aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeLines
...
| * | | | | Remove placeholder in shorten callGravatar Numerlor2021-03-05-1/+1
| | | | | |
| * | | | | Use clearer branchingGravatar Numerlor2021-03-05-5/+3
| | | | | | | | | | | | | | | | | | | | | | | | Co-authored-by: MarkKoz <[email protected]>
| * | | | | Add comments for purpose of DocItem attributesGravatar Numerlor2021-03-05-5/+5
| | | | | |
| * | | | | Do _item_futures cleanup in _parse_queue instead of a concurrent taskGravatar Numerlor2021-03-05-33/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The doc_cache coro was changed to be awaited directly instead of creating a task to ensure the cache is populated before the item is deleted
| * | | | | Set the result of the future instead of an exception when avaialbleGravatar Numerlor2021-03-05-1/+5
| | | | | |
| * | | | | Do not set redis results in get_symbol_embedGravatar Numerlor2021-03-05-3/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The redis results are already being set in the BatchParser for all symbols
| * | | | | Typo and docstring style changesGravatar Numerlor2021-03-05-11/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Co-authored-by: MarkKoz <[email protected]> fixup! Docstring typos and style changes
| * | | | | Move the seconds of a week timedelta into a constantGravatar Numerlor2021-03-05-1/+3
| | | | | |
| * | | | | Use deque instead of a listGravatar Numerlor2021-03-05-5/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As we extend the left side of a list now, using a deque that provides a direct method for it is more fitting. fixup! Use deque instead of a list
| * | | | | Use common check for early exitGravatar Numerlor2021-03-05-11/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This introduces a possibly redundant check for the doc_item being None but results in flatter code with less duplication
| * | | | | Add new symbols to front of queue instead of extending the endGravatar Numerlor2021-02-28-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Using extend caused old items, including the ones that were requested by users and pushed to the front, to be pushed back by all of the items on the new page, possibly significantly delaying their parsing
| * | | | | Add digits to package covnerterGravatar Numerlor2021-02-27-2/+2
| | | | | |
| * | | | | Get the last index instead of using maxGravatar Numerlor2021-02-27-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The last index will always be the largest one so there's no need for max to search for it
| * | | | | Use 4 spaces for hanging indentGravatar Numerlor2021-02-24-19/+19
| | | | | |
| * | | | | Allow wildcard match for clear cache commandGravatar Numerlor2021-02-24-3/+7
| | | | | |
| * | | | | Update docstringsGravatar Numerlor2021-02-23-2/+2
| | | | | |
| * | | | | Remove redundant group checkGravatar Numerlor2021-02-23-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As we check for non dt symbol names to be sure in case something pops up , we can skip the initial group check as all the symbols from those groups should point to non dt tags.
| * | | | | Correct length adjustmentGravatar Numerlor2021-02-23-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the placeholder contains a space and is actually 4 chars because of that with a comma that adds up to 5 characters in the signature instead of 4
| * | | | | Add unittests for _split_signatureGravatar Numerlor2021-02-23-0/+59
| | | | | |
| * | | | | Add lt and gt angle brackets to bracket pairsGravatar Numerlor2021-02-23-0/+1
| | | | | |
| * | | | | Add deletedoccache alias to the clear cache commandGravatar Numerlor2021-02-23-1/+1
| | | | | |
| * | | | | cleanup->clean_upGravatar Numerlor2021-02-23-2/+2
| | | | | |
| * | | | | Remove unnecessary commentsGravatar Numerlor2021-02-23-10/+0
| | | | | | | | | | | | | | | | | | | | | | | | The comments explain things that should be clear, or basic concepts
| * | | | | Use "inventories" when referring to the cog's collection of sphinx invsGravatar Numerlor2021-02-23-9/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously in some context inventory referred both to a single "inventory" that we got from a remote objects.inv and to the internal cog inventories. Always referring to the cog's inventories as plural The update_single docstring was also changed from rebuild to build, as the method doesn't handle anything with a preexisting inventory with the same symbols being in the cog's inventories
| * | | | | Simplify conditionGravatar Numerlor2021-02-23-1/+1
| | | | | |
| * | | | | Simplify the _split_parameters implementationGravatar Numerlor2021-02-23-30/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The main simplification was getting rid of keeping track of string depth which was unnecessary, as we can just always skip them as was being done for strings inside of brackets. The branching was also simplified to make sure less unnecessary checks were being done with a bit less confusing elifs.
| * | | | | Fix docstring typosGravatar Numerlor2021-02-23-10/+10
| | | | | |
| * | | | | Remove unnecessary use of partialGravatar Numerlor2021-02-05-6/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | run_in_executor can provide args to the func it's passed in, making the use of partial unnecessary. This will also make it more convenient to move to asyncio.to_thread when the codebase is switched to python 3.9
| * | | | | Avoid from import on _batch_parserGravatar Numerlor2021-02-03-3/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The tests import the modules the other way around causing a circular import
| * | | | | Use create_task utilGravatar Numerlor2021-02-03-2/+3
| | | | | |
| * | | | | Create a task for setting the redis result instead of awaitingGravatar Numerlor2021-02-03-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The queue parsing doesn't depend on anything with redis, so the await only delays the result being set on the future.
