10.9, but yes. It seems reasonable.
The requirement change to macOS 10.11+ is discussed here:
https://github.com/libsdl-org/SDL/issues/7636
Supposedly before 2.24.0, SDL supported mac OS 10.7+, not 10.6. (IIRC macOS
10.6 has an incomplete set of 64-bit APIs so it's common to say 64-bit apps
Closed #1242 as completed.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1242#event-10301276323
You are receiving this because you are subscribed to this thread.
Message ID: ___
Ohrrpgce mailing list
Hmm. Well no need for workarounds, we can just update the SDL builds (from
steam-runtime) by running `linux/update_steam_runtime_libs.sh`, which I've just
done. steam-runtime is still using SDL 2.26.5.
--
Reply to this email directly or view it on GitHub:
BTW, adding a constructor here isn't necessary, as FB always
zero-initialises all variables & members (unless you write = ANY).
Storing the slice ptr in GraphSlice was just a kludge I found necessary
there (but not in MenuDefSlice) but it probably is cleaner than passing in
the Slice ptr to the
Interesting, it seems to be specific to certain SDL2 versions, or maybe builds.
Can you reproduce it when running game.sh? When I run game.sh (which uses SDL
2.0.22) it happens about 80-90% of the time. When I use my distro's SDL 2.0.26
build it never happens. Probably an SDL2 bug that was
Looking through the SDL changelog I notice:
2.26.5 (Apr 6 2023)
> The minimum deployment target on macOS is now 10.11, due to changes in the
> latest Xcode update
2.26.1 (Dec 2, 2022)
> Fixed building with older Xcode and macOS SDK
2.26.0 (Nov 22, 2022)
> Implemented vsync synchronization on
I see that you upgraded (`get_sdl_frameworks.sh`) from SDL 2.0.14 to 2.24.0 a
year ago, which was presumably when we dropped support for OS 10.6-10.9 without
realising it.
--
Reply to this email directly or view it on GitHub:
Oh good. And now I remember! I wanted to call this "inn menu" to match
other commands. We have:
main menu
items menu
spells menu
equip menu
save/load menu
order/team menu
debug menu
On the other hand the only command with "screen" in the name is "status
screen". (So "status menu" would be a good
nts in a
> for loop
>
Definitely.
>
> On Wed, Oct 11, 2023, 9:30 AM Ralph Versteegen via Ohrrpgce <
> ohrrpgce@lists.motherhamster.org> wrote:
>
>> A very useful command that's been sorely lacking!
>> The number of arguments is really unfortunate though, and it seems
A very useful command that's been sorely lacking!
The number of arguments is really unfortunate though, and it seems likely
that we'd want to add even more in future! I wanted to add a command to
concatenate one array of extras onto another. If we add such a command
(even reusing some of the code
Glad to see it's working again.
Is the From munging needed to prevent bounces, or is it to fix the spam
binning and impersonation warnings gmail puts on forwarded messages from
Github issues? It's kind of annoying, but the problems with Github emails
are worse, so hopefully they're fixed.
On Tue,
The actual part of the build log which shows the root cause of the error is
right at the top:
scons: Reading SConscript files ...
Using target: darwin arch: x86 fbc: fbc (1.06.0 (04-29-2016)) fbcc:
/opt/local/bin/gcc-mp-4.7 (gcc 4.7.4) cc: clang (LLVM 8.0.0) cctarget:
Also, this is the reason that the Mac build logs have lots of spurious
warnings: -Wno-missing-braces is not passed.
It's pretty poor that we're still using such an old forked FB on Mac, I
really need to fix the last couple things in trunk FB's Mac support (mostly
just crt.bi headers) so we can
I spoke too soon, looks like it was actually passing -fno-pie that caused
the problem, though I don't really understand it. Easy enough to just omit
it.
On Fri, 10 Nov 2023 at 19:15, Ralph Versteegen wrote:
> The actual part of the build log which shows the root cause of the error
> is right at
Haha! Yes, oops, sorry, I just haven't finished the code to actually invoke
zip_exec yet. Was doing 5 things at once.
Also it turns out that I already had another solution implemented in
distribmenu.bas but commented out, which was to use ziptool to rename
OHRRPGCE-Game.app inside an existing .zip
Thanks for the bug report. Does the problem go away once you close the game
and reopen it? If so, I think I know what caused this -- relaunching Steam
likely caused the controller to disappear and reappear, and I already knew
the code for handling that is wrong, although I didn't see it crash
Wobbler said he hasn't seen it since, and there doesn't seem to be any clue in
this report as to what the root cause was, so closing "works for me".
