Handle missing turbojpeg DLL on Windows#53
Closed
Dj-Shortcut wants to merge 6 commits intoAlanRockefeller:mainfrom
Closed
Handle missing turbojpeg DLL on Windows#53Dj-Shortcut wants to merge 6 commits intoAlanRockefeller:mainfrom
Dj-Shortcut wants to merge 6 commits intoAlanRockefeller:mainfrom
Conversation
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
WalkthroughCentralizes TurboJPEG discovery/initialization in a new imaging.turbo module, updates imaging and thumbnail code to use it (falling back to Pillow), refactors app-data directory selection and guarded file-logging setup, adds Windows README notes, test coverage, and pytest config/gitignore updates. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant Consumer as Consumer (jpeg / prefetcher)
participant TurboMod as faststack.imaging.turbo
participant Env as OS/Env
participant FS as Filesystem
participant Native as native turbojpeg.dll
participant Pillow as Pillow decoder
Consumer->>TurboMod: call create_turbojpeg()
TurboMod->>Env: read env vars (FASTSTACK_TURBOJPEG_LIB/ROOT, PATH, ProgramFiles...)
TurboMod->>FS: build & probe candidate paths
TurboMod->>Native: attempt TurboJPEG(candidate) init
alt init succeeds
Native-->>TurboMod: decoder instance
TurboMod-->>Consumer: return (decoder, True)
Consumer->>Native: perform fast decode/resize
else all attempts fail or turbojpeg missing
TurboMod-->>Consumer: return (None, False)
Consumer->>Pillow: fallback decode/thumbnail via Pillow
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
faststack/logging_setup.py (1)
9-19: Side effect: directories are created during writability probe.The
_is_writable_dirfunction creates directories (viamkdir(parents=True)) for every candidate it checks, not just the one ultimately selected. This leaves behind empty directories in locations like~/.faststackorcwd/var/appdataeven when a higher-priority candidate (e.g.,APPDATA) is ultimately used.Consider checking
candidate.exists()andos.access(candidate, os.W_OK)first, and only create + probe if the directory doesn't exist or the quick check passes.♻️ Suggested improvement to reduce side effects
def _is_writable_dir(path: Path) -> bool: """Return True when the directory can be created and written to.""" try: + if path.exists(): + # Quick check before probing + if not path.is_dir(): + return False path.mkdir(parents=True, exist_ok=True) probe = path / ".write_test" with probe.open("w", encoding="utf-8") as f: f.write("ok") probe.unlink(missing_ok=True) return True except OSError: return False🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/logging_setup.py` around lines 9 - 19, The helper _is_writable_dir currently always creates the directory when probing, causing unwanted side-effect directories; change it to first check if the Path exists and is writable using Path.exists() and os.access(path, os.W_OK) (or path.stat + permission checks) and ONLY if that quick check fails perform the safe create-and-probe sequence (mkdir + write a temp file and cleanup). Keep the function signature and return semantics the same and ensure you only call path.mkdir(parents=True, exist_ok=True) when the directory does not exist or the access check indicates writability cannot be determined.faststack/imaging/turbo.py (1)
76-77: Consider logging a warning when all discovery attempts fail.When
create_turbojpeg()returns(None, False)after trying all candidates, only debug-level logs are emitted. Since this affects performance (Pillow fallback is slower), an info or warning log summarizing the failure would help users troubleshoot without enabling debug logging.💡 Suggested addition before the final return
for candidate in _candidate_library_paths(): if not candidate.exists(): continue try: decoder = TurboJPEG(lib_path=str(candidate)) log.info("Using libjpeg-turbo at %s", candidate) return decoder, True except Exception: log.debug("Failed to initialize turbojpeg from %s", candidate) + log.info( + "TurboJPEG native library not found; JPEG decoding will use Pillow (slower). " + "See README for installation instructions." + ) return None, False🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/imaging/turbo.py` around lines 76 - 77, The function create_turbojpeg currently returns (None, False) silently after trying all candidates; add a higher-severity log (logger.warning or logging.warning) just before that final return to summarize that all turbojpeg discovery attempts failed, include which candidates were tried and that Pillow fallback will be used, and reference the create_turbojpeg function and the candidates variable to locate where to insert the warning.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@faststack/logging_setup.py`:
- Around line 41-47: get_app_data_dir() currently returns the first writable
candidate or falls back to candidates[0] even if unwritable which can change
across runs; update get_app_data_dir() to first scan all candidates (using the
same _is_writable_dir check) for an existing faststack.ini (referencing
AppConfig and the filename faststack.ini) and return that candidate if found,
and if none exists then either return the first writable candidate as before or,
when falling back to an unwritable candidates[0], emit a warning via the logger
indicating the chosen fallback differs from higher-priority candidates so
callers know the path may be unstable.
