fix: create httpx.AsyncClient off event loop (Python 3.13 / HA blocking SSL detection)#46
Open
danielbrunt57 wants to merge 19 commits intoalandtse:mainfrom
Conversation
…ng SSL detec - Home Assistant on Python 3.13 flags blocking SSL initialization -(SSLContext.load_verify_locations) when httpx.AsyncClient() is created on the event loop thread. - This change lazily instantiates the session using asyncio.to_thread() and ensures it exists before request handling. - Prevents Detected blocking call to load_verify_locations warnings while preserving existing session_factory behavior.
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 inconclusive)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
📝 Coding Plan
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 Tip CodeRabbit can suggest fixes for GitHub Check annotations.Configure the |
danielbrunt57
changed the title
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
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)
authcaptureproxy/auth_capture_proxy.py (1)
192-204:⚠️ Potential issue | 🟠 MajorReturn type mismatch and dead local variable.
reset_datais typed-> Nonebut line 203 returns aweb.Response. Callers (e.g.,change_host_urlat line 190) discard the return value, so the error response is silently lost.- The
sessionlocal on line 201 is assigned but never used after the None-guard — it's dead code.If the intent is purely defensive logging, raise or log instead of returning a response nobody will see:
Proposed fix
self.session = None await self._ensure_session() - session = self.session - if session is None: # pragma: no cover - return await self._build_response(text="Internal error: HTTP session not initialized") + if self.session is None: # pragma: no cover + raise RuntimeError("Failed to initialize HTTP session")🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/auth_capture_proxy.py` around lines 192 - 204, The reset_data method currently declares -> None but returns a web.Response and assigns a dead local variable `session`; fix it by removing the unused local `session = self.session` assignment and replacing the defensive `if session is None` return with an explicit error handling path (either log the error and raise a RuntimeError or just raise RuntimeError("Internal error: HTTP session not initialized")) so the error is surfaced instead of silently discarded by callers like change_host_url; ensure you keep the method signature `async def reset_data(self) -> None` and use the class logger (or raise) rather than returning a web.Response.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 362-364: In all_handler, after awaiting self._ensure_session(),
assign the created session to a local variable (e.g., session = self.session) so
the subsequent uses of session (the bare name referenced later in all_handler)
refer to the instance session; update the all_handler function to read from
self.session after _ensure_session() and ensure all downstream references (the
places in all_handler that currently use session) use that local variable or
self.session consistently so the NameError is resolved.
- Around line 213-221: The _ensure_session method can race during concurrent
calls and create duplicate clients; add an asyncio.Lock instance (e.g.,
self._session_lock = asyncio.Lock()) in __init__ and use it to guard session
creation in _ensure_session: acquire the lock before calling await
asyncio.to_thread(self.session_factory), re-check self.session after acquiring
to avoid duplicate creation, assign the newly created client to self.session
only if still None, and release the lock; ensure you reference
self._session_lock, self.session_factory, and _ensure_session when implementing
this change.
---
Outside diff comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 192-204: The reset_data method currently declares -> None but
returns a web.Response and assigns a dead local variable `session`; fix it by
removing the unused local `session = self.session` assignment and replacing the
defensive `if session is None` return with an explicit error handling path
(either log the error and raise a RuntimeError or just raise
RuntimeError("Internal error: HTTP session not initialized")) so the error is
surfaced instead of silently discarded by callers like change_host_url; ensure
you keep the method signature `async def reset_data(self) -> None` and use the
class logger (or raise) rather than returning a web.Response.
---
Duplicate comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 554-588: The six uses of session in the request-sending branch are
failing because session was not assigned; after calling _ensure_session() assign
the instance session to the local variable by adding session = self.session so
subsequent calls like getattr(session, method)(...) and other session usages
resolve; locate the function that calls _ensure_session() (the block containing
the multipart/raw/json/data branches) and insert the assignment immediately
after _ensure_session() is invoked.