(Amusing: I noticed I marked this "works for me" back in March but didn't close
it, so I went looking for and found an open tab where I'd written
Closed #25.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/25#event-6642362013
You are receiving this because you are subscribed to this thread.
Message ID: ___
Ohrrpgce mailing list
In battles, the state of heroes, enemies, attack graphics and the weapon are
stored in the bslot() array, which is an array of BattleSprites. bslot 0-3 are
heroes, 4-11 enemies, 12-23 attack sprites (just 12 is used for most attacks),
and bslot(24) is the weapon. (BattleSprite has a huge amount
There really isn't anything we can do on backends other than gfx_sdl2 and
possible gfx_directx (which uses DirectInput which might sometimes provide
useful button metadata, but probably not) but I would not spend any time on
DirectInput.
On Fri, 20 May 2022 at 02:31, James Paige wrote:
> I like
There is currently no way to get the actual label of a button on the
player's controller. The "get scancode name" command can be used for
joystick buttons but returns a fixed name. There's also a separate system
in Custom for defining OS-specific button names which can be embedded in
strings with
here were some power outages on the weekend that
>> would have messed with the wifi, but that can't be the only problem
>>
>> On Mon., May 23, 2022, 10:34 p.m. Ralph Versteegen via Ohrrpgce, <
>> ohrrpgce@lists.motherhamster.org> wrote:
>>
>>> There seem
Do you know whether this has changed recently? I imagine it's because of the
recent switch to slices for the targeting cursor.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1240#issuecomment-1137991214
You are receiving this because you are
There seem to be a lot of problems with nightlies lately. Looking over the
last couple weeks almost every day one or more of the nightly build emails
is missing, the Android one especially. I think they haven't been running
rather than not sending email. 32 and 64-bit Linux builds haven't built in
Windows nightly builds now ship with SDL2.dll 2.0.22, containing scores more
gamepad fixes, so I would like to hear results of retesting this.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1215#issuecomment-1135383990
You are receiving this
Oh, also, AFAIU the default button mapping of all 8BitDo (and Switch and
GameCube and some other) gamepads has changed in recent SDL to match actual
button labels rather than based on layout (what they would be labelled on an
XBox controller). Maybe that's relevant.
--
Reply to this email
This bug has just been fixed in SDL_mixer git. There hasn't been an SDL_mixer
release in quite a while.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1196#issuecomment-1140127909
You are receiving this because you are subscribed to this
blem on my main computer, but it runs Debian 11 and
>> the build VMs are using an older Debian I think (10?)
>>
>> On Tue., May 24, 2022, 6:47 a.m. James Paige,
>> wrote:
>>
>>> I can check the VMs. There were some power outages on the weekend that
&g
Appears to have been fixed a few hours ago!
https://github.com/libsdl-org/SDL_mixer/issues/392
Note that the fix isn't in the current 2.5.1 RC but it looks like SDL_mixer
2.6.0 will soon be released.
--
Reply to this email directly or view it on GitHub:
Hey James,
I'm currently cleaning up licence files.
LICENSE-binary.txt links to http://HamsterRepublic.com/ohrrpgce/source.php
but that link no longer works, which is surprising because
http://HamsterRepublic.com/ohrrpgce/buglist.php (which is in
README-custom.txt) is a working redirect. I think
Sorry, I don't know how I missed this!
Ichorescent is way overdue and I'm not making much progress. I've been
planning on "this month" since last year!
"Blocker" is so narrowly defined! I have things I wanted to do which I
wouldn't strictly call blockers. I trimmed down my release TODO list by
Ah. But `docker/ohrrpgce-build-env-wine/Dockerfile` seems to download the
specific toolchain that's actually used on the build machine, as far as I can
tell.
Another two possible solutions:
* use a mingw.org toolchain instead
* If the build machine is upgraded to a prebuilt mingw-w64 7.0.0
Also, all newer versions of mingw-w64 presumably have the same problem with
`AddVectoredExceptionHandler`, because the code which calls that hasn't changed
since.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1241#issuecomment-1172908692
You
In nightlies (wasn't in hróðvitnir) using gfx_sdl2 on Linux/X11/xfce4, when
switching to fullscreen sometimes the image is not stretched to cover the
screen, but appears at the top-left of the screen at the same zoom level as in
windowed mode, and the rest of the screen is filled with garbage
Support for Windows 95 through 2000 is broken in nightlies, even when using
gfx_sdl builds (SDL2 requires Win XP+ anyway). This is due to a bug in the
version of mingw-w64 used on the build machine. A couple people asked me to get
it working again.
