| Date | Commit message (Collapse) |
|---|
|
No reason to leave that (usually) empty file open after killing off
"cat-file --batch-check". This wasn't an unbound leak, though,
as respawning the --batch-check process would've clobbered the
old err_c file.
|
|
A constant stream of traffic to either httpd/nntpd would mean
git-cat-file processes never expire. Things can go bad after a
full repack, as a full repack will unlink old pack indices and
git-cat-file does not currently detect unlinked files.
We could do something complicated by recursively stat-ing
objects/pack of every git directory and alternate;
but that's probably not worth the trouble compared to
occasionally restarting the cat-file process.
So simplify the code and let httpd/nntpd expire them
periodically, since spawning a "git-cat-file --batch" process
isn't too expensive. We already spawn for every request which
hits git-http-backend, cgit, and git-apply.
In the future, we may optionally support the Git::Raw module
to avoid IPC; but we must remain careful to not leave lingering
FDs open to unlinked files after repack.
|
|
Taking one step out of setting up a performant deployment could
make setup and administration easier (at the cost of installing
an extra-but-common XS module). This can also be useful for
the day NNTP servers see hug-of-death events.
|
|
This is worth a 1-2% speedup in t/perf-msgview.t rendering 2620
messages currently in https://public-inbox.org/meta/
|
|
* origin/v2-noop-speedup:
v2writable: short-circuit is_ancestor check on equality
v2writable: avoid mm_tmp creation without regen
v2writable: hoist out index_epoch sub
v2writable: split off unindex_range mapping
|
|
I don't have time to check and train spam for all these
projects.
Spam filtering is especially difficult on ruby-core: it
enters via Redmine, so it doesn't have a distinct Received:
chain, and also gets mixed with non-spam bug-report text,
throwing off Bayes training.
And I'm not sure if those mirrors did anybody any good, even;
so lets not say its' a "service" to anybody :P
The actual mirrors remain up, for now, but who knows...
I care about decentralization too much to ask anybody
to trust me to keep anything up :P
|
|
We don't need to use git to check ancestry if object IDs
match on a string comparison.
This saves 100ms or so and brings down the ~0.5s no-op time on
lore.kernel.org/lkml down to ~0.4s.
|
|
Creating mm_tmp is an expensive operation with large inboxes
and can be avoided if there are no new messages to process.
Since git-fetch(1) currently lacks an --exit-code option(*),
mirrors will run `public-inbox-index' unconditionally after
fetch, which is an expensive op if it needs to duplicate
a large SQLite DB.
This speeds up the mirror case of:
git --git-dir=git/$EPOCH.git fetch && public-inbox-index
This reduces the no-op `public-inbox-index' time from over 8s to
~0.5s on a (currently) 7-epoch clone of https://lore.kernel.org/lkml/
on my system.
(*) WIP --exit-code for git-fetch:
https://public-inbox.org/git/87ftphw7mv.fsf@evledraar.gmail.com/
|
|
This will make future changes easier-to-follow.
|
|
It'll make it easier to detect if we have anything to
unindex and run git-log on, at all.
|
|
And use it from Admin.
It's easy to tell what indexlevel=basic is from unconfigured
inboxes, but distinguishing between 'medium' and 'full' would
require stat()-ing position.* files which is fragile and
Xapian-implementation-dependent.
So use the metadata facility of Xapian and store it in the main
partition so Admin tools can deal better with unconfigured
inboxes copied using generic tools like cp(1) or rsync(1).
|
|
It's annoying for people using "git fetch && public-inbox-index"
as one user; and running -httpd/-nntpd as a different user
(where users see different config files).
|
|
* v2-idx-progress:
v2writable: show progress updates for index_sync
index: support --verbose option
v2writable: move index_sync options to sync state
v2writable: use prototypes for internal subs
v2writable: localize unindex-range.$EPOCH to $sync state
v2writable: move {ranges} into $sync state
v2writable: move {regen} into $sync state
v2writable: move {reindex} field to $sync state
v2writable: sync: move delete markers into $sync state
v2writable: introduce $sync state and put mm_tmp in it
|
|
We already "use warnings" everywhere, but could miss some spots.
This ought to cover that, and usually Perl module authors are
consistent about avoiding warnings that we won't clutter our
test outputs.
|
|
This can useful for limiting test resource use without relying
on remembering the variable command-line.
|
|
We can show progress whenever we commit changes to the FS.
|
|
It doesn't implement progress of batches, yet, but it wires
up the parsing of the command-line while preserving output
compatibility.
This output is NOT meant to be stable.
|
|
And use singular `opt' to be consistent with the common name
of 'getopt'.
|
|
Hopefully this improves maintainability by allowing Perl
to do some arg checking for us.
|
|
We don't need to stuff that into $self (V2Writable) which can be
longer-lived than a ->index_sync invocation.
|
|
Yet another temporary variable with no use outside of index_sync.
|
|
regen is always enabled for index_sync nowadays (and has
been for a while).