| * | | | | Move BeautifulSoup parsing into an executorGravatar Numerlor2021-02-03-1/+4
| | | | | |
| * | | | | Remove url lockGravatar Numerlor2021-02-03-3/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The items are added to the futures dict before a context switch can occur, making the subsequent requests to the url skip the queue extend and suspend at the future await
| * | | | | Use more consistent naming for doc items and symbolsGravatar Numerlor2021-02-03-71/+71
| | | | | | | | | | | | | | | | | | | | | | | | Trailing commas were also added where missing
| * | | | | Import module directly to avoid circular dependency and use of typeGravatar Numerlor2021-02-03-13/+11
| | | | | |
| * | | | | Attempt to resolve first part of the argument in case of a failed lookupGravatar Numerlor2021-02-02-2/+12
| | | | | |
| * | | | | Require a set of names to ignore instead of a blanket ignoringGravatar Numerlor2021-01-30-16/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As work is done on the modules the wrapper and wrapped functions are in more conflicts can occur, previously this could be missed as the info log that was done in case the error was suppressed was done when modules were being initialized during which there is a logging spam.
| * | | | | Cancel current task inside coro to not keep track of attempts manuallyGravatar Numerlor2021-01-23-8/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The scheduler shields the coroutine from cancellation so we can cancel the scheduler's tasks inside of it to avoid the error from multiple tasks with the same id trying to be scheduled which the manual tracking of attempts solved Co-authored-by: MarkKoz <[email protected]>
| * | | | | Wrap whole string in shortenGravatar Numerlor2021-01-23-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This helps avoid subtracting the length of "Moved: " from the shorten index Co-authored-by: MarkKoz <[email protected]>
| * | | | | Avoid errors when the first element is longer than the truncation indexGravatar Numerlor2021-01-23-3/+10
| | | | | |
| * | | | | lock markdownify version to 0.6.1Gravatar Numerlor2021-01-23-5/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the 0.6.3 update brought a change that fails to ignore newlines in the html, introducing unnecessary lines into the output
| * | | | | Create a namespace varGravatar Numerlor2021-01-23-8/+9
| | | | | |
| * | | | | Properly handle cache being clearedGravatar Numerlor2021-01-23-13/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously the code deleted the entry of all of the DocItems of the page after its contents were requested once, but this caused problems when the cache was cleared when it expired. Instead of deleting the entry to check if it should be queued on the next item request, we keep it and create an entry in the _item_futures dict for all items again and check for containment there. To avoid populating the queue multiple times with the same item in some cases the futures cleanup task will now only run when the queue is empty
| * | | | | Merge remote-tracking branch 'upstream/master' into doc-impGravatar Numerlor2021-01-22-323/+1314
| |\ \ \ \ \
| * | | | | | Fix issues with multiple gets being suspended when a refresh startsGravatar Numerlor2021-01-22-6/+34
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | With a normal event, if multiple gets were suspended and a refresh started, we'd continue the refresh after the first get finished and set the event which would be the same behaviour as the one it tried to fix. This is avoided by using a counter that's incremented every time a context manager is entered around an event and only setting the event when that counter reaches a zero after everything exited the context mgr
| * | | | | | Use inspect.unwrap instead of manually unwrappingGravatar Numerlor2021-01-21-3/+1
| | | | | | |
| * | | | | | Use a defaultdict for item futuresGravatar Numerlor2021-01-15-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | To be able to set the attribute at the start of the coro we need to be able to access the item's future before we know about all the other items. This also saves us from having to add them all as the queue parser or get_markdown will create the futures for us dynamically
| * | | | | | Prevent an inventory refresh while waiting for item cacheGravatar Numerlor2021-01-14-0/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If an inventory refresh was started while the symbol embed coroutine was suspended, it could cause the parser to try to fetch a non existent future if the markdown was requested after it was cleared but before new inventories were loaded in.
| * | | | | | Raise an error or log a warning if there's a global name conflictGravatar Numerlor2021-01-14-5/+35
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When wrapper uses a global name, which conflicts with a global name from wrapped's module that wrapped uses for its annotations, we run into a situation that can't be solved without changing one of the names, so an error is raised to give this clearer meaning. The check may be erroneous in some edge cases or the objects the conflicting names refer to can be functionally identical, so the error can be turned into a logged warning.
| * | | | | | Ensure footer is actually max 100 charsGravatar Numerlor2021-01-14-2/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Shortening the renamed symbols string to 100 chars is not accurate as the footer also contains a string before that, subtracting its length fixes this.
generated by cgit v1.2.3 (git 2.39.1) at 2025-10-26 11:52:46 +0000