about summary refs log tree commit homepage
DateCommit message (Collapse)
2017-01-11qspawn: fix bad error reporting on errors
Oops :x
2017-01-10introduce PublicInbox::MIME wrapper class
This should fix problems with multipart messages where text/plain parts lack a header. cf. git clone --mirror https://github.com/rjbs/Email-MIME.git refs/pull/28/head In the future, we may still introduce as streaming interface to reduce memory usage on large emails.
2017-01-08githttpbackend: use psgi_return shortcut
This drastically cuts down the amount of duplicate code we have in this branch.
2017-01-08httpd/async: remove needless sysread wrapper
We don't appear to be using it anywhere
2017-01-08Merge remote-tracking branch 'origin/master' into repobrowse
* origin/master: inbox: properly register cleanup timer for git processes search: remove subject_summary searchmsg: favor direct hash access over accessor methods remove incorrect comment about strftime + locales config: allow per-inbox nntpserver inbox: eliminate weaken usage entirely inbox: describe the full key name config: remove unused get() method config: always use namespaced "publicinboxlimiter" qspawn: prepare to support runtime reloading of Limiter http: remove weaken usage, reduce anonsub capture scope httpd/async: remove weaken usage http: fix spelling error watch: watchspam affects all configured inboxes doc: minor updates to design notes
2017-01-07initial git async work
This will allow us to handle network operations while waiting on "git cat-file" to seek and unpack things.
2017-01-07inbox: drop $ref arg for writing destination buffer
We never used this feature, so lets drop it for now since we can have fine-grained memory release with reference counting, anyways.
2017-01-07inbox: properly register cleanup timer for git processes
We still need to cleanup git processes occasionally, since "git cat-file --batch" does not release old packs (and git processes are fairly expensive). For SQLite and Xapian file handles, they should be capable of managing themselves without too much trouble, so lets try keeping them for the lifetime of a process.
2017-01-07search: remove subject_summary
Apparently it never actually got used, and the world seems fine without it, so we can drop it. While we're at it, consider removing our subject_path usage from existence, too. We are not using fancy subject-line based URLs, here.
2017-01-07searchmsg: favor direct hash access over accessor methods
This is faster, smaller, and more straighforward to me with fewer layers of indirection.
2017-01-07remove incorrect comment about strftime + locales
We only need strftime to be locale-independent when generating dates for email and HTTP headers. Purely numeric dates can use strftime for ease-of-readability.
2017-01-07config: allow per-inbox nntpserver
This allows certain inboxes to override the global nntpserver (perhaps under a different domain).
2017-01-07inbox: eliminate weaken usage entirely
We can do a better job initializing the data structure so we no longer need to rely on weak references to cleanup when we ditch the config on reload.
2017-01-07inbox: describe the full key name
Hopefully make this easier for future generations to understand.
2017-01-07config: remove unused get() method
This seems like an unnecessary abstraction, or an abstraction on the wrong level.
2017-01-07config: always use namespaced "publicinboxlimiter"
I'm not sure if we'll ever support sharing a config file with other tools, but maybe we will, and "limiter" is too generic.
2017-01-07qspawn: prepare to support runtime reloading of Limiter
We may allow the {max} value of a limiter to be changed in the future, so lets start accounting for it before we spawn followup processes.
2017-01-04http: remove weaken usage, reduce anonsub capture scope
Avoiding weaken here is no more dangerous than the existing circular refs (e.g. psgix.io) we create and manage throughout the lifetime of the connection. So, trust ourselves to maintain the data structure properly and avoid triggering extra memory usage. While we're at it, avoid having anonymous subroutines capture more variables than necessary to simplify reference auditing.
2017-01-04httpd/async: remove weaken usage
We do not need to use weaken() here, so avoid it to simplify our interactions with Perl; as weaken requires additional storage and (it seems) time complexity.
2017-01-04http: fix spelling error
Oops. And we'll be fixing circular references from now...
2017-01-02watch: watchspam affects all configured inboxes
If a message is spam in one mailbox, it is spam in all others a particular user/group will care about.
2017-01-01repobrowse: avoid empty pathspecs for future git compatibility
At the moment, we always set expath, so it will always be defined.
2016-12-26doc: minor updates to design notes
ssoma is not worth marketing, but perhaps our mirror of the git mailing list archives is...
2016-12-26spawn: remove non-blocking support, here
It is never used, and inappropriate to support in generic code. HTTPD::Async already sets non-blocking, and it's better to do it in -httpd-specific code since we know our -httpd can handle it.
2016-12-26repobrowse: port git snapshot over to qspawn
This is expensive, so we will utilize the qspawn system to prevent excessive overhead.
2016-12-26repobrowse: port patch generation over to qspawn
And start generalizing the qspawn usage code for PSGI with psgi_return.