New builds needed
=
gt;>> scons: *** [ohrrpgce-game] Error 1
>>>>
>>>> I don't have this problem on my main computer, but it runs Debian 11
>>>> and the build VMs are using an older Debian I think (10?)
>>>>
>>>> On Tue., May 24, 2022, 6:47 a.m. James Paige,
>&
(10?)
>>>
>>> On Tue., May 24, 2022, 6:47 a.m. James Paige,
>>> wrote:
>>>
>>>> I can check the VMs. There were some power outages on the weekend that
>>>> would have messed with the wifi, but that can't be the only problem
>>>
I think the simplest way to add this would be as an option in the map resize
submenu to trim parts of the map that are totally blank.
However, I think this isn't a feature that would actually help much or make the
editor more ergonomic, so while anyone is welcome to implement it I wouldn't
I started trying to add win95 full and player builds to distrib-win.bat
(and distrib-wine.sh!) but it's a disaster, I would have to copy-paste most
of the file. This distrib script insanity has gone on too long! Maybe now
is the right time to deduplicate our umpteen packaging scripts as a single
Please file unrelated feature suggestions (and bugs) as multiple issues.
What do you mean by "basic" attacks? The hero's weapon attack? You can remove
that from the hero's battle menu in the hero editor. You'll need to edit each
hero's battle menu separately.
As for multiple types of MP,
Oh, I was meaning to add an option to blank out a sprite set. That's on my
secret ichorescent todo list!
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1238#issuecomment-1185485937
You are receiving this because you are subscribed to this
Also, I notice a lot of inconsistencies in docs/* packaging. In addition to
plotdict.xml, the plotdict icons and docs/more-docs.txt:
-Linux tarballs have docs/plotdictionary.html, hamsterspeak.html and
plotscripttutor.html (which are just links)
-Mac packages have docs/*.html
-Windows nightlies
Hey James,
I noticed (thanks to Rue's whatsnew discord bot, buggy though it may be)
that svn revisions aren't being mirrored to github and that all nightly
builds haven't happened for the last week.
___
Ohrrpgce mailing list
Recently Wobbler found that if you package a Mac app .zip from Windows it
doesn't run on a Mac, and today I see Spendidland has [hit the same
problem](https://splendidland.itch.io/franken). Distributing from Linux works.
Although I don't currently have access to a Mac to test, I assume the
This is related to #963 which maybe should already be closed, but seems to list
multiple bugs.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1239#issuecomment-1114666402
You are receiving this because you are subscribed to this thread.
オップナー2608 tried out compiling using FB 1.09. First problem is that the FB
bootstrap package doesn't include Darwin, so you need an existing fbc to
compile it, or need to create your own bootstrap. After that, it turns out FB
is, understandably, missing this extremely kludgy commit from my mac
I committed a 'fix' in r12173 but didn't close this issue. I don't know whether
anyone continued to suffer this issue (I think both Kiefkrack and Ravancloak
switched to Windows). However, looking back at it, I can't see how the 'fix'
can possibly be correct. `spawn_and_wait` calls `system()` to
> stats can always be increased up to INT_MAX
This is incorrect, stats are capped to 32767 (and below by -32768) in
`safesubtract` called from `inflict` when a stat is targetted by an attack (but
not when modified by another attack effect such as Absorb, "Reset target stat
to max before hit",
The build logs from the Mac VM are missing on some days including the most
recent ones when it should have pulled commits from svn and rebuilt.
If you look at the Mac VM could you please also upgrade SDL2.framework and
SDL2_mixer.framework to the latest releases.
Also, could you please downgrade
Oh, I fixed the script in r13071 for the new names. But the old version of
the script is still being run. Looks like it's run from ohr_nightly_vm.sh.
I think that may be the only cron script not run inside a VM, hence the
lack of auto-updating itself?
On Sun, 28 Aug 2022 at 00:58, James Paige via
Since James just downgraded to Euphoria 4.0.5 I've just tried out the latest
nightly in a Win 95 (OSR2) VM and found we're not there yet:
* game/custom don't run: `linked to missing export
KERNEL32.DLL:IsDebuggerPresent`
* hspeak.exe doesn't run: `linked to missing export
I found RTLD_NOLOAD in the headers for android-21 (Android 5.0) but not
android-19, but I guess it's not available because of our low
minSdkVersion. Anyway, it's not needed at all on Android so I got rid of it
there.