---
Nitpick comments:
In `@faststack/imaging/turbo.py`:
- Around line 76-77: The function create_turbojpeg currently returns (None,
False) silently after trying all candidates; add a higher-severity log
(logger.warning or logging.warning) just before that final return to summarize
that all turbojpeg discovery attempts failed, include which candidates were
tried and that Pillow fallback will be used, and reference the create_turbojpeg
function and the candidates variable to locate where to insert the warning.
In `@faststack/logging_setup.py`:
- Around line 9-19: The helper _is_writable_dir currently always creates the
directory when probing, causing unwanted side-effect directories; change it to
first check if the Path exists and is writable using Path.exists() and
os.access(path, os.W_OK) (or path.stat + permission checks) and ONLY if that
quick check fails perform the safe create-and-probe sequence (mkdir + write a
temp file and cleanup). Keep the function signature and return semantics the
same and ensure you only call path.mkdir(parents=True, exist_ok=True) when the
directory does not exist or the access check indicates writability cannot be
determined.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 0a44eb70-366c-460c-916a-48a72e3d784f
📒 Files selected for processing (5)
README.mdfaststack/imaging/jpeg.pyfaststack/imaging/turbo.pyfaststack/logging_setup.pyfaststack/thumbnail_view/prefetcher.py
Dj-Shortcut
force-pushed
the
mijn-volgende-branch
branch
from
8b35bc8 to
8d583f3
Compare
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
faststack/thumbnail_view/model.py (1)
612-622:⚠️ Potential issue | 🟠 MajorSame key format concern applies here.
This
metadata_map.get(normalize_path_key(img.path), {})lookup has the same potential mismatch with sidecar-generated keys as noted above.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/thumbnail_view/model.py` around lines 612 - 622, The metadata_map lookup uses normalize_path_key(img.path) which may not match keys produced by the sidecar generator; update the lookup in the block around metadata_map/get so it uses the same key format as the sidecar (or try both forms). Specifically, in the code that does metadata_map.get(normalize_path_key(img.path), {}), replace or augment that call to use the canonical sidecar key (the same normalizer used when writing sidecars) or attempt alternate keys (e.g., raw img.path, the existing normalize_path_key, and the sidecar-normalized key) before falling back to {} so metadata_map and self._get_metadata(img.path) resolve consistently. Ensure you reference normalize_path_key, metadata_map, img.path and _get_metadata when making the change.
🧹 Nitpick comments (2)
faststack/app.py (1)
2125-2155: De-duplicate the default metadata payload to prevent drift.Line 2130-2137 and Line 2148-2155 duplicate the same fallback dict. Keeping one source avoids silent divergence when fields evolve.
