMCX0DMJAENWX1615CTN2&hash=item3bcfa76982:g:QP4AAeSwTMFn9XER&itmprp=enc%3AAQAKAAAA8FkggFvd1GGDu0w3yXCmi1cR39FTrb1K%2FoTGPK8zBgRqgJAawd%2FcHe9CBKNYzTFScIX2oYOj9V%2F49NOkosbs5cbBxzQclqa0cXTmhBwtoh%2Fmyj6%2BHpRZXDIscsil83UUWpXcI2q4ZfsMOfc%2F%2Bg7BlLdSVx0kq7j5UjkJEuFfk3lAc1H8oUHjURSgtqBor75hReM6wjh0VVEDlMZbV6EEubzp3FCwPPT5wCnSA4EoJxQAp3ULLJQX0eMkTVppLFBDxr92Akt
[13:10] <grogbot> [discord] <strangerke> for the fair price of 350eur...
[13:10] <grogbot> [discord] <strangerke> Eek
[13:34] <-- sev left irc: Quit: This computer has gone to sleep
[13:35] <grogbot> [discord] <.eriktorbjorn> And I don't have a floppy drive any more, as I said.
[13:39] <grogbot> [discord] <.eriktorbjorn> I don't remember what I paid for Gold Rush! but it wasn't that, even though I got it in a box. It would have contained floppies, but I asked them if there was any alternative... so mine has a CD-R in it instead. Later I bought a regular CD release, and put it in the box. And now I can add some old manuals to it as well.
[13:39] <grogbot> [discord] <.eriktorbjorn> I'm sure there are collectors who would scream in horror at this abomination, but at least it's pretty unique at this point. 🙂
[13:39] <grogbot> [discord] <.eriktorbjorn> https://cdn.discordapp.com/attachments/581224061091446795/1458817494965027040/california_gold_collection.jpg?ex=6961052a&is=695fb3aa&hm=d0c327ed7292cfc91b0a6ed54fa5b0664d9290555dd4b9e44d2d4f360c806f5e&
[13:52] --> sev joined #scummvm.
[13:52] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[14:51] <Scummette> [scummvm] whoozle pushed 1 new commits to master: https://is.gd/Rav2QW
[14:51] <Scummette> scummvm/master af75af9 ccawley2011: AGDS: Add detection for the demo version of Black Mirror (#7131)
[14:51] <Scummette> [scummvm] whoozle closed pull request #7131: AGDS: Add detection for the demo version of Black Mirror (master...black-mirror-demo) https://is.gd/4XLwlV
[14:56] <grogbot> [discord] <bosca_bo> Thank you SO SO much for the thorough analysis, appreciate it! 😄 I'm not in front of the code right now but:
[14:56] <grogbot> [discord] <bosca_bo> I know there's a function within the sound routines which detects whether the current speech file is of another endianness and swaps it. Could it be causing issues?
[14:56] <grogbot> [discord] <bosca_bo> Also, could you please recommend a low cost real BE system I could buy on ebay so that I could be of better help in these cases?
[15:00] <-- sev left irc: Quit: This computer has gone to sleep
[15:02] --> sev joined #scummvm.
[15:02] <-- sev left irc: Changing host
[15:02] --> sev joined #scummvm.
[15:02] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[15:18] --> unilynx joined #scummvm.
[15:32] <-- sev left irc: Quit: This computer has gone to sleep
[15:41] <grogbot> [discord] <dwa> you're welcome!
[15:41] <grogbot> [discord] <dwa> > I know there's a function within the sound routines which detects whether the current speech file is of another endianness and swaps it. Could it be causing issues?
[15:41] <grogbot> [discord] <dwa> that part seems not to be triggered at all in the GOG release with the Vorbis files
[15:41] <grogbot> [discord] <dwa> > Also, could you please recommend a low cost real BE system I could buy on ebay so that I could be of better help in these cases?
[15:41] <grogbot> [discord] <dwa> My most convenient big-endian machine is a PowerBook5,6 (1 CPU @ 1.66 GHz), but putting an SSD (and a compatible IDE adaptator) in it is a bit time consuming
[15:41] <grogbot> [discord] <dwa> all OSes for it have their problems: OSXPPC debugger is very broken on C++11. Debian 8 (used in the VM) is way out of date, but stable. Modern Debian powerpc breaks more and more. OpenBSD is stable (except for the general Xorg/Mesa regressions on big-endian), but slow and the tooling is old-school.
