about summary refs log tree commit homepage
DateCommit message (Collapse)
2021-07-31extindex: -xcpdb and -compact support
Since extindex uses Xapian shards in a similar way to v2 inboxes, we'll support -xcpdb (reshard+upgrade) and -compact all the same to give admins tuning+upgrade options.
2021-07-31admin: index_inbox: drop unnecessary check
No callers pass an unblessed pathname to index_inbox, only Inbox object refs.
2021-07-28listener: maximize listen(2) backlog
This helps avoid errors from script/lei dying on ECONNRESET when a single lei-daemon is serving all tests when run via "make check-run". Instead of using some arbitrary limit, use INT_MAX and let the kernel clamp it (both Linux and FreeBSD do). There's no need to call listen() in LEI.pm, either, since Listener->new takes care of it.
2021-07-28lei: die on ECONNRESET
ECONNRESET should be rare on a private local socket, and if we hit it, it's because we're hitting the listen() limit.
2021-07-28treewide: s/sequential_shard/sequential-shard/g
The underscore variant was never documented and maintaining the difference between the command-line and internal hash is not worth it.
2021-07-25extindex: support --jobs/-j properly on creation for shard count
This wasn't wired up properly, but Xapian appears to suffer from I/O amplification problems as DB shards get larger: https://lists.xapian.org/pipermail/xapian-discuss/2019-February/009727.html <23640.32170.703368.841021@y.dockes.com> Of course, we shouldn't have too many shards, either; because performance problems with too many shards was the entire reason extindex was created: https://lists.xapian.org/pipermail/xapian-discuss/2020-August/009823.html <20200826064728.GA32239@dcvr>
2021-07-25doc: lei-{p2q,rediff}: note implicit --stdin
lei actually uses implicit --stdin everywhere, but I thing these patch-related commands are the most common use of them.
2021-07-25t/lei-watch.t: improve test reliability
On single CPU (and overloaded SMP) systems, we can't rely on inotify in lei-daemon firing before a "lei note-event done" client hits it. So force in a single tick() to ensure the scheduler can yield to lei-daemon and see the inotify wakeup before "lei note-event done" to commit the write.
2021-07-25init: support git <2.30 for "-c KEY=VALUE" args
It turns out `--fixed-value' is a relatively new git-config(1) feature in git 2.30+ (December 2020). So use the quotemeta perlop for now since it seems compatible-enough for POSIX ERE used by git.
2021-07-25lei_mail_sync: locations_for API uses oidbin for comparisons
Favor oidbin use internally to reduce internal memory traffic.
2021-07-25lei_inspect: fix typo
Not sure how this wasn't caught, earlier...
2021-07-25lei_search: favor binary OID comparisons
Reduce memory traffic and code, too.
2021-07-25extsearchidx: favor binary comparison in common case
We'll use 20-byte SHA-1 comparisons instead of 40-byte hex representations for a minor reduction in memory traffic.
2021-07-25extsearchidx: use more appropriate max for dedupe
The over.msgid table may contain ghost Message-IDs and also Message-IDs of deleted spam messages, so over->max isn't a good aproproximation of dedupe progress.
2021-07-25extindex: improve comment around git->async_wait_all
I found myself tempted to remove this, but it appears impossible due to odd messages which have multiple Message-IDs.
2021-07-25extindex: support --dedupe[=MSGID]
Sometimes I just want to dedupe a single Message-ID to test something, and this lets me do it. This patch appears to do what its supposed to. But it also appears to be finding duplicates that were previously missed. That's a good thing, but I wish I understood what seems to be fixed :x I'm not sure why the previous ExtSearchIdx.pm (blob 357312b8) was causing messages to be missed, even, and why this patch seems to fix it... And it's not infinite looping, either. Anyways, before this patch, "-extindex --dedupe" was taking ~5 min to no-op every message (after the initial full --dedupe run which took over a day to run). No-op --dedupes now take just under 2 hours to scan every single cross-posted message for a no-op dedupe. The initial dedupe took nearly 44 hours on my system for <https://yhbt.net/lore/all/> due to SATA-2 TLC SSD latency on 3 gigantic Xapian shards. Running --dedupe with this change seems to prevent /BUG\?.*?not deduplicated properly/ stderr messages from being triggered by View.pm. Current versions of -extindex do not seem susceptible to introducing duplicates.
2021-07-25lei rm-watch: new command to support removing watches
Pretty trivial since it just invokes "git-config". It's mainly intended to make shell completion easier.