While that fixed the Android libapplication.so compile, I'm not actually
able to
Note that to find the download link Custom's build date and release name is
used. So if a +1 release occurs, or if a package is rebuilt with a new
release date and quietly uploaded as happened with
ohrrpgce-mac-minimal-2022-01-01-hrodvitnir-x86.tar.gz then the bugfix
release won't be used, and any
Unfortunately if you attempt to download a nonexistent file under
https://hamsterrepublic.com/ohrrpgce/nightly/ or
https://hamsterrepublic.com/ohrrpgce/archive/ (or
http://hamsterrepublic.com/ohrrpgce/symbols-archive/) you don't get a 404
error document, you get redirected to the wiki Main Page
Coincidental for it to (retroactively) break right after I rewrote
spawn_and_wait... (but that isn't used for Test Game anyway)
What are the last lines in c_debug.txt when importing scripts when it freezes?
(Importing prints more than Test Game does)
--
Reply to this email directly or view it
Game/Custom
--
As well as `IsDebuggerPresent`, winpthreads also uses `SetProcessAffinityMask`
& `TryEnterCriticalSection` which are missing in kernel32.dll, and the latter
can't be removed.
Now I understand that the 'win32' threading support in mingw-w64 is the
original one and
Wow there are a huge number of edge cases and special cases here, I regret
looking into this...
I see the problems were introduced in:
r12685 "Fix "Jump" (and the other new hiding attacks) to properly prevent
targetting when used by heroes".
r12874 "Fix check_for_unhittable_invisible_foe() to
Could even add a little javascript to show a tooltip on hover (pure CSS
isn't possible), but all that XSL was quite enough for me today.
On Tue, 13 Sept 2022 at 15:50, subversion--- via Ohrrpgce <
ohrrpgce@lists.motherhamster.org> wrote:
> teeemcee
> 2022-09-12 20:50:09 -0700 (Mon, 12 Sep 2022)
At least it's a convention, used consistently by all of two commands (I
think) :)
Making up new jargon seems to be unavoidable.
On Sat, 10 Sept 2022 at 12:22, James Paige via Ohrrpgce <
ohrrpgce@lists.motherhamster.org> wrote:
> This might also be my new favorite terrible naming convention :D
>
That sounds problematic for grid and layout slices: although you could
adjust the child's x/y so that it doesn't move, all the other children
would, and if you similarly adjust all of their positions so that they
don't move either, well, the result is quite absurd. It would be a far
better idea to
I'm aware of one thing that causes spawned programs to freeze (which then
freezes Custom under Unix if it's waiting): if the program prints to stdout/err
and it fills up the pipe buffer.
Try attaching gdb or seeing what the last thing in c_debug.txt is (since hspeak
is invoked twice)
--
Hah, that's amusingly dumb. That's a decent solution, but also a timeout would
be reasonable; `waitforkeyrelease` was only intended to wait a couple ticks
anyway.
`waitforkeyrelease` waits for a joystick analog stick exactly when it would
register as a keypress in-game; it calls
And don't be so afraid of using gdb! I don't use a debugger enough either. Used
to be more in the habit years ago...
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1245#issuecomment-1242939176
You are receiving this because you are subscribed
Never thought of it.
I suggest calling it "lookup ancestor" instead. Although there aren't any
existing commands named using the terms "ancestor" or "descendant",
documentation for plenty of commands does, so people should know what they
mean and I think we should just use them more often (plenty
I think this calls for a change to the NPC editor to indicate the vehicle's
speed when mounted.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1247#issuecomment-1263059479
You are receiving this because you are subscribed to this thread.
I've heard of two or three more people complaining of this bug recently. It's a
problem not just because it causes stuck battles, but also it's bad for
on-death effects, e.g. to transmog an enemy into a dead version.
The fundamental problem is the timing of when poison/regen happens in a
I tested in test.rpg, and found that the mounted vehicle speed is what's set as
'Speed' in the Vehicle editor, not what the NPC's speed is. I had completely
forgotten that there was such a setting, and thought the NPC speed would be
used. Maybe you did too?
--
Reply to this email directly or
Sounds like a bug worth mentioning in whatsnew!
On Mon, 10 Oct 2022 at 09:06, subversion--- via Ohrrpgce <
ohrrpgce@lists.motherhamster.org> wrote:
> james
> 2022-10-09 13:05:54 -0700 (Sun, 09 Oct 2022)
> 293
> Fix a bug that allowed NPC activation when riding a vehicle if you clicked
> other
.
