Eric Wong [Sun, 23 Feb 2020 22:47:15 +0000 (22:47 +0000)]
searchview: set obfuscation inbox properly
We never lookup `$ctx->{-obfuscate}' anywhere, as the
correct key is `$ctx->{-obfs_ibx}' since some of the
address obfuscation stuff is inbox-specific.
Note: some of the obfuscation stuff still needs tests,
but it's low-priority at the moment since I don't think
it's a good feature after all.
Eric Wong [Mon, 17 Feb 2020 11:00:48 +0000 (05:00 -0600)]
view: shorten life of MIME object for permalink
We don't need to hold onto the Email::MIME object across
multiple WwwResponse->getline calls, instead we can stuff
the rendered HTML of the first (and hopefully only) message
of the buffer into ctx->{-html_tip}.
Eric Wong [Sun, 16 Feb 2020 10:24:35 +0000 (10:24 +0000)]
doc: design_www: document solver endpoint
The blob regeneration (solving) part has been stable and
performant for over a year with no problems, even with web
crawlers constantly hitting it without needing rate limits.
All the other stuff is open to bikeshedding (as long as
my crappy hardware supports it :P)
Eric Wong [Sat, 15 Feb 2020 09:46:40 +0000 (09:46 +0000)]
view: remove last Hval->new caller
The object-oriented Hval API turned out to be less useful and
more clunky than I envisioned years ago, so get rid of it.
We'll no longer strip trailing whitespace from From: headers in
the HTML display, but I doubt anybody cares.
Eric Wong [Sat, 15 Feb 2020 09:46:39 +0000 (09:46 +0000)]
view: escape ampersand in Message-IDs
We need to escape ampersands (and some other characters for href
attributes), so introduce a `mid_href' sub to do just that.
'<', '>' and '"' were always escaped, so there's no risk of tag
or attribute injection, but creative Message-IDs could cause
confusion for some parsers and generate invalid URLs.
Start getting rid of the bloated, over-engineered OO Hval API
while we're at it, I only noticed this bug because I started
killing off Hval->new* callers.
Eric Wong [Sat, 15 Feb 2020 09:46:37 +0000 (09:46 +0000)]
view,searchview: avoid smsg method calls when using SQLite/Xapian
We already pre-populate the hashref when loading $smsg
(PublicInbox::SearchMsg) objects out of over.sqlite3 or Xapian,
so making expensive method calls isn't necessary in those cases.
We only need to use the method calls when SQLite or Xapian are
not available or are being populated (such as during indexing).
Eric Wong [Sat, 15 Feb 2020 09:46:36 +0000 (09:46 +0000)]
view: cleanup topic accumulation and dumping
Avoid needlessly normalizing the subject when dumping, since
it's pushed into the @$topic array during accumulation in
normalized form.
We can also safely treat $smsg as a hashref and avoid
calling "->ds" as a method since we know we've got that
loaded via Over||Search and won't have to use Email::MIME
header lookup methods.
Eric Wong [Fri, 7 Feb 2020 08:31:57 +0000 (02:31 -0600)]
doc: mark some TODO items as done
NNTP TLS and COMPRESS support and cgit spawning from
the WWW interface were implemented last year.
Given the lack of syscall number stability guarantee on the
OpenBSD and FreeBSD, I don't think supporting a pure-Perl kevent
is feasible. Inline::C may still be an option since IO::KQueue
is abandoned, though, as it is for some Linux-only syscalls and
maybe some POSIX ones not covered by POSIX.pm.
Eric Wong [Thu, 6 Feb 2020 20:59:45 +0000 (20:59 +0000)]
tests: switch to XML::TreePP for testing Atom feeds
XML::Feed pulls in a lot of dependencies, some of which XS.
That makes testing with blead or any non-OS-supplied Perl
installations more time consuming and more difficult because
of the need to have development headers and libraries for
libexpat1 or libxml2.
Performance from libexpat1 or libxml2 for our small tests cases
isn't relevant, either, and the pure Perl XML::TreePP seems up
to the task. It's also available in CentOS 7.x, FreeBSD 11.x,
and Debian, at least.
Eric Wong [Thu, 6 Feb 2020 08:49:36 +0000 (08:49 +0000)]
syscall: support Linux x32 ABI
The x32 ABI allows users to take advantage of the extra
registers on x86-64 without the bloat of 64-bit pointers and
longs.
