Repository navigation
fix(build): a compile error in a source file no longer translates the game again - #2675
Conversation
… game again A failed compile made the next build translate the whole game again, whatever failed: on Windows a one-line typo in host/hos_update.c cost 301 s on the build after it, 276 s of them the lift, against 26 s for a normal host edit. The lift is forgotten after a failure for the cases in issues 1197 and 1176, a lifted file cut short or lifted C another builder left, and those fail in the translated game's own files. make_game.py now takes the files build.py could not compile, alone too, from its "FAILED <name>" lines. When every one is a source of the repository, not a lifted file (code_*) and none the lift or build.py writes (sa/gen, sa/build/rt64, sa/build/gen), forget_failed_build keeps the translation, and the next build compiles only the files not compiled yet. A build stopped without such a line, one with a generated file among them, and a failed link forget it as before. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Upstream NotRexed/Hookshot fixed four builder problems the fork shares, on the path its Flutter builder and Rust engine use. Ported semantically, with the Rust engine's own layout and failure handling: - e7f5581: scripts/oot_saves.py finds Lichii's own data folder (~/.local/share/lichii, %APPDATA%\lichii) before Eden's, where a Lichii before build 150 still keeps its saves. `list` groups programs by the folder import and export use; the fork's `saves` list for the Flutter launcher follows the same rule, so Lichii is not offered from Eden's. - 6c529a2: a Btrfs folder has a device number no disk has, so the hard drive warning could not tell; nx2/disk.py and builder/rust/src/disk.rs now read /proc/self/mountinfo for the deepest mount and every disk of a multi-disk Btrfs (/sys/fs/btrfs/<id>/devices). - c544828 (+ NotRexed#2686, NotRexed#2675): the lift key is taken again after the lift and split_hot, from image/chunks.txt that the first lift writes, so a new work folder's second build no longer translates the game again; the new key is kept only when nothing else the lift reads changed while it ran. A compile that failed only in repository sources keeps the translation (a host typo no longer costs a lift); a lifted or generated file that failed still forgets it. The Rust driver never had a key but had no such distinction: compile failures now follow the same rule, a link failure forgets the lift, and a failed host source keeps it. - a740d67: host_mtune()/host::mtune() detect a Steam Deck APU (0405 or 0932) or RX_STEAMDECK=1 and add -mtune=znver2 beside -march, in nx2/build.py and the Rust driver's game and host compiles; RX_MTUNE names a tune. The fork's lift and host builds both get it. Tests: the upstream ones adapted (test_oot_saves, test_disk_warning, test_cpu_target, test_lift_cache, test_lift_cleanup_error), new tests/test_failed_compile_keeps_lift.py, and Rust tests for the Btrfs mount walk, the lift failure rule and the Zen 2 tune. No game files or generated game data. Assisted-by: opencode (DeepSeek V4.1 Flash 1017)
Joey-Magnus
left a comment
Joey-Magnus
left a comment
There was a problem hiding this comment.
I merged this into current main and ran tests/test_failed_compile_keeps_objects.py (passes) and run_all.py make_game lift failed build compile. The only failures were test_lift_cache (fails on main here too) and two Windows-only checks the name filter forced to run (test_portable_local_builder, test_precompile_compute_guard), which fail the same way on main. CI is green.
The classification is sound. build.py prints FAILED <basename> ( only for files that failed alone too (build.py:1281), and the regex doesn't match the earlier "not compiled yet" line. A failure in code_* or anything in sa/gen, sa/build/rt64 or sa/build/gen still forgets the lift (make_game.py:1675). A link failure still forgets it, since linked is checked first.
The one case I looked for: host sources compile with -I gen, so a cut-short lifted header could make host files fail. But the lifted chunks include the same headers and every file is compiled before build.py gives up, so a code_ name would be in the list and the lift is still forgotten. The 301 s to 21 s numbers come from your own Windows build log, which is the kind of measurement we want.
A one-line typo in a host file makes the next build translate the whole game again. On Windows, the build after a typo in
host/hos_update.ctook 301 s, 276 s of it the lift, against 26 s for a normal host edit.Why the lift was forgotten
After a failed build,
forget_failed_builddrops the lift key for the cases in issues 1197 and 1176: a lifted file cut short, and lifted C left by another builder. Both fail in the translated game's own files. A compile error in a repository source isn't one of them.Change
make_game.pynow takes the filesbuild.pycould not compile, even alone, from itsFAILED <name> (...)lines.forget_failed_buildkeeps the translation, and the next build compiles only the files not compiled yet. A repository source here means no lifted file (code_*) and nothing the lift orbuild.pywrites (sa/gen,sa/build/rt64,sa/build/gen).the build failed in hos_update.c, not in the translated game: the next one keeps the translation and compiles only the files not compiled yet.Testing
tests/test_failed_compile_keeps_objects.py: a repository source that failed keeps the translation. A lifted or generated file among the failed files (code_021.c,code_hot_3.c,hos_image.cinsa/gen,a64_rt.cinrt64) forgets it. TheFAILEDpattern matchesbuild.py's own line, and not its first-try "not compiled yet" line. The new cases fail without the change, and all six pass with it.mainat 0ec31ff plus this change): the typo build failed in 16 s with the new message, and the build after it, with the typo removed, took 21 s and skipped the lift. Without the change, that build took 301 s.run_all.pyon Windows: every check that readsmake_game.py(52) passes, apart fromtest_cpp_preflight,test_linux_updateandtest_prepare_stall, which fail the same way on a cleanmainhere, andtest_lift_cache, which fails here only because of a space in the checkout's path (fixed in the lift-key PR).Part of the dev-loop proposal (#2636, A7). Classifying link failures, the rest of A7, is left for later.
🤖 Generated with Claude Code