♻️ Suggested cleanup
def _get_metadata_dict(self, image_path: Path | str) -> dict: """Get metadata for an image path as a dict for thumbnail model.""" + default_meta = { + "stacked": False, + "uploaded": False, + "edited": False, + "restacked": False, + "favorite": False, + "todo": False, + } try: meta = self.sidecar.get_metadata(image_path, create=False) if meta is None: - return { - "stacked": False, - "uploaded": False, - "edited": False, - "restacked": False, - "favorite": False, - "todo": False, - } + return default_meta.copy() return { "stacked": meta.stacked, "uploaded": meta.uploaded, "edited": meta.edited, "restacked": meta.restacked, "favorite": meta.favorite, "todo": meta.todo, } except Exception as e: # Broad catch for UI plumbing - don't crash grid view log.debug("Failed to get metadata for %s: %s", image_path, e) - return { - "stacked": False, - "uploaded": False, - "edited": False, - "restacked": False, - "favorite": False, - "todo": False, - } + return default_meta.copy()🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/app.py` around lines 2125 - 2155, The _get_metadata_dict function duplicates the same default fallback dict in two places; extract a single canonical default (e.g., DEFAULT_METADATA or a helper function like _default_metadata_dict) and use it for both the None and exception return paths to avoid drift, and when returning it ensure you return a fresh dict (copy) so callers can't mutate the shared constant; update references in _get_metadata_dict and any related code that constructs the metadata using sidecar.get_metadata accordingly.faststack/io/sidecar.py (1)
164-173: Avoid the full-entry scan on every miss.When neither the stable key nor the candidate keys hit, this fallback walks the entire sidecar and normalizes every stored key. On large folders, a batch of new images turns
get_metadata()misses into quadratic work. Consider building a one-time alias map duringload()or gating this scan behind a “legacy keys present” flag so steady-state lookups stay O(1).🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/io/sidecar.py` around lines 164 - 173, The code in get_metadata (the block using self.data.entries and _stable_key_from_key) does a full scan on misses; instead, add a one-time alias map or legacy-keys flag during load() to make lookups O(1): populate self._alias_map (mapping stable_key -> existing_key/meta) or set self._has_legacy_keys by iterating self.data.entries once in load(), and then change get_metadata to consult self._alias_map or short-circuit when self._has_legacy_keys is false; only if the alias_map misses (or the flag indicates legacy keys) perform a targeted update/scan and update alias_map accordingly. Ensure you update any code that currently mutates self.data.entries to also update self._alias_map and the flag so steady-state lookups avoid the full-entry scan.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@faststack/app.py`:
- Line 2230: The INFO logs currently emit full absolute file paths (image_path),
which can leak user directory info; change those log.info calls to log the
filename only by using os.path.basename(image_path) (e.g., replace
log.info("Toggled uploaded flag to %s for %s", meta.uploaded, image_path) with
log.info("Toggled uploaded flag to %s for %s", meta.uploaded,
os.path.basename(image_path))) and add a debug-level log that retains the full
path (e.g., log.debug("Full image path: %s", image_path)) so the detailed path
remains available for debugging; apply the same change to the other similar
occurrences that use image_path in INFO logs (the other log.info lines noted in
the comment).
In `@faststack/io/sidecar.py`:
- Around line 196-213: The code incorrectly treats strings with dots as paths;
in _lookup_keys and _stable_key_from_key remove the check that uses "." in
Path(value).name and instead only treat a value as path-like when it contains
path separators (os.path.sep, "/" or "\\") or when Path(value) resolves to an
existing file under self.directory; when path-like, call
metadata_key_for_path(path) and _metadata_filename_key(path) and use path.stem
as before, otherwise treat the value as a semantic key (return value and
[value]); ensure you add an existence check (Path(self.directory) /
value).exists() or similar to decide path inference so keys like "sample.v1" are
preserved.
---
Outside diff comments:
In `@faststack/thumbnail_view/model.py`:
- Around line 612-622: The metadata_map lookup uses normalize_path_key(img.path)
which may not match keys produced by the sidecar generator; update the lookup in
the block around metadata_map/get so it uses the same key format as the sidecar
(or try both forms). Specifically, in the code that does
metadata_map.get(normalize_path_key(img.path), {}), replace or augment that call
to use the canonical sidecar key (the same normalizer used when writing
sidecars) or attempt alternate keys (e.g., raw img.path, the existing
normalize_path_key, and the sidecar-normalized key) before falling back to {} so
metadata_map and self._get_metadata(img.path) resolve consistently. Ensure you
reference normalize_path_key, metadata_map, img.path and _get_metadata when
making the change.