This ought to be significant since Perl was designed when 32-bit
was prevalent; and the common structs for ops, hashes, scalars,
and arrays use longs (SSize_t/Size_t) for things which should
never need 64-bits when processing emails.
Debian's x32 port seems to work quite nicely under a chroot
on an amd64 Linux system. All tests pass under x32, now.
Eric Wong [Thu, 6 Feb 2020 04:03:47 +0000 (22:03 -0600)]
doc: remove .x/ subdirectory for Xapian manpages
There's no need to keep Xapian manpage renderings in a separate
subdirectory, after all.
Eliminating this difference between the local FS and URL path
will allow relative URLs to the Xapian manpages in our local
HTML documentation to work smoothly, since there was never any
".x/" path component for files served from public-inbox.org
Eric Wong [Tue, 4 Feb 2020 21:44:04 +0000 (21:44 +0000)]
doc: recommend -compact after --reindex
It's likely a user will be low on space after running --reindex,
so recommend the use of public-inbox-compact afterwards.
And add a few more notes about using public-inbox-compact to
clarify it's for inboxes-only (and not any old Xapian DBs) that
using xapian-compact(1) directly is error-prone and likely to
break things.
Eric Wong [Tue, 4 Feb 2020 04:44:23 +0000 (04:44 +0000)]
inbox: simplify ->description and ->cloneurl
We can use "//=" from Perl 5.10 to simplify the logic for these
methods. The use of chomp() in ->cloneurl was also unnecessary
since split(/\s+/s,...) already removes newlines.
Eric Wong [Tue, 4 Feb 2020 04:44:21 +0000 (04:44 +0000)]
www: stricter regexp for 405 errors
We want to match "GET" and "HEAD" exactly, not requests which
start with "GET" or end with "HEAD". This doesn't seem like
a real problem for public-inboxes which are actually public
data anyways.
Eric Wong [Tue, 4 Feb 2020 11:15:15 +0000 (11:15 +0000)]
doc: spellling fixes for manpages
The wording for publicinbox.nntpserver was awkward, too, and I
took this as opportunity to hopefully clarify it and favor
"hostname" for Internet addresses, because we already use
"address" to mean "email address" in the config.
Eric Wong [Sun, 2 Feb 2020 17:51:26 +0000 (17:51 +0000)]
spawn: actually die on (vfork|fork) failures
Commit 9f5a583694396f84 ("spawn (and thus popen_rd) die on failure")
was incomplete in that it only removed error checking for spawn
failures for non-(vfork|fork) calls, but the actual (vfork|fork)
PID result could still be undef.
Fixes: 9f5a583694396f84 ("spawn (and thus popen_rd) die on failure")
Eric Wong [Sun, 2 Feb 2020 06:52:18 +0000 (06:52 +0000)]
searchidxshard: rely on autoflush instead of ->flush
It reduces the number of ops and simplifies the code, slightly.
Add a missing IO::Handle import while we're at it, to be
explicit about which methods we use.
Eric Wong [Sun, 2 Feb 2020 06:52:15 +0000 (06:52 +0000)]
v2writable: nproc_shards: subtract 1 from given value
This is to be consistent with the `nproc(1)' code path. It also
quiets down a warning from Admin when "-j $JOBS" is specified,
since the master process (which distributes work to shards and
handles OverIdx and Msgmap) is considered a job on its own.
Eric Wong [Sat, 1 Feb 2020 09:12:13 +0000 (09:12 +0000)]
wwwtext: give "url" examples in sample config
inbox.$NAME.url is a common parameter and set by
public-inbox-init(1), so ensure we have lines for it and
emphasize it can be multi-value for .onion hidden services or
otherwise mirrored and available under multiple URLs.
Eric Wong [Thu, 30 Jan 2020 07:51:06 +0000 (07:51 +0000)]
convert: preserve highwater mark from v1 msgmap
If we're reusing the msgmap from a v1 inbox, we also need to
ensure the highwater mark doesn't get doubled in the v1->v2
conversion by internally triggering the equivalent of
"--reindex" on a fresh v2 inbox.
This was needed to convert an indexed v1 inbox which featured
messages with multiple Message-IDs in it. Fresh, unindexed
clones of v1 inboxes would not have been affected by this.
Eric Wong [Sun, 26 Jan 2020 23:41:02 +0000 (23:41 +0000)]
v2writable: newest epochs go first in alternates
New epochs are the most likely to have loose objects. git won't
be able to take advantage of pack indices and needs to scan
every alternate for the loose object via open/openat syscalls.
