P

plugin-Library-staging

Settings | Report Duplicate

1

I Use This!

Inactive

Commits : Listings

Analyzed about 22 hours ago. based on code collected about 22 hours ago.

Commit Message	Contributor	Files Modified	Lines Added	Lines Removed	Code Location	Date
Feb 26, 2025 — Feb 26, 2026 Showing page 9 of 37 Search / Filter on:
Add new class BoundedPriorityBlockingQueue. We were using PriorityBlockingQueue, which isn't bounded, so of course we got tens of thousands of btree's in the queue, and ran out of memory. Use it, and impose a limit on proc_val.out as well.	Matthew Toseland	More...				almost 16 years ago
Even more nasty hacking related to not having proper async!	Matthew Toseland	More...				almost 16 years ago
Maybe fix serialiser	Matthew Toseland	More...				almost 16 years ago
Fix pull?	Matthew Toseland	More...				almost 16 years ago
Don't try to insert anyway.	Matthew Toseland	More...				almost 16 years ago
Fix critical bug: we were starting from scratch after every restart, because we did not pull the previous index!	Matthew Toseland	More...				almost 16 years ago
Logging	Matthew Toseland	More...				almost 16 years ago
Show full trace if logMINOR enabled.	Matthew Toseland	More...				almost 16 years ago
Logging	Matthew Toseland	More...				almost 16 years ago
Doh	Matthew Toseland	More...				almost 16 years ago
Fix failure to detect from MIME type race condition.	Matthew Toseland	More...				almost 16 years ago
Sanity check.	Matthew Toseland	More...				almost 16 years ago
Fix FIXMEs	Matthew Toseland	More...				almost 16 years ago
More comments explaining two-phase pull.	Matthew Toseland	More...				almost 16 years ago
Comments.	Matthew Toseland	More...				almost 16 years ago
Reduce output queue on puller to minimise stuff getting bogged down in memory when it can't be processed immediately.	Matthew Toseland	More...				almost 16 years ago
Severely throttle the number of deflates running in parallel. Because inserts are asynchronous this should not harm anything. Because things that generate deflate nodes mostly run inline, this should not result in high memory usage backlogs downstream either.	Matthew Toseland	More...				almost 16 years ago
Fix progress	Matthew Toseland	More...				almost 16 years ago
Don't ignore async push when progress == null	Matthew Toseland	More...				almost 16 years ago
Individually limit parallelism for proc_deflate	Matthew Toseland	More...				almost 16 years ago
Clear the data after we have set the metadata.	Matthew Toseland	More...				almost 16 years ago
Make sure we use the same FreenetArchiver here.	Matthew Toseland	More...				almost 16 years ago
More dirty hacking, document thoroughly. push() now returns, and tells SimpleProgress that it is done, as soon as it has a URI. The ultimate caller must call waitForAsyncInserts() before finishing. IterableSerialiser should be split up, but these hacks allow us to minimise memory usage without that, because we never have tons of memory pinned and are waiting for an insert.	Matthew Toseland	More...				almost 16 years ago
Null out data once we have the URI.	Matthew Toseland	More...				almost 16 years ago
Refactor, keep SimpleProgress up to date even when async. Caller can then join(), so this is more or less asynchronous.	Matthew Toseland	More...				almost 16 years ago
Logging: timing.	Matthew Toseland	More...				almost 16 years ago
Don't loop if broken.	Matthew Toseland	More...				almost 16 years ago
Fix NPE	Matthew Toseland	More...				almost 16 years ago
Don't leak.	Matthew Toseland	More...				almost 16 years ago
Do the inserts asynchronously - at least once they've generated URIs (using EarlyEncode).	Matthew Toseland	More...				almost 16 years ago

←
1
2
…
5
6
7
8
9
10
11
12
13
…
36
37
→