Eric Wong [Thu, 23 May 2019 09:36:49 +0000 (09:36 +0000)]
xcpdb: implement using Perl bindings
By avoid copydatabase(1) entirely, we can make further changes
to avoid locking the entire inbox for a long operation and
switch to fine-grained locking.
Eric Wong [Thu, 23 May 2019 09:36:47 +0000 (09:36 +0000)]
xapcmd: do not cleanup on errors
We move the old directory into the new directory, so avoid the
situation where a bug or error could cause the tempdir cleanup to run
and destroy both our old and new directories.
Eric Wong [Thu, 23 May 2019 09:36:46 +0000 (09:36 +0000)]
xcpdb: new tool which wraps Xapian's copydatabase(1)
copydatabase(1) is an existing Xapian tool which is the
recommended way to upgrade existing DBs to the latest Xapian
database format (currently "glass" for stable/released
versions). Our use of Xapian relies on preserving document IDs,
so we'll wrap it like we do xapian-compact(1) and use the
"--no-renumber" switch.
I could not name the tool "public-inbox-copydatabase" since it
would be ambiguous as to which DB it's actually copying. So, I
abbreviated the suffix to "xcpdb" (Xapian CoPy DataBase), which
I hope is acceptable and unambiguous.
Eric Wong [Thu, 23 May 2019 09:36:45 +0000 (09:36 +0000)]
xapcmd: support spawn options
copydatabase(1) is exceptionally noisy and it's output is
confusing when run in parallel. Support redirects at least, and
env while we're at it to give us future options.
We can also stuff a -jobs parameter into the options to limit
parallelism since it can be useful for low-priority upgrade
jobs.
Eric Wong [Thu, 23 May 2019 09:36:40 +0000 (09:36 +0000)]
v1writable: retire in favor of InboxWritable
In retrospect, introducing V1Writable was unnecessary and
InboxWritable->importer is in a better position to abstract
away differences between v1 and v2 writers.
So teach InboxWritable to initialize inboxes and get rid
of V1Writable.
Eric Wong [Wed, 22 May 2019 21:54:21 +0000 (21:54 +0000)]
Merge branch 'ds-cleanup'
* ds-cleanup:
DS: warn on deprecations
DS: remove IPPROTO_TCP import
DS: drop $VERSION var
DS: remove support OtherFds code
DS: get rid of unused methods and aliases
Eric Wong [Wed, 22 May 2019 21:46:56 +0000 (21:46 +0000)]
DS: remove support OtherFds code
It's easy enough to wrap FDs in classes that can use
all of the functionality of the event loop, not just
the read-only interface AddOtherFds provided.
Eric Wong [Tue, 21 May 2019 23:22:05 +0000 (23:22 +0000)]
Merge remote-tracking branch 'origin/xap-optional' into master
* origin/xap-optional:
admin: improve warnings and errors for missing modules
searchidx: do not create empty Xapian partitions for basic
lazy load Xapian and make it optional for v2
www: use Inbox->over where appropriate
nntp: use Inbox->over directly
inbox: add ->over method to ease access
Eric Wong [Thu, 16 May 2019 01:23:29 +0000 (01:23 +0000)]
view: perform highlighting for space-prefixed diffs
"git format-patch --interdiff" and similar can prefix diffs
with leading white space. Teach our diff parser to account
for it and set appropriate CSS classes for them.
Eric Wong [Wed, 15 May 2019 01:18:07 +0000 (01:18 +0000)]
www: use Inbox->over where appropriate
We don't need to rely on Xapian search functionality for the
majority of the WWW code, even. subject_normalized is moved to
SearchMsg, where it (probably) makes more sense, anyways.
Eric Wong [Tue, 14 May 2019 03:32:00 +0000 (03:32 +0000)]
tests: remove unnecessary loading of ::DS and Socket
PublicInbox::DS works for every platform we we care about,
nowadays; so checking for it is a waste of time. Cleanup a
few POSIX and Socket imports while we're in the area.
Eric Wong [Tue, 14 May 2019 02:04:43 +0000 (02:04 +0000)]
searchidx: fix incremental index with indexlevel=basic on v1
We were reindexing the full history every invocation of -index
when Xapian was not used because we were incorrectly relying on
'last_commit' metadata stored in Xapian.
Rewrite the indexing logic to be less confusing while we're
at it, since we rely on `git merge-base --is-ancestor' nowadays.
Furthermore, we need to handle message removals from the
overview index correctly when Xapian is not in use.
Co-authored-by: Eric W. Biederman <ebiederm@xmission.com>
Eric Wong [Tue, 14 May 2019 02:04:41 +0000 (02:04 +0000)]
v1writable: new wrapper which is closer to v2writable
Import initialization is a little strange from history, but we
also can't change it too much because it's technically a public
API which external code may rely on...