---
Nitpick comments:
In `@faststack/app.py`:
- Around line 2125-2155: The _get_metadata_dict function duplicates the same
default fallback dict in two places; extract a single canonical default (e.g.,
DEFAULT_METADATA or a helper function like _default_metadata_dict) and use it
for both the None and exception return paths to avoid drift, and when returning
it ensure you return a fresh dict (copy) so callers can't mutate the shared
constant; update references in _get_metadata_dict and any related code that
constructs the metadata using sidecar.get_metadata accordingly.
In `@faststack/io/sidecar.py`:
- Around line 164-173: The code in get_metadata (the block using
self.data.entries and _stable_key_from_key) does a full scan on misses; instead,
add a one-time alias map or legacy-keys flag during load() to make lookups O(1):
populate self._alias_map (mapping stable_key -> existing_key/meta) or set
self._has_legacy_keys by iterating self.data.entries once in load(), and then
change get_metadata to consult self._alias_map or short-circuit when
self._has_legacy_keys is false; only if the alias_map misses (or the flag
indicates legacy keys) perform a targeted update/scan and update alias_map
accordingly. Ensure you update any code that currently mutates self.data.entries
to also update self._alias_map and the flag so steady-state lookups avoid the
full-entry scan.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: cfe59918-0f33-4162-9382-49b1d0e3ff17
📒 Files selected for processing (9)
8faststack/app.pyfaststack/io/sidecar.pyfaststack/logging_setup.pyfaststack/tests/test_jump_to_last_uploaded.pyfaststack/tests/test_sidecar.pyfaststack/tests/thumbnail_view/test_folder_stats.pyfaststack/thumbnail_view/folder_stats.pyfaststack/thumbnail_view/model.py
| status = "uploaded" if meta.uploaded else "not uploaded" | ||
| self.update_status_message(f"Marked as {status}") | ||
| log.info("Toggled uploaded flag to %s for %s", meta.uploaded, stem) | ||
| log.info("Toggled uploaded flag to %s for %s", meta.uploaded, image_path) |
There was a problem hiding this comment.
Avoid full absolute paths in INFO logs.
These log lines now emit full file paths, which can leak user-identifying directory details (e.g., usernames/home paths). Prefer filename-only at INFO and keep full path for DEBUG when needed.
🔐 Safer logging patch
- log.info("Toggled uploaded flag to %s for %s", meta.uploaded, image_path)
+ log.info("Toggled uploaded flag to %s for %s", meta.uploaded, image_path.name)
- log.info("Toggled todo flag to %s for %s", meta.todo, image_path)
+ log.info("Toggled todo flag to %s for %s", meta.todo, image_path.name)
- log.info("Toggled edited flag to %s for %s", meta.edited, image_path)
+ log.info("Toggled edited flag to %s for %s", meta.edited, image_path.name)
- log.info("Toggled restacked flag to %s for %s", meta.restacked, image_path)
+ log.info("Toggled restacked flag to %s for %s", meta.restacked, image_path.name)
- log.info("Toggled favorite flag to %s for %s", meta.favorite, image_path)
+ log.info("Toggled favorite flag to %s for %s", meta.favorite, image_path.name)
- log.info("Toggled stacked flag to %s for %s", meta.stacked, image_path)
+ log.info("Toggled stacked flag to %s for %s", meta.stacked, image_path.name)Also applies to: 2253-2253, 2276-2276, 2299-2299, 2317-2317, 2340-2340
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@faststack/app.py` at line 2230, The INFO logs currently emit full absolute
file paths (image_path), which can leak user directory info; change those
log.info calls to log the filename only by using os.path.basename(image_path)
(e.g., replace log.info("Toggled uploaded flag to %s for %s", meta.uploaded,
image_path) with log.info("Toggled uploaded flag to %s for %s", meta.uploaded,
os.path.basename(image_path))) and add a debug-level log that retains the full
path (e.g., log.debug("Full image path: %s", image_path)) so the detailed path
remains available for debugging; apply the same change to the other similar
occurrences that use image_path in INFO logs (the other log.info lines noted in
the comment).