On Mon, 10 Oct 2022 at 12:12, James Paige wrote:
> I'm not certain, but I think this bug was introduced after hrodvitnir.
>
> On Sun, Oct 9, 2022, 6:46 PM Ralph Versteegen via Ohrrpgce <
> ohrrpgce@lists.motherhamster.org> wrote:
>
>> Sounds like a bug worth mentionin
I've set up fixed symlinks as described in ohrstable.sh and also created a
temp ohrrpgce-player-win-win95.zip symlink to
ohrrpgce-player-win-2020-05-02-gorgonzola.zip.
http://hamsterrepublic.com/dl/ohrrpgce-player-win.zip was symlinked to
ohrrpgce-minimal-2020-05-02-gorgonzola.zip instead of to
Ah ha. I tried out old versions of dash through 2019 and earlier from
backup disk snapshots and can reproduce the problem, and see that my fix
will work.
On Sun, 7 Aug 2022 at 03:30, James Paige via Ohrrpgce <
ohrrpgce@lists.motherhamster.org> wrote:
> I just checked. The nightly build boxes are
> all the
> [formats](https://github.com/libxmp/libxmp/blob/master/docs/formats.txt)
> supported by libxmp (I don't know if SDL_mixer allows playing them though).
Looking at SDL_mixer's source, it's a bit convoluted... but basically it first
looks at the file extension to figure out which
Paige wrote:
> For my computer it is stuck down at all times. I spent about 90 seconds
> searching my keyboard for a NumLock key before I gave up and changed the
> code
>
> Yeah, I think skipping caps lock and scrollock too is probably a good idea
>
> On Sat, Aug 6, 2022, 11:1
I'm surprised to discover that the latest SDL_mixer 2.6 release adds support
for stb_vorbis, dr_mp3 and dr_flac and makes them the default (statically
linked). These are very popular libraries because they're much smaller than the
standard ones, so I'm very pleased with that, especially because
pent about 90 seconds
>>> searching my keyboard for a NumLock key before I gave up and changed the
>>> code
>>>
>>> Yeah, I think skipping caps lock and scrollock too is probably a good
>>> idea
>>>
>>> On Sat, Aug 6, 2022, 11:13 PM R
I decided to try taking libraries from steam-runtime (v1), which is based on
Ubuntu 12.04 (it relies on some of the host's libraries, so using an old Ubuntu
as base allows running on old linux distros up to 10 years old;
linux_portability_check.py agrees) with libraries such as SDL that are
There are four different problems mentioned in this old bug report:
* (solved) Symlinks in OHRRPGCE-Game.app weren't preserved when distributing to
Mac from Windows. We worked around this problem by just deleting all the
symlinks in OHRRPGCE-Game.app in `bundle-apps.sh`. Ten years ago we also
If we wished, it would now be a small step to switch from svn to git on the
nightly build machines. svn is still exclusively used only for checking for
updates, updating the repo and distrib script, and reverting plotdict.xml.
Note that the linux nightly build machines will continue to use two
Although I've just finished building portable libmodplug and libSDL2_mixer for
linux, I suppose if we switch to libxmp now it'll still save time later.
I forgot that we had some additional test module music in testfiles/. And my
shell history tells me I previously listened to these with the
found any way
> to turn it off. I also have neither ScrollLock nor Pause.
>
> On Sun, Aug 7, 2022 at 8:37 PM Ralph Versteegen via Ohrrpgce <
> ohrrpgce@lists.motherhamster.org> wrote:
>
>> You were searching your keyboard but more likely you should have been
>> searching
After going to some trouble to get a working SDL_mixer 1.2 for linux x86 I
realised that if we're shipping our own libraries there's no real reason not to
use SDL2 instead for x86 too (currently we use SDL_mixer 1.2 on the dubious
theory that an obsolete 32-bit Linux is less likely to have SDL2
Also I appealed for testing on Windows at
https://www.slimesalad.com/forum/viewtopic.php?t=8342
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1202#issuecomment-1208875342
You are receiving this because you are subscribed to this thread.
Stuck down even if you press numlock to turn it off? In what situations?
Is it appropriate to ignore capslock and scrolllock too? All three are
ignored pretty much everywhere else that we scan for a pressed key.
On Sun, 7 Aug 2022 at 07:45, subversion--- via Ohrrpgce <
Closed #1237 as completed.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1237#event-7972959009
You are receiving this because you are subscribed to this thread.