And we may need to support v1 repos indefinitely. This should
make it easier to write tests for both formats.
Eric Wong [Tue, 7 May 2019 09:09:54 +0000 (09:09 +0000)]
ci: include some scripts to make CI easier
This should make it easier to test a bunch of package
installation profiles across whatever OS isolation
one chooses (chroots, containers, jails, VMs).
Eric Wong [Wed, 8 May 2019 08:34:36 +0000 (08:34 +0000)]
doc/include.mk: remove git use and redundant declarations
And document that we still have GNU-isms in that
include.mk Makefile (and may continue to do so).
Finally, take advantage of GNU-isms to warn users
to run "gmake" to build all manpages.
Eric Wong [Wed, 8 May 2019 02:47:20 +0000 (02:47 +0000)]
INSTALL: depend on Encode instead of Encode::MIME::Header
Email::MIME uses Encode::MIME::Header and depends on that
appropriately; however we depend on other parts of the Encode
distribution, but that's bundled with Perl by upstream, anyways;
and should place no additional burden on users.
Eric Wong [Sun, 5 May 2019 01:37:00 +0000 (01:37 +0000)]
INSTALL: update with FreeBSD pkg names and reorder
Fortunately, there is a pattern to most of these package names
in all distros I've tested (and hopefully other BSDs have them,
too).
Then, reorder the INSTALL document to mention the top-level
modules, first, so users can avoid needing to type extra
dependencies. However, we still list some implicit dependencies
in case the upstream package drops dependencies independently of
us.
Finally, Devel::Peek is not a dependency worth making optional
since it's bundled by Perl upstream. Fedora/RH-based distros
are the only one which turn it into a non-standard package when
Perl5 is installed.
* origin/danga-bundle:
DS: epoll: fix misordered EPOLL_CTL_DEL call
DS: drop unused "_undef" sub
syscall: drop readahead wrapper
build: do not manify DS and Syscall pods
DS: handle EINTR in IO::Poll path, too
DS: workaround IO::Kqueue EINTR (mis-)handling
DS: drop profiling support
DS: remove unused fields and functions
listener: use EPOLLEXCLUSIVE for listen sockets
bundle Danga::Socket and Sys::Syscall
Eric Wong [Wed, 8 May 2019 19:04:17 +0000 (19:04 +0000)]
DS: epoll: fix misordered EPOLL_CTL_DEL call
Any operations on an fd after POSIX::close() are invalid, so
epoll_ctl will fail. Worse off, in a multi-threaded Perl, the
fd may be reused by another thread and EPOLL_CTL_DEL can hit the
wrong file description as a result.
Eric Wong [Mon, 6 May 2019 23:42:31 +0000 (23:42 +0000)]
spawn (Inline::C): fix off-by-one error
Noticed while testing on FreeBSD 11.2 amd64 with the optional
Inline::C extension using clang 6.0.0. The end result on
FreeBSD was spawning processes failed badly and things were
immediately unusable with this enabled.
av_len is a misleading API, and I failed to read the API
comments in perl:/av.c which state:
> Note that, unlike what the name implies, it returns
> the highest index in the array, so to get the size of
> the array you need to use "av_len(av) + 1".
> This is unlike "sv_len", which returns what you would expect.
If this bug affected anybody, it would've only affected users
using both the optional Inline::C module AND set the
PERL_INLINE_DIRECTORY environment variable.
That said, I've never seen any evidence of it on Debian
GNU/Linux + gcc on any x86 variant. That includes full 64-bit
systems, a full 32-bit system, a 64-bit system with 32-bit
userspace, across multiple gcc versions since 2016.
Eric Wong [Fri, 3 May 2019 10:34:10 +0000 (10:34 +0000)]
DS: remove unused fields and functions
More will likely be dropped in the future, but drop the obvious
ones we aren't using, for now; especially since some of them are
set at ->new time and unavoidable.
This saves 579 bytes per-client on my 64-bit Debian stable
system as measured by Devel::Size::total_size from
PublicInbox::HTTP::event_read. This adds up in C10K or C100K
situations.
Things we drop are:
* corked - MSG_MORE requires fewer syscalls
* read_push_back - tried to use it, ate CPU with slow clients
* IP/port fields - accept() already returns what we care about
Eric Wong [Fri, 3 May 2019 10:34:09 +0000 (10:34 +0000)]
listener: use EPOLLEXCLUSIVE for listen sockets
Since our listen sockets are non-blocking and we may run
multiple httpd|nntpd processes; we need a way to avoid
thundering herds when there are multiple httpd|nntpd worker
processes.