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@faststack/imaging/jpeg.py`:
- Around line 13-18: The assignment from create_turbojpeg() to jpeg_decoder and
TURBO_AVAILABLE is being overwritten by the legacy initialization block that
sets JPEG_DECODER and TURBO_AVAILABLE later, making jpeg_decoder unused; update
the module to consistently use the discovery result by either (a) replacing the
legacy initialization block with a single assignment that sets uppercase
module-level names (set JPEG_DECODER = jpeg_decoder and TURBO_AVAILABLE =
TURBO_AVAILABLE returned by create_turbojpeg()), or (b) update all decode call
sites to reference the lowercase jpeg_decoder and a single TURBO_AVAILABLE
variable; ensure only one initialization path remains so create_turbojpeg()'s
detection controls availability on Windows and all decoder references
(JPEG_DECODER or jpeg_decoder) are consistently used.
In `@faststack/logging_setup.py`:
- Around line 45-47: get_app_data_dir() currently may return an unwritable path
which causes setup_logging() to attempt unguarded directory/file creation and
crash; modify get_app_data_dir() to verify writability (e.g., attempt to create
and remove a temp file or ensure os.access with write and try mkdir with
exist_ok) for the chosen candidate, and if none are writable fall back to a
known writable location such as a subdirectory under tempfile.gettempdir() (or
create a dedicated writable fallback), then return that verified path; also
update setup_logging() to handle exceptions from directory/file creation
gracefully (catch OSError when creating log dirs/files and fallback to stderr or
the writable temp fallback) so startup cannot fail on permission errors.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 0adabc91-bc80-417e-851a-b47cd70f13bf
📒 Files selected for processing (5)
README.mdfaststack/imaging/jpeg.pyfaststack/imaging/turbo.pyfaststack/logging_setup.pyfaststack/thumbnail_view/prefetcher.py
🚧 Files skipped from review as they are similar to previous changes (1)
- README.md
Author
|
Addressed the merge conflict and pushed the updated branch. The remaining sidecar key/path work is being handled |
There was a problem hiding this comment.
🧹 Nitpick comments (1)
faststack/tests/test_turbo.py (1)
108-126: Consider saving and restoring original handlers for better test isolation.The test correctly validates graceful degradation when file logging fails. However, the cleanup in
finallyclears all handlers rather than restoring the original state, which could affect subsequent tests if any had pre-existing handlers.♻️ Optional: Save and restore handlers
logging_setup.setup_logging(debug=True) root_logger = logging.getLogger() + original_handlers = root_logger.handlers.copy() try: assert any( isinstance(handler, logging.StreamHandler) for handler in root_logger.handlers ) finally: - root_logger.handlers.clear() + root_logger.handlers[:] = original_handlersNote: If the root logger starts empty in your test environment, the current approach is fine.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/tests/test_turbo.py` around lines 108 - 126, Update the test_test_setup_logging_keeps_console_handler_when_file_logging_fails to save the current root logger handlers before calling logging_setup.setup_logging and restore them in the finally block instead of clearing all handlers; capture root_logger.handlers (or a copy via list(root_logger.handlers)) prior to setup_logging, run the assertions using logging_setup.setup_logging, and in finally replace root_logger.handlers with the saved list so pre-existing handlers are preserved for other tests.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@faststack/tests/test_turbo.py`:
- Around line 108-126: Update the
test_test_setup_logging_keeps_console_handler_when_file_logging_fails to save
the current root logger handlers before calling logging_setup.setup_logging and
restore them in the finally block instead of clearing all handlers; capture
