[Coastrnaut]

Milkdrop reads the numbered code lines of a preset (per_frame_1=, per_frame_2=, …), removes // and \\ comments up to the end of each line and then joins the lines without any separator before compiling the expressions (ReadCode() and StripLinefeedCharsAndComments() in state.cpp). Its editor wrapped long lines when saving presets, often in the middle of a variable name or number:

per_pixel_1=rot=rot*atan2(-rad,sin(ang*20-ang*20*ang*10*atan2(above(bass,bass_Att),above(Treb,t
per_pixel_2=reb_Att))))*sin(Rad*ang*above(Bass,bass_Att))*rad-ang;

projectM joined the lines with a newline, which split these tokens (t / reb_Att, is_ / beat, …), so projectm-eval rejected the code and the whole preset failed to load. The documentation of GetCode() already said equations are returned as a single line with comments stripped.

This adds PresetFileParser::GetEquationCode(), which joins lines like Milkdrop does, and uses it for all expression code of presets and user sprites. Shader code still uses GetCode() and keeps its line breaks.

Results

• 83 of the 100 presets that failed to load in the four packs below load now. The remaining 17 contain code that Milkdrop can't compile either (e.g. 0.99l, .-.4); Skip equation code that doesn't compile instead of failing the preset #1069 handles those.
• For all 14,436 presets that already loaded, shader compile/link results and all rendered frames are identical to master.
• New PresetFileParser tests, including the wrapped-token cases. 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 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: 170/170 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). 83 presets that failed to load now load; no preset gets worse. All frames before the first of them are identical to master; later ones differ because loading a preset changes shared state such as the expression evaluator's random generator.
• 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.