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
authcaptureproxy/auth_capture_proxy.py (1)
207-218:⚠️ Potential issue | 🟡 Minor
reset_data()return type isNonebut the guard returnsweb.Response.The type annotation
-> Noneconflicts withreturn await self._build_response(...)on line 218, which producesweb.Response. Mypy/pyright will flag this. Additionally,change_host_url(line 205) callsawait self.reset_data()and silently discards the return value, so any caller relying onreset_datacompleting successfully has no signal when the guard is hit.The guard is marked
# pragma: no coverbecause_ensure_session()re-raises on failure (no silentNonepath), making it effectively dead code. Either remove it or fix the type:🛠️ Option A – remove the unreachable guard (preferred if _ensure_session always raises on failure)
- self.session = None - await self._ensure_session() - session = self.session - if session is None: # pragma: no cover - return await self._build_response(text="Internal error: HTTP session not initialized") + self.session = None + await self._ensure_session()🛠️ Option B – fix the return type annotation if the guard is intentionally kept
- async def reset_data(self) -> None: + async def reset_data(self) -> Optional[web.Response]:🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/auth_capture_proxy.py` around lines 207 - 218, reset_data() is annotated to return None but contains a guard that returns await self._build_response(...) (a web.Response), causing a type mismatch; either remove the unreachable guard (preferred if _ensure_session() always raises on failure) by deleting the "if session is None" block so reset_data remains -> None, or explicitly make the function return Optional[web.Response] and update callers (e.g., change_host_url which currently awaits reset_data() and ignores its return) to check/propagate the web.Response; reference symbols: reset_data, _ensure_session, _build_response, change_host_url.
🧹 Nitpick comments (1)
authcaptureproxy/auth_capture_proxy.py (1)
74-79: Redundantin_event_loop = Falsein theexceptblock.
in_event_loopis alreadyFalseat line 74; the re-assignment in theexceptbranch is a no-op.♻️ Proposed cleanup
- in_event_loop = False try: asyncio.get_running_loop() in_event_loop = True except RuntimeError: - in_event_loop = False + passOr more idiomatically:
- in_event_loop = False - try: - asyncio.get_running_loop() - in_event_loop = True - except RuntimeError: - in_event_loop = False + try: + asyncio.get_running_loop() + in_event_loop = True + except RuntimeError: + in_event_loop = FalseThe simplest form:
- in_event_loop = False - try: - asyncio.get_running_loop() - in_event_loop = True - except RuntimeError: - in_event_loop = False + try: + asyncio.get_running_loop() + in_event_loop = True + except RuntimeError: + in_event_loop = FalseEven simpler one-liner:
- in_event_loop = False - try: - asyncio.get_running_loop() - in_event_loop = True - except RuntimeError: - in_event_loop = False + try: + asyncio.get_running_loop() + in_event_loop = True + except RuntimeError: + in_event_loop = FalseCleanest version:
- in_event_loop = False try: asyncio.get_running_loop() in_event_loop = True except RuntimeError: in_event_loop = False🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/auth_capture_proxy.py` around lines 74 - 79, The variable in_event_loop is initialized to False then set to True in the try branch but redundantly reassigned to False in the except branch; remove the no-op assignment from the except block so the logic becomes: initialize in_event_loop = False, call asyncio.get_running_loop() inside try and set in_event_loop = True on success, and leave the except RuntimeError block empty (or just pass). Update the code around the asyncio.get_running_loop() usage and the in_event_loop variable so only the try branch sets True and the except does not reassign False.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Outside diff comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 207-218: reset_data() is annotated to return None but contains a
guard that returns await self._build_response(...) (a web.Response), causing a
type mismatch; either remove the unreachable guard (preferred if
_ensure_session() always raises on failure) by deleting the "if session is None"
block so reset_data remains -> None, or explicitly make the function return
Optional[web.Response] and update callers (e.g., change_host_url which currently
awaits reset_data() and ignores its return) to check/propagate the web.Response;
reference symbols: reset_data, _ensure_session, _build_response,
change_host_url.
---
Duplicate comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 228-236: Add an asyncio.Lock to prevent concurrent creation of
duplicate httpx.AsyncClient instances: in __init__ create self._session_lock =
asyncio.Lock(), then update _ensure_session to acquire that lock before
checking/creating the session so only one coroutine runs
asyncio.to_thread(self.session_factory) and assigns self.session; after
acquiring the lock, re-check if self.session is still None, create the client
via asyncio.to_thread(self.session_factory) only if needed, assign to
self.session, and release the lock (using async with self._session_lock) to
avoid leaking the overwritten client.
---
Nitpick comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 74-79: The variable in_event_loop is initialized to False then set
to True in the try branch but redundantly reassigned to False in the except
branch; remove the no-op assignment from the except block so the logic becomes:
initialize in_event_loop = False, call asyncio.get_running_loop() inside try and
set in_event_loop = True on success, and leave the except RuntimeError block
empty (or just pass). Update the code around the asyncio.get_running_loop()
usage and the in_event_loop variable so only the try branch sets True and the
except does not reassign False.
