Skip to content

Free, open-source AI video generator for Mac. Run LTX-2.5 and MiniMax H3 (Hailuo) locally on Apple Silicon: text-to-video and image-to-video with sound, lip-sync, character LoRA training, songs from lyrics (YuE2) and image editing (Qwen-Image-Edit). One-click install via Pinokio. Offline: no cloud, no API key, no NVIDIA GPU.

patch_ltx_codec: add upgrade path for old codec patches missing +faststart

mrbizarrod6e6f41c9560
diff --git a/patch_ltx_codec.py b/patch_ltx_codec.py
index 133722d..a8a0f73 100644
--- a/patch_ltx_codec.py
+++ b/patch_ltx_codec.py
@@ -142,18 +142,63 @@ def _atomic_write(target: Path, text: str) -> None:
raise
-def apply_patch(target: Path, old: str, new: str, marker: str, label: str) -> str:
+def apply_patch(target: Path, old: str, new: str, marker: str, label: str,
+ upgrade_marker: str | None = None) -> str:
"""Idempotently apply old→new replacement on `target`. Returns one of
OUTCOME_APPLIED / OUTCOME_ALREADY / OUTCOME_MISSING / OUTCOME_DRIFT —
deep-review fix to surface upstream drift loudly instead of silently
- no-op'ing the patch and shipping a broken install."""
+ no-op'ing the patch and shipping a broken install.
+
+ `upgrade_marker` (optional): a substring that exists in the NEW patch
+ but not in the OLD already-applied version. If `marker` is found but
+ `upgrade_marker` is NOT, an older version of our own patch is on disk
+ — re-write the file to the latest content. Used for shipping fixes to
+ users who already have an earlier patch applied (e.g. adding +faststart
+ to the codec patch without forcing a venv rebuild)."""
if target is None or not target.exists():
print(f" [{label}] MISSING — target file not found", file=sys.stderr)
return OUTCOME_MISSING
text = target.read_text()
if marker in text:
- print(f" [{label}] already patched: {target}")
- return OUTCOME_ALREADY
+ # Marker present → some version of our patch is on disk. If the
+ # caller didn't supply an upgrade_marker we treat it as ALREADY.
+ if upgrade_marker is None or upgrade_marker in text:
+ print(f" [{label}] already patched: {target}")
+ return OUTCOME_ALREADY
+ # Marker but no upgrade_marker → old patch version on disk. The
+ # surrounding lines were rewritten by the previous patch, so the
+ # OLD raw upstream string isn't there to find. Restore from import
+ # of the function's source text by REVERTING our previous patch
+ # back to OLD using an embedded undo, then re-apply NEW. We do
+ # this by re-reading the upstream commit-pinned source.
+ print(f" [{label}] upgrading older patch: {target}")
+ # Naive but works for our patches: re-write the entire file by
+ # mapping NEW back to itself isn't useful. Instead, find the
+ # specific patched line and replace with NEW — provided NEW
+ # contains a unique-enough head/tail anchor. We rely on caller
+ # setting OLD to a substring still present after patching.
+ # SIMPLER ROUTE: write the upstream OLD string anywhere it would
+ # have been. Since we can't easily reconstruct the un-patched
+ # form, we instead just edit the patched site-packages directly
+ # at runtime via direct Edit — this branch is reachable only
+ # for the codec patch where +faststart was added. The narrow
+ # marker-pair (old=cmd.extend without faststart, new=cmd.extend
+ # with faststart) lets us find and replace.
+ # Find the old line (without faststart) and replace with new line.
+ old_one_liner = ('cmd.extend(["-c:v", "libx264", "-pix_fmt", _pix, '
+ '"-crf", _crf, output_path])')
+ new_one_liner = ('cmd.extend(["-c:v", "libx264", "-pix_fmt", _pix, '
+ '"-crf", _crf,\n "-movflags", '
+ '"+faststart", output_path])')
+ if old_one_liner in text:
+ _atomic_write(target, text.replace(old_one_liner, new_one_liner))
+ print(f" [{label}] upgrade applied: {target}")
+ return OUTCOME_APPLIED
+ print(
+ f" [{label}] upgrade target text not found — patch shape may have "
+ f"changed. Manual inspection needed.", file=sys.stderr,
+ )
+ return OUTCOME_DRIFT
if old not in text:
print(
f" [{label}] DRIFT — expected text not found in {target}. "
@@ -173,11 +218,18 @@ def main() -> int:
outcomes: list[tuple[str, str]] = []
# Patch 1: codec
+ # `upgrade_marker="+faststart"` lets us upgrade installs where the
+ # earlier version of this patch was applied (LTX_OUTPUT_PIX_FMT marker
+ # present, but the +faststart movflag missing). Without the upgrade
+ # path, those installs would never get the moov-at-front fix that lets
+ # gallery thumbnails render the first frame without downloading the
+ # full clip.
codec_target = _find("ltx_core_mlx/model/video_vae/video_vae.py")
- outcomes.append(("codec (yuv444p crf 0)", apply_patch(
+ outcomes.append(("codec (yuv444p crf 0 + faststart)", apply_patch(
codec_target, PATCH_CODEC_OLD, PATCH_CODEC_NEW,
marker="LTX_OUTPUT_PIX_FMT",
- label="codec (yuv444p crf 0)",
+ upgrade_marker="+faststart",
+ label="codec (yuv444p crf 0 + faststart)",
)))
# Patch 2: I2V OOM cleanup before decode