Message ID: ___
Ohrrpgce mailing list
Fixed in r13151 (3a36b2d9) using a Layout slice as I suggested. Honestly it was
pretty difficult to set up the Layout slice right to replicate the old
behaviour, it's a messy solution.
Shifting the text up a pixel was and is done independently for each hero. If a
hero doesn't have MP the text
Topken reports:
> I am unable to install android nightlies as I get this error
> `INSTALL_PARSE_FAILED_NO_CERTIFICATES: Scanning Failed.: No signature found
> in package of version 2 or newer for package
> com.hamsterrepublic.ohrrpgce.game` so not sure how to proceed on my Samsung
> Galaxy Tab
This is a nice feature.
But I notice that it spawns an enemy for each hit on each target, e.g. 2
hits each on 3 targets spawns 6 enemies. Enemy's "spawn on [non]elemental
hit" settings act the same way, but each each enemy has independent
settings for spawn it makes sense for them to all be
Regarding testing whether an .apk is correctly signed without needing an
Android 11 installation, the link I gave suggested using apksigner (included in
the Android SDK) to check:
```
> /opt/android-sdk-linux/build-tools/28.0.3/apksigner verify --verbose
>
I see that error happened while running ohrpackage on one of the Linux VMs.
When you invoke scons, it makes sure the python path is set up to include
SCons. I think SCons should normally also be visible in the oridinary
python environment. Maybe it's because it's running scons with python 2 and
On my machine I'm able to `import SCons` in python3 (but not in python2,
although I also have scons installed for python2). I don't know what the
difference is.
The workaround you added is a correct fix, except the comment isn't.
Transpiling will still work because the import should succeed
Closed #1250 as completed.
--
Reply to this email directly or view it on GitHub:
https://github.com/ohrrpgce/ohrrpgce/issues/1250#event-8726675780
You are receiving this because you are subscribed to this thread.
Message ID: ___
Ohrrpgce mailing list
Actually, the clipping in 8-bit mode is redundant to clipping done in
map_rgb_to_masterpal(), so that 2% is just noise.
On Sun, 26 Mar 2023 at 13:51, subversion--- via Ohrrpgce <
ohrrpgce@lists.motherhamster.org> wrote:
> teeemcee
> 2023-03-25 17:51:36 -0700 (Sat, 25 Mar 2023)
> 147
> Speed up
I'm confused. Yes, that log is definitely from a Windows build under
Proton. But wasn't the idea to get the Linux build running on the Deck?
Didn't you report that it didn't run? What build is it using now after
switching to Desktop mode and back?
On Thu, 20 Apr 2023 at 01:10, Adam Perry wrote:
Thanks. But that's a Windows build of the OHR running under Proton, not the
Linux one we're trying to get working (since the Windows one suffers from
battle slowdowns). Maybe switching to Desktop mode for some reason caused
Steam to switch to the Windows build? I've tried searching the web but
Alright so now that it's working... it's running the windows build, which
Steam has under its own volition decided to download instead of the linux
one? And you don't know whether it had a linux or windows build at the time
that it wasn't launching?
On Fri, 21 Apr 2023 at 01:09, Adam Perry
Oh, good find.
There are commandline flags we can run games with to get debug logs either
from Proton or from the Steam Runtime. Adam, I assume you suffer the battle
slowdown and memory leak bug when running games under Proton? If so, while
you have a Windows build, maybe you could run with
We just had a report about the same slow battles and crashing bug from
someone running (the False Skies demo) under Proton, but for the first time
on a Linux desktop, so it's not Steam Deck specific afterall. But linux
native builds (maybe?) not running on the Deck might still be.
On Sat, 22 Apr
If we want to force Steam to use a certain build (e.g. the Linux one) for
testing, one way to do it is to make it a private (passworded) beta. (I
don't know any other way.)
On Sat, 22 Apr 2023 at 02:02, Adam Perry wrote:
> Hmm... yep, sounds about right.
>
> On Fri, Apr 21, 2023 at 2:33 AM
Update: they reported that the problem doesn't happen with a Linux native
build of the OHR. Also, they're using Manjaro, which is the same Linux
distribution as the Steam Deck, so I think it must be a Manjaro/Arch +
Proton specific problem.
On Sat, 22 Apr 2023 at 22:52, Ralph Versteegen wrote:
Kiefer sent me these icons 3 years ago, but I didn't use them because I
wanted to first add markup code to embed icons into text. I still want to
do that, and add a new Icon sprite type (though I'd love to allow embedding
slices too! Text slices could position their children, and the text wraps
1 - 100 of 141 matches
Mail list logo