[15:41] <grogbot> [discord] <dwa> buying a Raspberry Pi 3 (not newer) is another option; but NetBSD is the only system proving a big-endian ARM build for it
[15:42] <grogbot> [discord] <lephilousophe> ah! that's the info I missed. 🙂 I wondered how you made run Linux in BE on a RPi
[15:42] <grogbot> [discord] <criezy> The original mac version may have speech data either as BE or LE depending on the language / version / unknown criteria. This is the reason why we have an heuristic to detect the endianness, but it is only used (to set `_bigEndianSpeech`) for the mac version. For all other versions it assumes LE data I think.
[15:42] <grogbot> [discord] <dwa> I have one RaspPi running it, but only the base OS. I didn't test the X11/toolchain yet
[15:43] <grogbot> [discord] <criezy> If the issue is with the Vorbis speech, isn't the issue in the `CowVorbis` case in `expandSpeech` with the VorbisStream provinding BE data?
[15:46] <grogbot> [discord] <criezy> If the issue is with the Vorbis speech, isn't the issue in the `CowVorbis` case in `expandSpeech` with the VorbisStream provinding BE data (or native endianess data?) instead of LE data? (edited)
[15:49] <grogbot> [discord] <criezy> Ah indeed, the `AudioStream::readBuffer` doc indicates "Data must be in native endianness, 16 bits per sample, signed." So I guess in `expandSpeech` for the cases where it is using an AudioStream (MP3, Vorbis, FLAC) we need to convert from native endianess to LE. Or we need for thoses cases (and only thoses cases) in playSpeech to indicate the data is in native endianess (and in other cases in LE).
[15:53] --> sev joined #scummvm.
[15:53] <-- sev left irc: Changing host
[15:53] --> sev joined #scummvm.
[15:53] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[15:53] <grogbot> [discord] <criezy> Note that for the original mac versions currently if _bigEndianSpeech is true it converts the data to LE. So in all cases we have LE data. But I guess this could be changed to keep data BE and indicate that in playSpeech when we create the raw stream. I am not sure this would make much of a difference in term of performances through. So it may be easier to keep the assumption data is always LE (and convert BE to LE when needed).
[16:00] <-- sev left irc: Quit: This computer has gone to sleep
[16:56] unilynx (textual@2603:7081:8f0:4c0:2d80:9e24:4df2:26f5) left #scummvm.
[17:10] <grogbot> [discord] <bosca_bo> thanks for the pointer folks!
[17:10] <grogbot> [discord] <bosca_bo> would this patch work then? (cc @kebabounet )
[17:10] <grogbot> [discord] <bosca_bo> https://cdn.discordapp.com/attachments/581224061091446795/1458870445846823055/fix_bigendian_speech.patch?ex=6961367a&is=695fe4fa&hm=ed070709df5ef7f9414a7a772002bf38f5ed3217308b0961faceeb06219b217c&
[17:38] <grogbot> [discord] <Digitall> You can still buy Gold Rush - Economy Edition for 19.95 USD from here: https://www.softwarefarm.com/gr_economy.htm
[17:38] <grogbot> [discord] <Digitall> I ordered it a while back and it turned up fine
[17:39] <grogbot> [discord] <criezy> This needs to be tested, but this looks like it should fix the issue.
[17:39] <grogbot> [discord] <criezy> Also good spot for the serveSample function. It seems it should also have the same issue as speech for compressed music. It surprised me at first because I am sure that when we initially implemented the AIF audio (for mac version) I still had a PPC mac and it was working properly. But I forgot there was a bug rewite of the code since then (PR 5348) and that the issue was probably introduced at that time.
[17:39] <grogbot> [discord] <criezy> Also good spot for the serveSample function. It seems it should indeed also have the same issue as speech for compressed music. It surprised me at first because I am sure that when we initially implemented the AIF audio (for mac version) I still had a PPC mac and it was working properly. But I forgot there was a bug rewite of the code since then (PR 5348) and that the issue was probably introduced at that time. (edited)
[17:39] <grogbot> [discord] <criezy> Also good spot for the serveSample function. It seems it should indeed also have the same issue as speech for compressed music. It surprised me at first because I am sure that when we initially implemented the AIF audio (for mac version) I still had a PPC mac and it was working properly. But I forgot there was a big rewite of the code since then (PR 5348) and that the issue was probably introduced at that time. (edited)
[17:40] <grogbot> [discord] <bosca_bo> Indeed, when I rewrote the audio system I definitely didn't test it enough on BE
[17:41] <grogbot> [discord] <Digitall> To clear... Their idea of ecommerce is a phone number and e-mail address. I think I paid with Credit Card over phone after talking to them.
[17:42] <-- Scummette left irc: Ping timeout: 264 seconds
[17:56] --> Scummette joined #scummvm.