Rename `index_prepare' to `sync_prepare' to show it's for
->index_sync; and not the online indexing we do for ->add.
|
|
reindexing info is not used outside of the index_sync code path.
|
|
Another small step to reduce parameters passed to reindex_oid.
|
|
A first step towards making the v2 index_sync code
easier-to-follow. More fields to follow...
|
|
`public-inbox-index --reindex' could cause NNTP article number
gaps to form when it also has to deal with new,
never-before-seen commits in mirrors running off `git fetch'.
Fix this by running two distinct invocations of ->index_sync;
once to only reindex old commits, and a second time to index
new commits.
This does not appear to be a problem on v1 at the moment,
but I'll need more time to analyze this.
|
|
We can't pass an empty string to `git merge-base --is-ancestor'
AFAIK, this did NOT present issues in the current test suite.
|
|
It did not cause a test failure because the default fallback
is `indexlevel=full'
|
|
Streaming large blobs can take multiple iterations of the event
loop in our -httpd; so we must not let the File::Temp::Dir
result go out-of-scope when streaming large blobs created from
patches.
|
|
Fix a misspelling and ensure line context is printed by
`die' by leaving out the final '\n'. Also, `delete' was
pointless.
|
|
No reason to copyright colour schemes :P
|
|
Apparently it's never been used and we write to msgmap directly.
|
|
Don't hard-code "basic", since we already ran -init with the
intended indexlevel.
|
|
I have never not found double negatives to be confusing...
|
|
git bundles could/should make self-hosting easier.
Being able to configure synonym (and spelling) lists would make
some searches more useful.
Might as well dogfood kernel stuff, too, given the overlap and
history between this project, git and the Linux kernel. Would
be interesting to have *BSD folks throw their hat in the ring,
too.
Building/testing userspace stuff is often the most
time-consuming, but necessary to ensure future compatibility.
|
|
Oops :x
|
|
Since we go through the effort of hosting these manpages,
link to them.
|
|
In particular, the '--compact' switch is really useful since it
works without holding the inbox-wide lock for minutes at a time
on giant inboxes (inboxes where copies can take dozens, if not
hundreds of minutes).
|
|
They're nowhere to be found on Xapian.org, and links to
external services are either too long (for manpages.debian.org)
or have privacy-invasive tracking JS on them.
|
|
Otherwise timestamps for .html files get screwed up, too;
and that hurts caching.
|
|
It's not critical, but it's nice to have for cache-friendliness
(otherwise I would not have written it :P)
I guess I should follow up on getting it into 'git contrib/':
https://public-inbox.org/git/20100702033709.GA6818@burratino/
|
|
The nginx manpage is in section 8.
|
|
Oops :x
|
|
Some users (or bots :P) can trigger horrible queries which
the caller can choose to either log or ignore. This prevents
horrible queries from ExtMsg from logging confusing "ref: "
messages when $@ is not a Perl reference.
|
|
-index documentation avoid redundant v1 information and refers
readers to apropriate v1/v2 manpages. Search::Xapian can also
be optional, now, as only the PSGI search interface uses it.
Favor "INBOX_DIR" where appropriate, since "REPO_DIR" can be
confused for code repos which we also support.
XAPIAN_FLUSH_THRESHOLD is documented for all relevant
bulk commands.
|
|
To properly handle compact tmpdir cleanup in single process
situations, we need to carefully account for Xtmpdir not
being a singleton and ensuring we don't clobber signal
handlers which belong to other Xtmpdirs.
|
|
We don't have to be tied to the number of partitions in case
we made a bad choice at initialization. This doesn't affect
reindexing, but the copying phase is already intensive.
And optimize away the extra process when we only have a single
job which won't parallelize.
The wording for the (v2) reindexing phase could be improved,
later. I also plan to allow repartitioning of existing
Xapian DBs.
|
|
We should not have leftover junk on interrupted invocations.
|
|
Allow users to specify the --blocksize <B>, --no-full, --fuller
options for xapian-compact(1) for fine-tuning compact behavior
for low-traffic/inactive inboxes.
We also won't support --multipass, since it doesn't seem
compatible with our requirement to use --no-renumber.
We also won't support --single-file, since it only seems
intended for totally dead inboxes; and it doesn't seem
worth the support overhead when "totally dead" turns out
to be a misdiagnosis.
|
|
Since -xcpdb is a superset of -compact, we can reuse much of
that code used for driving compact.
For compact (only), this is slightly less memory efficient since
it requires an extra process per-partition, but we get to prefix
the output with the partition name for more readable output.
|