Owner
|
Why is your venv in git? |
Contributor
Author
I accidentally committed my virtual environment directory. |
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)
pyproject.toml (1)
72-72:⚠️ Potential issue | 🟠 Major
mypydev-dependency upper bound<1.0excludes all modern mypy releases.mypy 1.0 shipped in early 2023; the current release is 1.x. This constraint means the project can never install a mypy version released in the last two years, which may be the root cause of the CI mypy issues mentioned in the PR's commit messages.
🔧 Proposed fix
-mypy = ">=0.790, <1.0" +mypy = ">=1.0, <2.0"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@pyproject.toml` at line 72, The mypy dev-dependency constraint "mypy = \">=0.790, <1.0\"" prevents installing modern 1.x releases; update the mypy version specifier in pyproject.toml to allow current releases (for example change to "mypy = \">=0.790, <2.0\"" or remove the upper bound) so CI can install mypy 1.x, then run the type checks/CI to confirm compatibility; locate the mypy dependency line in pyproject.toml to make this edit.
🧹 Nitpick comments (3)
pyproject.toml (1)
169-171:target-versionin[tool.black]is inconsistent with the declared Python support.
target-version = ["py38"]whilerequires-python = ">=3.10"and classifiers list 3.10–3.13. Black won't use newer syntax features available from 3.10+ unless the target version reflects the actual minimum.♻️ Proposed fix
[tool.black] line-length = 100 -target-version = ["py38"] +target-version = ["py310", "py311", "py312", "py313"]🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@pyproject.toml` around lines 169 - 171, The Black config in pyproject.toml uses target-version = ["py38"] which contradicts requires-python = ">=3.10"; update the [tool.black] setting so Black targets the actual supported Python baseline (e.g. change target-version to ["py310"] or remove target-version entirely so Black infers from the environment), ensuring the symbol target-version in [tool.black] matches the package's requires-python and classifier Python versions.authcaptureproxy/auth_capture_proxy.py (1)
240-241: Missing blank line between_ensure_sessionandrefresh_tests.PEP 8 requires a blank line between method definitions in a class body. The current code elides it, which is at minimum a linter violation and visually merges two unrelated methods.
🔧 Proposed fix
- self.session = await asyncio.to_thread(self.session_factory) - def refresh_tests(self) -> None: + self.session = await asyncio.to_thread(self.session_factory) + + def refresh_tests(self) -> None:🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/auth_capture_proxy.py` around lines 240 - 241, Insert a single blank line separating the end of the preceding async method (the method that assigns self.session via "self.session = await asyncio.to_thread(self.session_factory)", i.e., _ensure_session) and the start of the next method "def refresh_tests(self) -> None:" so that there is one empty line between these two method definitions to satisfy PEP 8 and improve readability.authcaptureproxy/examples/modifiers.py (1)
219-219: Narrow the bareexcept Exceptioncatches toValueError.
yarl.URL()raisesValueErrorfor invalid input. CatchingExceptionmasks programming errors (e.g.,AttributeError,TypeError) and is flagged by Ruff BLE001.♻️ Proposed fix (both sites)
- except Exception: + except ValueError: url = NoneApply identically at line 219 (style branch) and line 239 (else branch).
Also applies to: 239-239
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/examples/modifiers.py` at line 219, Change the bare "except Exception" handlers that wrap yarl.URL() validation to catch ValueError only: locate the try/except blocks around the yarl.URL(...) calls (the handlers currently using "except Exception" at the two sites shown) and replace them with "except ValueError" so only invalid-URL errors are caught (apply the same change to both occurrences mentioned).
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 383-384: The guard that handles a missing HTTP session currently
returns self._build_response(text="Internal error: HTTP session not
initialized") which yields a 200 OK; update the call in auth_capture_proxy.py
(the branch checking session is None) to return an internal-server-error
response by either passing status=500 into _build_response or raising/returning
a web.HTTPInternalServerError; modify the invocation near the session check so
it uses _build_response(..., status=500) (or equivalent) to ensure clients
receive a 500 error.
In `@authcaptureproxy/examples/modifiers.py`:
- Around line 222-223: The code calls re.sub(pattern, replacement,
attribute_value) where replacement is obtained via replacement = await
run_func(modifier, name="", url=url) but modifiers (swap_url, prepend_url,
replace_empty_url) return URL objects, causing TypeError; fix by converting the
replacement to a string before calling re.sub (e.g., wrap replacement with
str(...) or ensure run_func returns a str), mirroring the approach used in
find_regex_urls; update the code around run_func/modifier (and variables
replacement, new_value, attribute_value, pattern) so re.sub always receives a
str or callable.