[17:57] <grogbot> [discord] <dwa> it seems to be fine on big-endian with this patch, yes!
[17:57] <grogbot> [discord] <dwa> for the GOG release, the Win demo, the Macintosh demo
[17:57] <grogbot> [discord] <dwa> (just quickly moved in a few rooms, no deep testing)
[17:58] <grogbot> [discord] <dwa> the PSX demo crash when building with assert() on, though... attaching the GDB backtrace

[17:59] <grogbot> [discord] <dwa> https://cdn.discordapp.com/attachments/581224061091446795/1458882730153611409/gdb-assert-sword1-demo-psx-debian-powerpc-with-big-endian-patch.txt?ex=696141eb&is=695ff06b&hm=a72fc04a3d39e9cb5afcde1674362775782ac595d9bf23f6ec044ad117cbdb5f&
[18:03] --> sev joined #scummvm.
[18:03] <-- sev left irc: Changing host
[18:03] --> sev joined #scummvm.
[18:03] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[18:10] <grogbot> [discord] <.eriktorbjorn> There was a regular CD release too, but of course that was years ago. It was nice because it included some concept art. I think those are also included with "Gold Rush! Classic" on Steam. (Not to be confused with "Gold Rush! Anniversary" which is the remake, or "Gold Rush! 2".)
[18:10] <grogbot> [discord] <.eriktorbjorn> There's even more in the "Special Edition" of the remake, but many of them are pretty faded. (They're drawings, not paintings.)
[18:14] <grogbot> [discord] <bosca_bo> Woohoo!
[18:15] <grogbot> [discord] <bosca_bo> Dang
[18:15] <grogbot> [discord] <bosca_bo> was psx demo like this even before the patch?
[18:19] <Scummette> [scummvm] whoozle pushed 1 new commits to master: https://is.gd/LRe2qy
[18:19] <Scummette> scummvm/master 9e0d874 whoozle: AGDS: add support for the demo version of Black Mirror
[18:46] <Scummette> [scummvm] tag2015 pushed 1 new commits to master: https://is.gd/wE8DMV
[18:46] <Scummette> scummvm/master f585e6a tag2015: AGS: Update blackwell2 / gemini rue entries
[18:56] <grogbot> [discord] <dwa> Yes, that crash is not introduced by the new patch
[18:56] <grogbot> [discord] <dwa> for some reason, that crash only happens on that Debian powerpc env, hm
[18:56] <grogbot> [discord] <dwa> (same files, same Git branch&)
[19:05] <grogbot> [discord] <dwa> I thought it could be due to a file descriptor leak, especially when it deals with the "brokent tunes.dat file of the PSX demo" (which is the file being handled when it crashes)
[19:05] <grogbot> [discord] <dwa> and I think this could fix it:
[19:05] <grogbot> [discord] <dwa> ```diff
[19:05] <grogbot> [discord] <dwa> diff --git a/engines/sword1/sound.cpp b/engines/sword1/sound.cpp
[19:05] <grogbot> [discord] <dwa> index d38ce46512c..43df52bdad0 100644
[19:05] <grogbot> [discord] <dwa> --- a/engines/sword1/sound.cpp
[19:05] <grogbot> [discord] <dwa> +++ b/engines/sword1/sound.cpp
[19:05] <grogbot> [discord] <dwa> @@ -516,13 +516,13 @@ void Sound::startSpeech(uint16 roomNo, uint16 localNo) {
[19:05] <grogbot> [discord] <dwa> break;
[19:05] <grogbot> [discord] <dwa> }
[19:05] <grogbot> [discord] <dwa>
[19:05] <grogbot> [discord] <dwa> + file.close();
[19:05] <grogbot> [discord] <dwa> +
[19:05] <grogbot> [discord] <dwa> if (locIndex == 0xFFFFFFFF) {
[19:05] <grogbot> [discord] <dwa> warning("Could not find local number %d in room %d in speech.inf", roomNo, localNo);
[19:05] <grogbot> [discord] <dwa> return;
[19:05] <grogbot> [discord] <dwa> }
[19:05] <grogbot> [discord] <dwa>
[19:05] <grogbot> [discord] <dwa> - file.close();
[19:05] <grogbot> [discord] <dwa> -
[19:05] <grogbot> [discord] <dwa> index = _cowHeader[(roomOffset + locIndex) * 2];
[19:05] <grogbot> [discord] <dwa> sampleFileSize = _cowHeader[(roomOffset + locIndex) * 2 + 1];
[19:05] <grogbot> [discord] <dwa> } else {
[19:05] <grogbot> [discord] <dwa> @@ -1025,8 +1025,10 @@ bool Sound::prepareMusicStreaming(const Common::Path &filename, int newHandleId,
[19:05] <grogbot> [discord] <dwa> // The PSX demo has a broken/truncated tunes.tab. So we check here
[19:05] <grogbot> [discord] <dwa> // that the offset is not beyond the end of the file.