2021-07-25lei: avoid SQLite COUNT() for dedupe
SQLite COUNT() is a slow operation that does a full table scan with no conditions. There's no need for it, since lei dedupe only needs to know if it's empty or not to decide between new/ and cur/ for Maildir outputs.
2021-07-25t/lei*: check error messages on failures
I just hit an unreproducible failure in t/lei-p2q.t and lacked $lei_err information to diagnose it. Hopefully this helps track down odd failures in the future.
2021-07-22t/solver_git: use like() to improve error reporting
I hit a test failure here, but haven't been able to reproduce it...
2021-07-22lei: auto-refresh watches in config, cancel missing
This makes behavior less surprising on restarts as we no longer lose state on restarts, so there's no need to manually run "lei add-watch" to re-enable watches. This also allows us to transparently handle changes if somebody edits the lei config file directly or via git-config(1).
2021-07-22lei: start implementing inotify Maildir support
This allows lei to automatically note keyword (message flag) changes made to a Maildir and propagate it into lei/store: lei add-watch --state=tag-ro /path/to/Maildir This doesn't persist across restarts, yet. In the future, it will be applied automatically to "lei q" output Maildirs by default (with an option to disable it). State values of tag-rw, index-<ro|rw>, import-<ro|rw> will all be supported for Maildir. This represents a fairly major internal change that's fairly intrusive, but the whole daemon-oriented design was to facilitate being able to automatically monitor (and propagate) Maildir/IMAP flag changes.
2021-07-22init: allow arbitrary key-values via -c KEY=VALUE
This won't blindly append identical key=values, but allows specifying multiple, different key=value pairs as long as the values are different.
2021-07-22extsearch: support publicinbox.*.boost parameter
This behaves identically the lei external "boost" parameter in prioritizing raw messages for extindex. Relying exclusively on the config file order doesn't work well for mirrors since it's impossible to guarantee config file ordering via grokmirror hooks. Config file ordering remains the default if boost is unconfigured, or in case of ties. Note: I chose the name "boost" rather than "priority" or "rank" since I always get confused by whether higher or lower numbers take precedence when it comes to kernel scheduling. "weight" is also a part of Xapian API terminology, which we currently do not expose to configuration (but may in the future).
2021-07-20httpd: fix SIGHUP by invalidating cache on reload
Since we require separate PublicInbox::HTTPD instances for each listen socket address (in order to support {SERVER_<NAME|PORT>} for PSGI env), the old cache needed to be invalidated on rare app refreshes. SIGHUP has always been broken in -httpd (but not -imapd or -nntpd) due to this cache. Update the daemon documentation and 5.10.1-ize some bits while we're in the area.
2021-07-18config: s/_one_val/get_1/ for public use
We'll be using this in lei for watch configs.
2021-07-08extindex: dedupe: reduce SQLite contention and dirty data
Complex queries causes SQLite to block readers for longer than their retry period. For dedupe, it was also preventing us from making good use of checkpoints due to the query time. With many deduplications, checkpoints are necessary to maintain system health due to having too much data piled up.
2021-07-08extsearchidx: ignore Eml warnings across the board
There's nothing we can do about misformatted emails and headers we get from untrusted sources. They're too noisy and those messages already exist in public-inboxes, anyways, so just keep things quiet so we can spot real problems more easily.
2021-07-06extindex: --gc: avoid SQLite lock conflict on shard cleanup
Xapian shard cleanup only requires read-only access to over.sqlite3, so avoid opening it with read-write access since create_tables will hit lock conflicts on "INSERT OR IGNORE" statements.
2021-07-06extindex: implement --dedupe to fix old extindices
This is intended to fix older indices that had deduplication bugs for matching content. It'll also make dealing with future changes to ContentHash easier since that's never guaranteed stable. It also supports --dry-run to print changes only without making them.
2021-07-06eml: relax warn_ignore regexps for current Email::Address::XS
These seem needed with the data I'm currently working on, but I haven't changed my version of Email::Address::XS since my last Debian stable upgrade (to buster).
2021-07-05lei: drop workers on EOF from clients
Sometimes a user will be bored waiting for a command to finish, so ensure we drop disconnect workers in this case.
2021-07-03lei import: increase flags search batch size, display progress
IMAP flag-only synchronization doesn't fetch entire messages, so we can safely bump the batch size iff a user specified one for full messages to 10000 times that. Since I sometimes wonder why nothing happens for several seconds after starting "lei import $URL", we'll also show some progress during the flag synchronization phase.
2021-07-03lei inspect: help+completion for --dir option
It's the most generic name I could find for it since it can mean so many things...