---
Outside diff comments:
In `@pyproject.toml`:
- Line 72: The mypy dev-dependency constraint "mypy = \">=0.790, <1.0\""
prevents installing modern 1.x releases; update the mypy version specifier in
pyproject.toml to allow current releases (for example change to "mypy =
\">=0.790, <2.0\"" or remove the upper bound) so CI can install mypy 1.x, then
run the type checks/CI to confirm compatibility; locate the mypy dependency line
in pyproject.toml to make this edit.
---
Nitpick comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 240-241: Insert a single blank line separating the end of the
preceding async method (the method that assigns self.session via "self.session =
await asyncio.to_thread(self.session_factory)", i.e., _ensure_session) and the
start of the next method "def refresh_tests(self) -> None:" so that there is one
empty line between these two method definitions to satisfy PEP 8 and improve
readability.
In `@authcaptureproxy/examples/modifiers.py`:
- Line 219: Change the bare "except Exception" handlers that wrap yarl.URL()
validation to catch ValueError only: locate the try/except blocks around the
yarl.URL(...) calls (the handlers currently using "except Exception" at the two
sites shown) and replace them with "except ValueError" so only invalid-URL
errors are caught (apply the same change to both occurrences mentioned).
In `@pyproject.toml`:
- Around line 169-171: The Black config in pyproject.toml uses target-version =
["py38"] which contradicts requires-python = ">=3.10"; update the [tool.black]
setting so Black targets the actual supported Python baseline (e.g. change
target-version to ["py310"] or remove target-version entirely so Black infers
from the environment), ensuring the symbol target-version in [tool.black]
matches the package's requires-python and classifier Python versions.
There was a problem hiding this comment.
🧹 Nitpick comments (1)
authcaptureproxy/auth_capture_proxy.py (1)
239-240: Missing blank line between method definitions.PEP 8 expects at least one blank line between methods in a class body.
Proposed fix
self.session = await asyncio.to_thread(self.session_factory) + def refresh_tests(self) -> None:🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/auth_capture_proxy.py` around lines 239 - 240, There is no blank line separating the end of the previous async method and the next method definition; insert a single blank line between the statement "self.session = await asyncio.to_thread(self.session_factory)" and the following "def refresh_tests(self) -> None:" so the class methods comply with PEP 8 (one blank line between method definitions).
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Duplicate comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 382-383: The error guard for a missing session currently returns a
200 because _build_response defaults to status 200; update the check in the
method where "if session is None" occurs to call self._build_response with an
explicit server error status (e.g., await self._build_response(text="Internal
error: HTTP session not initialized", status=500)) so the response is a 500
instead of 200; ensure you modify the call to _build_response in that block
only.
---
Nitpick comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 239-240: There is no blank line separating the end of the previous
async method and the next method definition; insert a single blank line between
the statement "self.session = await asyncio.to_thread(self.session_factory)" and
the following "def refresh_tests(self) -> None:" so the class methods comply
with PEP 8 (one blank line between method definitions).
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@authcaptureproxy/auth_capture_proxy.py`:
- Around line 212-218: In reset_data, currently session is nulled then
_ensure_session is awaited and an early return occurs if session is still None,
which leaves self.last_resp, self.query, self.data, self._active and
self._all_handler_active stale; fix by unconditionally resetting those fields
(set self.last_resp = None, self.query = None, clear self.data, set self._active
= False, self._all_handler_active = False) before the "if self.session is None"
guard (i.e., after setting self.session = None and before awaiting/after
awaiting _ensure_session as appropriate) so that any race with stop_proxy or a
re-created session in all_handler starts from a clean state; keep references to
reset_data, _ensure_session and all_handler to locate the change.
alandtse
requested changes
Feb 20, 2026
alandtse
requested changes
Feb 28, 2026
Owner
alandtse
left a comment
alandtse
left a comment
There was a problem hiding this comment.
There's a lot random files touched and also unresolved so comments.