2016-12-26git: lazy initialization of error output
We may not keep this feature after all, but for now we'll hold off on creating it to cheapen instantiation.
2016-12-26Merge remote-tracking branch 'origin/master' into repobrowse
* origin/master: (25 commits) evcleanup: ensure deferred close from timers are handled ASAP httpd/async: improve variable naming githttpbackend: minor cleanups to improve readability githttpbackend: simplify compatibility code githttpbackend: minor readability improvement http: fix clobbering of $null_io linkify: modify argument in place view: do not modify array during iteration view: stop chomping off whitespace at ends of messages view: remove unused parameter search: lookup_mail handles modified DBs doc: various comments on async handling searchthread: simplify API and remove needless OO searchthread: update comment about loop prevention searchmsg: remove ensure_metadata tests: add thread-all testing for benchmarking searchmsg: do not memoize {date} field searchmsg: remove locale-dependency for ->date t/config.t: fix feedmax default wwwtext: link to RFC4685 (Atom Threading) ...
2016-12-26evcleanup: ensure deferred close from timers are handled ASAP
Danga::Socket defers close() syscalls until the end of the event loop to avoid FD recycling. Unfortunately, this is dependent on IO events firing and waking the process up from poll/kevent/epoll_wait. Without any I/O activity, a socket could remain in the @Danga::Socket::ToClose array indefinitely. Thus, we will trigger a fake IO event after running all timers to trigger the deferred close in Danga::Socket::PostEventLoop.
2016-12-26t/repobrowse_git_httpd: remove XS parser dependency
Relying on the XS parser has been optional since March 2016: commit 7dd78012da81d48e5e73e56c3255895dfa9de1f5 ("http: use Plack::HTTPParser for HTTP parsing")
2016-12-25httpd/async: improve variable naming
We only refer to PublicInbox::HTTP objects here, so '$io' was a bad name.
2016-12-25githttpbackend: minor cleanups to improve readability
Fewer returns improves readability and the diffstat agrees.
2016-12-25githttpbackend: simplify compatibility code
Fewer conditionals means theres fewer code paths to test and makes things easier-to-read.
2016-12-25githttpbackend: minor readability improvement
Use a more meaningful variable name for the Qspawn object, since this module is the reference for its use.
2016-12-25http: fix clobbering of $null_io
Oops, this would be disatrous if we started handling bigger request bodies or slow clients. Fixes: c008654229a9 ("avoid IO::File for anonymous temporary files")
2016-12-24linkify: modify argument in place
This results in over 1% speedup doing $MESSAGE_ID/T/ HTML generation for a 368-message thread.
2016-12-24view: do not modify array during iteration
This results in a half percent speedup or so doing $MESSAGE_ID/T/ HTML generation for a 368 message thread.
2016-12-24view: stop chomping off whitespace at ends of messages
This allows a 3-4% speedup in $MESSAGE_ID/T/ page generation speed for a 368+ message thread. It also more faithfully preserves the message as intended; even if the it makes the sender look like a space-wasting slob :P
2016-12-24view: remove unused parameter
And add a comment about it to remind our future selves.
2016-12-22search: lookup_mail handles modified DBs
We call lookup_mail all over the place, be sure we can handle database modifications in those cases.
2016-12-22doc: various comments on async handling
Notes for future developers (myself included) since we can't assume people can read my mind.
2016-12-22repobrowse: remove Plack::Request dependency
This does not make installation easier, but lightens runtime a bit. Plack::Request is unnecessary bloat and indirection which does things behind our back. $env has all the stuff we need.
2016-12-21searchthread: simplify API and remove needless OO
This simplifies callers to prevent errors and avoids needless object-orientation in favor of a single procedure call to handle threading and ordering.
2016-12-21searchthread: update comment about loop prevention
It definitely is necessary to prevent looping with the %seen hash.
2016-12-20repobrowse: git plain: fixup request handling
Leftover from commit d4cec3bc3e439ab6ea9a2f89da5f98033bd95b4f ("repobrowse: start reducing Plack::Request dependencies")
2016-12-20searchmsg: remove ensure_metadata
Instead, only preload the ->mid field for threading, as we only need ->thread and ->path once in Search->get_thread (but we will need the ->mid field repeatedly). This more than doubles View->load_results performance on according to thread-all on an inbox with over 300K messages.
2016-12-20tests: add thread-all testing for benchmarking
I'll be using this to improve message threading performance.
2016-12-17searchmsg: do not memoize {date} field
We only generate the ->date once in NNTP, so creating the hash entry is a waste.
2016-12-17searchmsg: remove locale-dependency for ->date
strftime is locale-dependent, which can cause surprising failures for some users.
2016-12-17t/config.t: fix feedmax default
Oops :x