Skip to content

Actions: MiSTerOrganize/MiSTer_OpenBOR_7533

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
771 workflow runs
771 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
refresh latest (handler: remove .snapshots entirely -- REPSTATE, prune and fallback The fallback could not fire for any take that can exist. A pre-payload take is container v1 and the reader rejects it at _mr_cv==2u before any of this runs; a v2 take always carries its payload, because container v2 and the payload shipped in the SAME binary with no build between them. So the branch that read the sidecar was unreachable, and the keep-20 prune had nothing left to prune. The .inp is now the whole take o... Refresh latest release #43: Commit ef4bd35 pushed by MiSTerOrganize
14s main
main
14s
handler: remove .snapshots entirely -- REPSTATE, prune and fallback The fallback could not fire for any take that can exist. A pre-payload take is container v1 and the reader rejects it at _mr_cv==2u before any of this runs; a v2 take always carries its payload, because container v2 and the payload shipped in the SAME binary with no build between them. So the branch that read the sidecar was unreachable, and the keep-20 prune had nothing left to prune. The .inp is now the whole take on both cores. T... Build MiSTer ARM #530: Commit ef4bd35 pushed by MiSTerOrganize
21m 22s main
main
21m 22s
DB dispatch: handler: remove .snapshots entirely -- REPSTATE, prune and fallback The fallback could not fire for any take that can exist. A pre-payload take is container v1 and the reader rejects it at _mr_cv==2u before any of this runs; a v2 take always carries its payload, because container v2 and the payload shipped in the SAME binary with no build between them. So the branch that read the sidecar was unreachable, and the keep-20 prune had nothing left to prune. The .inp is now the whole take on b... Dispatch DB rebuild #14: Commit ef4bd35 pushed by MiSTerOrganize
6s main
main
6s
refresh latest (workflow_dispatch) Refresh latest release #42: Manually run by MiSTerOrganize
13s main
main
13s
refresh latest (workflow_dispatch) Refresh latest release #41: Manually run by MiSTerOrganize
18s main
main
18s
openbor-diff-harness Diff Harness #60: Commit 655b161 pushed by MiSTerOrganize
1m 19s main
main
1m 19s
refresh latest (handler: stop creating .snapshots -- nothing writes it now The engine's sidecar write is gone, so creating the directory every launch would leave exactly the empty advertisement of a per-take store that PICO-8's dead one was. Both remaining users cope with its absence: the fallback tests [ -d "$_SNAP" ] and the prune's glob matches nothing. $REPSTATE stays DEFINED -- the fallback and the prune both reference the path to read directories recorded before the payload existed. Co-Authored-... Refresh latest release #40: Commit 655b161 pushed by MiSTerOrganize
14s main
main
14s
handler: stop creating .snapshots -- nothing writes it now The engine's sidecar write is gone, so creating the directory every launch would leave exactly the empty advertisement of a per-take store that PICO-8's dead one was. Both remaining users cope with its absence: the fallback tests [ -d "$_SNAP" ] and the prune's glob matches nothing. $REPSTATE stays DEFINED -- the fallback and the prune both reference the path to read directories recorded before the payload existed. Co-Authored-By: Claude Opus ... Build MiSTer ARM #529: Commit 655b161 pushed by MiSTerOrganize
20m 51s main
main
20m 51s
DB dispatch: handler: stop creating .snapshots -- nothing writes it now The engine's sidecar write is gone, so creating the directory every launch would leave exactly the empty advertisement of a per-take store that PICO-8's dead one was. Both remaining users cope with its absence: the fallback tests [ -d "$_SNAP" ] and the prune's glob matches nothing. $REPSTATE stays DEFINED -- the fallback and the prune both reference the path to read directories recorded before the payload existed. Co-Authored-By:... Dispatch DB rebuild #13: Commit 655b161 pushed by MiSTerOrganize
8s main
main
8s
openbor-diff-harness Diff Harness #59: Commit 532ee9c pushed by MiSTerOrganize
1m 28s main
main
1m 28s
engine: stop writing the .snapshots sidecar; the .inp already carries it Both copies came from MREC_ARMSNAP, in the same if(_mr_ok && rename(...)) success block: the payload embedded in the take, and a byte-identical sidecar dropped in .snapshots/<take>.<seed>/. The .inp is self-sufficient and is the only form that travels between machines, so the directory duplicated the file it sat beside. The HANDLER still reads .snapshots as a fallback, deliberately. A take recorded before the payload existed has... Build MiSTer ARM #528: Commit 532ee9c pushed by MiSTerOrganize
28m 25s main
main
28m 25s
refresh latest (workflow_dispatch) Refresh latest release #39: Manually run by MiSTerOrganize
13s main
main
13s