2021-07-03extsearchidx: extra assertions for deduplication flow
I haven't found any bugs from this (still looking for missed deduplication bugs), and it's a bit shorter and more likely to catch future bugs. Clean up an unnecessary ->{mid} array copy while we're at it, too.
2021-07-01lei inspect: support "mid:" (and "m:") prefix
Using this to track down deduplication failures in -extindex...
2021-07-01lei inspect: support automatic pager in output
All commands which output non-trivial amounts of data to the terminal should support this.
2021-07-01extsearchidx: lock before writing multi-pack-index
This avoids errors from git in case -extindex gets invoked in parallel.
2021-06-30extsearchidx: symlink .rev and .bitmap files into ALL.git
It's possible for these to exist and git can (or may eventually) take advantage of them to speed up functionality which affects us.
2021-06-30searchidx: default BATCH_BYTES to 8MB on 64-bit systems
This default seems closer to reasonable on 64-bit systems which are the norm these days. 32-bit systems gain 48K so it's an even 1 MB, but we need to keep 32-bit systems from using too much since there's still some ancient systems out there with small inboxes.
2021-06-29www: fix manifest.js.gz for default publicInbox.grokManifest
ManifestJsGz->response was not invoking the new "url_filter" method properly. Furthermore, fix url_filter for returning 404 responses. Reported-by: Kyle Meyer <kyle@kyleam.com> Link: https://public-inbox.org/meta/87fsx3128a.fsf@kyleam.com/ Fixes: 520be116e8a686cb ("www_listing: start updating for pagination + search")
2021-06-27extindex: maintain pack symlinks and use "git multi-pack-index"
This is a fair amount of complexity, but it speeds up "git cat-file --batch" startup by 3-4% with 50K packfiles with a hot kernel cache. This appears extremely sensitive to RAM available to the kernel page cache with my SATA 2 SSD. Faster storage and more RAM can bring loading pack. 2.60s vs 2.69s were the best cases on my workstation with and without the multi-pack-index, however times could be all over the place (even in the minutes) with more activity on my workstation. Getting sub-minute times requires a git patch to speed up alt_odb_usable(): <https://lore.kernel.org/20210624005806.12079-1-e@80x24.org/> Otherwise, prepare to wait several minutes.
2021-06-24www_listing: fix manifest.js.gz generation with extindex "all"
WwwListing and ManifestJsGz may be too different nowadays to be worth the code sharing between them. Update some comments and note we still needs better tests :x Fixes: 520be116e8a686cb ("www_listing: start updating for pagination + search")
2021-06-24v2writable: avoid spawning "git hash-object"
We have git_sha() nowadays that's used everywhere, so avoid process spawning overhead for "git hash-object".
2021-06-24doc: tuning: add a note about Linux sys.vm.max_map_count
git tends to die when mmap(2) fails on this limit, so let users know about it. Perhaps git could gracefully fallback.
2021-06-24favor git(1) rather than libgit2 for ExtSearch
While both git and libgit2 take around 16 minutes to load 100K alternates there's already a proposed patch to make git faster: <https://lore.kernel.org/git/20210624005806.12079-1-e@80x24.org/> It's also easier to patch and install git locally since the git.git build system defaults to prefix=$HOME and dealing with dynamic linking with libgit2 is more difficult for end users relying on Inline::C. libgit2 remains in use for the non-ALL.git case, but maybe it's not necessary (libgit2 is significantly slower than git in Debian 10 due to SHA-1 collision checking).
2021-06-23www: do not warn on blank query parameters
Sometimes users (or bots) may lead queries with '&' and trigger uninitialized variable warnings, just ignore them and give consumers a $ctx->{qp}->{''} entry. While we're in the area, pass a regexp rather than scalar string to the `split' perlop to prevent Perl from recompiling the regexp on every call.
2021-06-23www_listing: start updating for pagination + search
When dealing with thousands of inboxes, displaying all of them on a single page isn't going to work. So steal some pagination and search results code from the message search to generate some basic HTML output that looks good in w3m.
2021-06-23search: make xap_terms easier-to-use and use it more
This allows us to simplify callers throughout, and exceptions are can no longer be silently hidden. MiscSearch now uses xap_terms for looking up eidx_key terms for a code reduction. We also simplify LeiStore->_msg_kw for runtime use by moving the MsetIterator handling into t/lei_store.t test case.
2021-06-22t/hl_mod: accept "make" or "makefile" for Makefile
Version 4.0 of highlight has renamed the "make" language to "makefile". So just check the string starts with "make", to handle both 3.x and 4.x. I tested that public-inbox does actually work with highlight 4 -- it can highlight my Makefile fine. :)