[Coastrnaut]

If a code block of a preset doesn't compile, Milkdrop shows a warning, leaves that block disabled and still uses the preset (RecompileExpressions() in state.cpp). If the init code fails, it resets the q variables (or the t variables of a custom wave or shape) that the init code would have set.

projectM throws a MilkdropCompileException instead, so the whole preset fails to load. With the line joining fix in #1068, 17 presets in the four packs below still fail this way, all with code that Milkdrop rejects as well, for example:

shape_0_per_frame1=additive=above(sin(time*200),0.99l);
y=if(c,.-.4,y);
zoom=zoom+.10*sin(rad+.+15.15);

This logs a warning and skips the block instead of throwing, and resets the variables like Milkdrop does. It intentionally changes how projectM treats broken presets, so I kept it separate from the other fixes: apps that relied on these presets being skipped will now show them, like Milkdrop does.

Results

• The 17 presets load and render, with the broken block skipped and a warning logged.
• For all other presets, shader compile/link results and rendered frames are identical to master.
• New EquationCompileErrorTest. The per-frame context can be tested without a GL context; PresetState needs one, so the init code path is covered by the preset run. All unit tests pass.

How I tested it

I loaded all 14,536 presets from cream-of-the-crop, en-d, milkdrop-original and projectm-classic one after another (Windows, RTX A6000), with the preset folders as texture search path, rendered 30 frames each with fixed audio and frame times, and recorded shader compile/link results, failed loads, logged warnings and frame hashes. For the pixel comparison both builds had their random seeds fixed (a local patch only, identical on both sides). The presets this PR fixes are excluded from the pixel comparison, since loading them changes shared state (e.g. the expression evaluator's random generator) for every preset after them.

Retested on current master (2026-11-12)

• Unit tests: 166/166 libprojectM and 98/98 playlist tests pass.
• Presets: I loaded all 14,536 presets from cream-of-the-crop, en-d, milkdrop-original and projectm-classic one after another with this branch and with master (GL 3.3 core, RTX A6000, preset folders as texture search path, 30 frames each with fixed audio, frame times and random seeds; the seeds are a local patch only, identical on both sides). The 100 presets that failed to load now load; no preset gets worse. One of them, sonar cow, has a composite shader that doesn't compile without the other shader PRs, so it uses the default one here. Frames before the first of them are identical to master except for $$$ Royal - Mashup (257), which also renders differently between two runs of master.
• Together with my other PRs: All my open PRs (Fix calling convention of glGetStringi pointer in GLProbe #1059 to Build for Windows Store (UWP) without app-side workarounds #1065, Return int from the NaN-safe multiplication of two ints #1067 to Fix shaders that declare or name samplers through macros #1072, Build Apple frameworks for iOS, tvOS, watchOS and visionOS, and support EAGL contexts #1075, Don't shift the blur textures by Direct3D's half-pixel offset #1078 and Render with 10 bits per color channel if the target framebuffer has more than 8 #1079) merge with each other and with master without conflicts. Merged together (all except Render with 10 bits per color channel if the target framebuffer has more than 8 #1079, which I opened after this run), they build on desktop GL and GLES, all unit tests pass, and the corpus gives 0 failed loads, 0 shader compile/link failures and 0 GL errors in an OpenGL ES context, with and without textures.