root_logger.handlers (or a copy via list(root_logger.handlers)) prior to
setup_logging, run the assertions using logging_setup.setup_logging, and in
finally replace root_logger.handlers with the saved list so pre-existing
handlers are preserved for other tests.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 213719a4-8d8a-45b8-9bb4-263d15a9f576
📒 Files selected for processing (4)
faststack/imaging/jpeg.pyfaststack/imaging/turbo.pyfaststack/logging_setup.pyfaststack/tests/test_turbo.py
🚧 Files skipped from review as they are similar to previous changes (2)
- faststack/imaging/jpeg.py
- faststack/logging_setup.py
Author
Fixed in the current branch. I verified the current code first: the test was clearing root_logger.handlers in finally, which could remove pre- |
Author
|
Small note: this PR includes a tiny Windows-specific pytest test-infra hardening commit. On Windows, pytest can fail with PermissionError (WinError 5) due to inaccessible The change:
No production code or test logic was modified. If preferred, I can split this out into a separate PR, but I kept it here since it's |
Author
|
@CodeRabbit review |
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
🧹 Nitpick comments (1)
faststack/imaging/turbo.py (1)
60-63: Filter non-existent PATH DLL candidates before enqueueing.Every PATH directory currently becomes a TurboJPEG load attempt, which can create heavy exception/log noise on Windows fallback paths. Only add candidates when
turbojpeg.dllactually exists.♻️ Proposed change
- for path_dir in os.getenv("PATH", "").split(os.pathsep): - if path_dir: - candidates.append(str(Path(path_dir) / "turbojpeg.dll")) + for path_dir in os.getenv("PATH", "").split(os.pathsep): + if not path_dir: + continue + dll_path = Path(path_dir) / "turbojpeg.dll" + if dll_path.is_file(): + candidates.append(str(dll_path))🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@faststack/imaging/turbo.py` around lines 60 - 63, The PATH scanning loop is appending every PATH directory's turbojpeg.dll to the candidates list even if the file doesn't exist, causing noisy load attempts; update the loop that builds candidates (the for path_dir in os.getenv("PATH"... block) to check Path(path_dir) / "turbojpeg.dll".exists() before calling candidates.append(...) so only existing DLL paths are enqueued for loading.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@faststack/imaging/turbo.py`:
- Around line 60-63: The PATH scanning loop is appending every PATH directory's
turbojpeg.dll to the candidates list even if the file doesn't exist, causing
noisy load attempts; update the loop that builds candidates (the for path_dir in
os.getenv("PATH"... block) to check Path(path_dir) / "turbojpeg.dll".exists()
before calling candidates.append(...) so only existing DLL paths are enqueued
for loading.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 2eb706e3-dfa0-4f32-a660-993ee6bd9001
📒 Files selected for processing (6)
.gitignorefaststack/imaging/jpeg.pyfaststack/imaging/turbo.pyfaststack/logging_setup.pyfaststack/tests/test_turbo.pypytest.ini
✅ Files skipped from review due to trivial changes (3)
- .gitignore
- pytest.ini
- faststack/tests/test_turbo.py
🚧 Files skipped from review as they are similar to previous changes (1)
- faststack/logging_setup.py
Dj-Shortcut
force-pushed
the
mijn-volgende-branch
branch
from
740397b to
fc05bc4
Compare
Owner
|
Accepted changes from this PR (with modifications):
Reject:
The good parts of this PR will be merged with main in PR # 58 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fix Windows startup issues when turbojpeg.dll is missing and document the setup for Windows users.
This PR makes FastStack fall back cleanly when the native libjpeg-turbo library is not installed, instead of failing during startup. It also adds shared
TurboJPEG discovery logic, improves app-data directory fallback handling, and documents the Windows setup and troubleshooting steps in the README.
Summary by CodeRabbit
New Features
Documentation
Improvements
Tests
Chores