DB dispatch: handler: log prune reads whole lines, per this file's own rule The snapshot prune two hundred lines above states it: xargs splits on whitespace, so a path containing a space reaches rm -f as two fragments, the second resolved relative to cd "$GAMEDIR" -- inside the user's PAK library. The log prune still used xargs. Safe today only because $LOGDIR and the generated names happen to contain no spaces, i.e. correct by accident. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Dispatch DB rebuild #12: Commit cf4727b pushed by MiSTerOrganize
9s main
main
9s
openbor-diff-harness Diff Harness #58: Commit cf4727b pushed by MiSTerOrganize
1m 30s main
main
1m 30s
refresh latest (handler: log prune reads whole lines, per this file's own rule The snapshot prune two hundred lines above states it: xargs splits on whitespace, so a path containing a space reaches rm -f as two fragments, the second resolved relative to cd "$GAMEDIR" -- inside the user's PAK library. The log prune still used xargs. Safe today only because $LOGDIR and the generated names happen to contain no spaces, i.e. correct by accident. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.... Refresh latest release #38: Commit cf4727b pushed by MiSTerOrganize
13s main
main
13s
handler: log prune reads whole lines, per this file's own rule The snapshot prune two hundred lines above states it: xargs splits on whitespace, so a path containing a space reaches rm -f as two fragments, the second resolved relative to cd "$GAMEDIR" -- inside the user's PAK library. The log prune still used xargs. Safe today only because $LOGDIR and the generated names happen to contain no spaces, i.e. correct by accident. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Build MiSTer ARM #527: Commit cf4727b pushed by MiSTerOrganize
26m 9s main
main
26m 9s
refresh latest (workflow_dispatch) Refresh latest release #37: Manually run by MiSTerOrganize
16s main
main
16s
refresh latest (workflow_dispatch) Refresh latest release #36: Manually run by MiSTerOrganize
16s main
main
16s
CI: widen the diagnostic gate, and stop taking the engine clone on trust The gate scanned .github/scripts/apply_patches.py but not patch_sdl_dummy.py -- 376 lines that inject the DDR3 mmap, the keepalive thread and the frame writer into SDL's dummy driver, all linked into the shipped binary. That is the same omission that let a 60-line PPM frame-capture diagnostic through in 2026-05-22. Scope is now .github/scripts entire. Verified: no false positive on today's tree, and a synthetic marker placed in pat... Build MiSTer ARM #526: Commit 4812454 pushed by MiSTerOrganize
26m 24s main
main
26m 24s
openbor-diff-harness Diff Harness #57: Commit 4812454 pushed by MiSTerOrganize
1m 12s main
main
1m 12s
openbor-diff-harness Diff Harness #56: Commit d0b6d43 pushed by MiSTerOrganize
1m 14s main
main
1m 14s
refresh latest (handler: the 4086 replays guard was defeated three lines later ee4f780 wrapped only the .snapshots mkdir. The .scratch mkdir below it ran unconditionally and re-created $REPLAYS as a parent, so loading 4086 still produced /media/fat/replays/OpenBOR_4086 on every launch -- same directory, a different hidden child. The guard read as correct and did nothing. Replaced the local if with a single HAS_RECORDER flag tested at every site that creates something under $REPLAYS. The remaining sit... Refresh latest release #35: Commit d0b6d43 pushed by MiSTerOrganize
14s main
main
14s
DB dispatch: handler: the 4086 replays guard was defeated three lines later ee4f780 wrapped only the .snapshots mkdir. The .scratch mkdir below it ran unconditionally and re-created $REPLAYS as a parent, so loading 4086 still produced /media/fat/replays/OpenBOR_4086 on every launch -- same directory, a different hidden child. The guard read as correct and did nothing. Replaced the local if with a single HAS_RECORDER flag tested at every site that creates something under $REPLAYS. The remaining site (... Dispatch DB rebuild #11: Commit d0b6d43 pushed by MiSTerOrganize
8s main
main
8s
handler: the 4086 replays guard was defeated three lines later ee4f780 wrapped only the .snapshots mkdir. The .scratch mkdir below it ran unconditionally and re-created $REPLAYS as a parent, so loading 4086 still produced /media/fat/replays/OpenBOR_4086 on every launch -- same directory, a different hidden child. The guard read as correct and did nothing. Replaced the local if with a single HAS_RECORDER flag tested at every site that creates something under $REPLAYS. The remaining site (.armsnap, in ... Build MiSTer ARM #525: Commit d0b6d43 pushed by MiSTerOrganize
27m 52s main
main
27m 52s
handler: do not create a replays tree for 4086, which has no recorder REPLAYS is derived from $BINARY and was mkdir'd unconditionally, so loading 4086 created an empty /media/fat/replays/OpenBOR_4086/.snapshots on every launch -- advertising a feature that build does not have. 4086 never got the recorder (deliberate divergence; the repo is archived). The comment above that mkdir explains why it is load-bearing: MiSTer only remembers a browse directory while it still exists, so if the tree vanished "Loa... Build MiSTer ARM #524: Commit ee4f780 pushed by MiSTerOrganize
22m 23s main
main
22m 23s