Myracle GUI

[mar]

the UI scale is clamped to 100..125% and it's on top of OS settings. The dialogs are designed to fit with 125% extra scale, so you cannot go
higher.

I already query DisplayWidthMM() using X11 API - so it seems it doesn't honor system scaling...?!
just tried on my 4k laptop and it's too small indeed...

there seems to be this this weird abomination, sigh:
https://stackoverflow.com/questions/490 ... tor-in-x11

X11 is one of the worst APIs out there, poorly designed and poorly documented (and Wayland seems to be built on top of sockets, so they made it even worse it seems)
some genius probably thought an async client-server arch for low level window management was a great idea...

as for high usage, I'll look. note that Myracle uses 100% CPU rendering, so that GPU usage is probably blitting

I'll try to report back once I find some time to look into the problems...

as for graph tabs: there needs to be a graph first (a game with engine output), otherwise the windows show nothing

[mar]

although the Linux version is marked as experimental, please let me report a few problems.
I'm using Ubuntu 26.04 on a 14" laptop with intel core ultra 258V cpu. Screen resolution is 2880x1800 and scaling is set to 200%. Running myracle 1.0 build 96.

- the fonts are to tiny, even when setting "UI scale" to 125% in Tools->Settings->Appearance. Manual changing "ui_scale" to 2.00 (in config/appearance.json) has no effect. Could you please consider the os screen scaling factor?

- after starting myracle (without running anything) the cpu load rises a lot here. Xwayland uses up to 95% and myracle up to 45% of the cpu. The gpu is also heavily utilized at ~90%. These loads depend on myracles window size!

ok, uploaded fixed linux builds, they still read as 97 so you may need to redownload (only the minimal build should do, replacing the binary only)

I used the weird way of reading dpi from xresources and highdpi scaling works now on my laptop
I've also changed the way I do blitting (now changed rects only and only of needed) - blit resource usage should be significantly
lower now, especially when idle

I still consider it experimental, however. I'm having a weird problem that sometimes a window fails to create and the app segfaults (typically at startup) - I have no idea what's going on, but it's some x11-related problem I was unable to resolve (yet?)

[Max]

Thanks Martin,
you fixed both bugs lightning fast! I really like your design aesthetic and the small footprint of the gui.

I still consider it experimental, however. I'm having a weird problem that sometimes a window fails to create and the app segfaults (typically at startup) - I have no idea what's going on, but it's some x11-related problem I was unable to resolve (yet?)

I run two short engine matches successively. In the second game of the first match (Stockfish 18 - Pawnoccio 2.0) myracle didn't react to mouse actions any more (like switching graph from "Eval" to "Depth"). But myracle finished the tournament. The second match (Crafty UCI - Clockwork) went without issues.

Max

[mar]

Thanks Martin,
you fixed both bugs lightning fast! I really like your design aesthetic and the small footprint of the gui.

thanks, the price for small footprint is having to dig through low level stuff myself, however.

I run two short engine matches successively. In the second game of the first match (Stockfish 18 - Pawnoccio 2.0) myracle didn't react to mouse actions any more (like switching graph from "Eval" to "Depth"). But myracle finished the tournament.

hmm, that's something new then...

I have another small bug report related to variations, it seems like a neverending story
hopefully this will settle at some point and then I could perhaps focus on the Linux build some more to make it a first class citizen

[mar]

build 98 is up:
- fix: game cut after user move didn't work well with variations
- can select of two primary sorting modes for crosstable in tournament settings (winrate vs points)