EPOLLEXCLUSIVE was added just for that in Linux 4.5
Eric Wong [Fri, 3 May 2019 10:34:08 +0000 (10:34 +0000)]
bundle Danga::Socket and Sys::Syscall
These modules are unmaintained upstream at the moment, but I'll
be able to help with the intended maintainer once/if CPAN
ownership is transferred. OTOH, we've been waiting for that
transfer for several years, now...
Changes I intend to make:
* EPOLLEXCLUSIVE for Linux
* remove unused fields wasting memory
* kqueue bugfixes e.g. https://rt.cpan.org/Ticket/Display.html?id=116615
* accept4 support
And some lower priority experiments:
* switch to EV_ONESHOT / EPOLLONESHOT (incompatible changes)
* nginx-style buffering to tmpfile instead of string array
* sendfile off tmpfile buffers
* io_uring maybe?
Eric Wong [Sat, 27 Apr 2019 21:46:54 +0000 (21:46 +0000)]
extmsg: escape ampersands in @EXT_URL array
We already escape the user-provided Message-IDs (so there's no
security problem AFAIK), but the URL templates which exist in
our source code were not escaped properly.
Eric Wong [Fri, 26 Apr 2019 07:42:24 +0000 (07:42 +0000)]
viewdiff: do not break out of DSTATE_CTX on /^$/
It seems a common case for mangled patches is editors or MUAs
dropping trailing whitespace, and lines matching /^ $/ gets
the space dropped to only match /^$/.
Eric Wong [Thu, 25 Apr 2019 07:57:06 +0000 (07:57 +0000)]
cgit: improve handling of cgit data path
Document `publicinbox.cgitdata' config directive, but allow it
to be unspecified and/or missing for installations which do not
wish to serve static data at all.
For users installing cgit from source to their home directory,
we can usually infer the cgit data path based on the cgit.cgi
binary path, even.
Eric Wong [Wed, 24 Apr 2019 19:52:14 +0000 (19:52 +0000)]
wwwhighlight: do linkification, too
Otherwise, there's no reason to use this API over highlight(1).
Maybe this can be an option in the future; but I'm struggling to
find a reason to not do it by default.
Eric Wong [Thu, 18 Apr 2019 08:25:56 +0000 (08:25 +0000)]
www: support listing of inboxes
We will still return a 404 by default to '/' for compatibility
with users of Plack::App::Cascade or similar. Inboxes are
sorted by modification times to help users detect activity
(similar to the /$INBOX/ topic view).
New configuration options:
* publicinbox.wwwlisting - configure the listing type
* publicinbox.<name>.hide - hide a particular inbox from the listing
See changes to public-inbox-config.pod for full descriptions
of the new options.
Eric Wong [Wed, 17 Apr 2019 22:05:38 +0000 (22:05 +0000)]
start depending on Perl 5.10.1+
I mainly want to start using the '//' (defined-or) operator to
simplify code, and Perl 5.10.1 is roughly a decade old at this
point.
"given/when" would've be nice, but it's future is in doubt AFAIK.
I also started using the 'parent' module in WwwHighlight, and
'autodie' in UserContent.pm, both of which were only distributed
with Perl since 5.10.1; and testing with ancient
versions/distros is time-consuming.
Anyways, I think this a small-enough jump to not break any
existing installations, given we already depend on fairly
recent versions of git and Xapian.
Maybe we can use more newish Perl features in the future...
Eric Wong [Thu, 18 Apr 2019 00:58:06 +0000 (00:58 +0000)]
inbox: add `modified' sub
For inboxes with SQLite enabled (all v2, and probably most v1);
we can use the overview DB to get the timestamp of the latest
message. It's faster than scanning git branches for commit times,
but not always the same.
Eric Wong [Wed, 17 Apr 2019 10:49:15 +0000 (10:49 +0000)]
git: calculate modified time of repository
This will be used for generating an HTML listing for v1 inboxes,
at least. The logic for this follows that of grokmirror,
and we may dynamically generate manifest.js.gz natively...
Eric Wong [Mon, 15 Apr 2019 07:32:49 +0000 (07:32 +0000)]
HACKING: stop advertising mailing list subscription info
It's still available, but no point in advertising something
which increases the dependency on a centralized subscriber list.
In fact, it's actively against the goal of promoting resistance
against centralization.
Eric Wong [Tue, 16 Apr 2019 01:10:09 +0000 (01:10 +0000)]
cleanup: use '$ibx' consistently when referring to Inbox refs
'$inbox' is more human-readable, so that is for the more
human-readable name in most cases. Making our variable naming
more consistent should make the code easier-to-review and
harder to screw up.