[19:05] <grogbot> [discord] <dwa> int32 tableOffset = (tuneId - 1) * 8;
[19:05] <grogbot> [discord] <dwa> - if (tableOffset >= tableFile.size())
[19:05] <grogbot> [discord] <dwa> + if (tableOffset >= tableFile.size()) {
[19:05] <grogbot> [discord] <dwa> + tableFile.close();
[19:05] <grogbot> [discord] <dwa> return false;
[19:09] <grogbot> [discord] <dwa> for some reason, when starting the PSX demo,
[19:09] <grogbot> [discord] <dwa> prepareMusicStreaming() first is for tuneId=7, on my regular desktop (when the PSX menu shows up)

[19:10] <grogbot> [discord] <dwa> ThreadSanitizer finds several data races on a regular desktop with the PSX demo, @bosca_bo
[19:10] <grogbot> [discord] <bosca_bo> 😥😥😥
[19:11] <grogbot> [discord] <bosca_bo> Thanks for the investigation, any indication about where they are triggered?
[19:12] <-- ldevulder left irc: Ping timeout: 240 seconds
[19:12] <grogbot> [discord] <dwa> the ThreadSanitizer data races? when the first scene shows up
[19:13] <grogbot> [discord] <dwa> I can send you the logs if you want. I don't know if MSVC can use ThreadSanitizer when using the Clang backend
[19:53] <Scummette> [scummvm] AndywinXp pushed 1 new commits to master: https://is.gd/KiZHpT
[19:53] <Scummette> scummvm/master 152445e AndywinXp: SWORD1: Attempt at fixing #16393
[19:56] <Scummette> [scummvm] AndywinXp pushed 1 new commits to branch-3-0: https://is.gd/SnygBr
[19:56] <Scummette> scummvm/branch-3-0 5d416fc AndywinXp: SWORD1: Attempt at fixing #16393
[20:39] <-- sev left irc: Quit: This computer has gone to sleep
[20:39] --> sev joined #scummvm.
[20:39] <-- sev left irc: Changing host
[20:39] --> sev joined #scummvm.
[20:39] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[20:42] <-- Scummette left irc: Ping timeout: 264 seconds
[20:55] --> Scummette joined #scummvm.
[21:10] --> ccawley2011 joined #scummvm.
[21:10] #scummvm: mode change '+o ccawley2011' by ChanServ!ChanServ@services.libera.chat
[21:36] <-- ccawley2011 left irc: Ping timeout: 244 seconds
[21:57] <-- sev left irc: Quit: This computer has gone to sleep
[21:58] <-- borosky left irc: Ping timeout: 240 seconds
[22:13] <grogbot> [discord] <railtaff> why does the font look weird
[22:13] <grogbot> [discord] <railtaff> https://cdn.discordapp.com/attachments/581224061091446795/1458946649022922970/image.png?ex=69617d73&is=69602bf3&hm=071b05c2c20149cc3702500ace8094a44bdda33e4d25351cddccce935d5ed2ec&
[22:34] --> sev joined #scummvm.
[22:34] #scummvm: mode change '+o sev' by ChanServ!ChanServ@services.libera.chat
[22:54] --> unilynx joined #scummvm.
[23:01] <grogbot> [discord] <sev____> please define "weird"
[23:02] <grogbot> [discord] <dwa> ah, found a more reliable way of triggering the crash -- "good" news is that it's then possible to trigger it on a regular PC 😄
[23:02] <grogbot> [discord] <dwa> I've tracked it in https://bugs.scummvm.org/ticket/16443 with a reproducer. Seems then totally unrelated to big-endian matters
[23:03] <grogbot> [discord] <railtaff> non bolded text is pixelated
[23:03] <grogbot> [discord] <railtaff> or is that intended
[23:03] <grogbot> [discord] <railtaff> well both font weights are pixelated actually
[23:03] <grogbot> [discord] <sev____> well, it is not intended, but it is one of the known problems indeed
[23:04] <grogbot> [discord] <sev____> this particular font has no antialiasing by design, it is an old Mac font, because when we switch here to TTF rendering, it is still not anti-aliased
[23:05] <grogbot> [discord] <sev____> something is missing there, most probably blending with background
[23:08] <-- unilynx left irc: Quit: Textual IRC Client: www.textualapp.com
[00:00] --- Fri Jan 9 2026