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Contributor
Author
|
@coderabbitai review |
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
authcaptureproxy/examples/modifiers.py (1)
217-220: Catch specific exceptions in URL parsing, notException.Lines 219 and 240 use bare
except Exception, which hides unrelated failures. ForURL(...)with string input, catch only the documented exceptions:TypeError(invalid type) andValueError(decode errors onSplitResult).Proposed fix
- except Exception: + except (TypeError, ValueError): url = None ... - except Exception: + except (TypeError, ValueError): url = None🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/examples/modifiers.py` around lines 217 - 220, The try/except around URL(match.group(0)) is too broad—replace the bare except Exception with a specific except catching TypeError and ValueError only; update both occurrences (the try that assigns url: Optional[URL] = URL(match.group(0)) and the similar try later around URL(...) at the other occurrence) so they catch (TypeError, ValueError) to avoid hiding unrelated errors while preserving expected parse failures.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@authcaptureproxy/examples/modifiers.py`:
- Around line 210-216: The current regex stored in pattern expects the full
attribute (style="...") but html_tag.get("style") returns only the style value,
so re.search never matches; change the regex used in this block (the pattern
variable) to match the style content itself (e.g., match "background-image:
url(...)" and capture the inner URL) and update the subsequent re.search/re.sub
usage to use that capture (or replace using a regex that targets
background-image:url(...) inside attribute_value) so replacements operate on
attribute_value returned by html_tag.get and not on the full attribute syntax.
---
Nitpick comments:
In `@authcaptureproxy/examples/modifiers.py`:
- Around line 217-220: The try/except around URL(match.group(0)) is too
broad—replace the bare except Exception with a specific except catching
TypeError and ValueError only; update both occurrences (the try that assigns
url: Optional[URL] = URL(match.group(0)) and the similar try later around
URL(...) at the other occurrence) so they catch (TypeError, ValueError) to avoid
hiding unrelated errors while preserving expected parse failures.
ℹ️ Review info
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
.gitignoreauthcaptureproxy/auth_capture_proxy.pyauthcaptureproxy/examples/modifiers.py
🚧 Files skipped from review as they are similar to previous changes (1)
- .gitignore
- except Exception:
+ except (TypeError, ValueError):
url = None
...
- except Exception:
+ except (TypeError, ValueError):
url = None
Fixed `except (TypeError, ValueError)::`
Contributor
Author
|
@coderabbitai resume |
✅ Actions performedReviews resumed. |
Updated regex pattern to match background-image URLs more accurately.
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
authcaptureproxy/examples/modifiers.py (1)
203-225:⚠️ Potential issue | 🟠 MajorInline-style URL rewriting still never runs.
Line 204 calls
soup.find_all(tag)withtag == "style=True"(a string literal). BeautifulSoup treats this as a tag name filter searching for tags literally named"style=True", which don't exist in HTML, so the loop never executes and the inline CSS background-image rewriting logic (lines 206–230) is unreachable. Use the keyword argumentsoup.find_all(style=True)to find tags with astyleattribute.Suggested fix
for tag, attribute in search.items(): - for html_tag in soup.find_all(tag): - if tag == "style=True": + html_tags = soup.find_all(style=True) if tag == "style=True" else soup.find_all(tag) + for html_tag in html_tags: + if tag == "style=True":🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@authcaptureproxy/examples/modifiers.py` around lines 203 - 225, The loop currently calls soup.find_all(tag) where tag can be the string literal "style=True", which never matches; change the logic that iterates search.items() to detect the special key "style=True" (or when tag == "style=True") and call soup.find_all(style=True) instead, and set attribute to "style" so the existing inline-CSS handling runs; keep the rest of the block (pattern, re.search, URL(...) parsing, run_func(modifier,...), and html_tag[attribute] replacement) unchanged so modifier is invoked for inline background-image URLs.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Outside diff comments:
In `@authcaptureproxy/examples/modifiers.py`:
- Around line 203-225: The loop currently calls soup.find_all(tag) where tag can
be the string literal "style=True", which never matches; change the logic that
iterates search.items() to detect the special key "style=True" (or when tag ==
"style=True") and call soup.find_all(style=True) instead, and set attribute to
"style" so the existing inline-CSS handling runs; keep the rest of the block
(pattern, re.search, URL(...) parsing, run_func(modifier,...), and
html_tag[attribute] replacement) unchanged so modifier is invoked for inline
background-image URLs.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 88377177-49b0-4c33-b04b-d7be966e8083
📒 Files selected for processing (1)
authcaptureproxy/examples/modifiers.py
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.
Fixes: #47
Summary by CodeRabbit
Bug Fixes
Refactor
Style
Chores