Those syscalls will add up some day when we've got hundreds or
thousands of epochs.
Eric Wong [Sun, 26 Jan 2020 10:29:24 +0000 (10:29 +0000)]
t/v2reindex.t: 5.10.1 glob compatibility
I'm not sure when `for (<"quoted string/glob/*">)' became
supported, and maybe it was inadvertant, but it fails
with Perl 5.10.1. Just use the glob() function to be
explicit.
Eric Wong [Sat, 25 Jan 2020 04:45:10 +0000 (04:45 +0000)]
viewdiff: rewrite and simplify
Instead of going line-by-line, use split() with a giant regexp
to capture groups of contiguous lines. This offloads state
management to the regexp itself and makes it FAR easier to
keep track of <span> and </span> pairings.
Performance seems roughly on par after this change for the
meta@public-inbox archives. It seems a tiny bit faster for
git@vger with xt/perf-msgview.t, likely due to the longer
messages and larger contiguous groups of lines having the same
prefix (or no prefix at all) and drastically reduces the number
of subroutine calls and Perl ops executed.
Eric Wong [Sat, 25 Jan 2020 04:45:08 +0000 (04:45 +0000)]
viewdiff: add "b=" param when missing "diff --git" line
<2841d2de-32ad-eae8-6039-9251a40bb00e@tngtech.com> as posted to
git@vger contained an otherwise valid diff without a "diff
--git" line. Generate a "b=" parameter in that case using the
"+++" line instead of the "diff --git" line. SearchIdx.pm no
longer uses the "diff --git" line for filename information,
either.
Eric Wong [Sat, 25 Jan 2020 04:45:07 +0000 (04:45 +0000)]
viewdiff: add "b=" param with non-standard diff prefix
<20180228012207.GB251290@aiede.svl.corp.google.com> (posted to
git@vger) uses "i" and "w" prefixes instead of the standard "a"
and "b" prefixes, ensure we emit a "b=$FILENAME" param for the
solver endpoint to improve search accuracy, syntax highlighting,
and information density in the URL itself.
Eric Wong [Sat, 25 Jan 2020 04:45:04 +0000 (04:45 +0000)]
linkify: move to_html over from ViewDiff
We use the same idiom in many places for doing two-step
linkification and HTML escaping. Get rid of an outdated
comment in flush_quote while we're at it.
Eric Wong [Sat, 25 Jan 2020 04:45:01 +0000 (04:45 +0000)]
xt/perf-msgview: switch to multipart_text_as_html
It's a more widely-used (but still internal) API which will
probably last longer than msg_html. It also reaches deeper into
the stack and avoids the overhead of ->getline via PSGI, so it's
faster and gives a more accurate measurement of lower-level parts.
Eric Wong [Sat, 25 Jan 2020 04:45:00 +0000 (04:45 +0000)]
tests: move the majority of t/view.t into t/plack.t
And some more into t/mid.t. PublicInbox::View::msg_html may
change internally, so lets rely on the stable PSGI interface
to test it, rather than a test which reaches deep into the
internals.
Eric Wong [Sat, 25 Jan 2020 04:44:56 +0000 (04:44 +0000)]
wwwstream: discard single-use $ctx fields after use
This should make it clear that we only use these elements
once and can discard them. While we're in the area, avoid
escaping '"' by using qq() instead of "" to quote strings
requiring interpolation.
Eric Wong [Sat, 25 Jan 2020 04:44:52 +0000 (04:44 +0000)]
searchview: keep $noop sub private to the package
It'll always be used as a callback, so there's no point in
giving it a name to be called non-anonymously. Making
assigments to it is slightly faster since there's no need
to repeatedly do a lookup by name.
Eric Wong [Sun, 26 Jan 2020 01:17:44 +0000 (01:17 +0000)]
xapcmd: increase scope of lock
The old lock scope was only sufficient for protecting against
concurrent modifications from the common -mda, -watch, or -learn
writers.
It was not sufficient for protecting against parallel -compact
or -xcpdb invocations from eager admins. Most of the time this
only leads to confusing and misleading warning messages, but
parallel xcpdb --reshard could lead to errors.
Eric Wong [Sun, 26 Jan 2020 01:17:43 +0000 (01:17 +0000)]
search: {version} => {ibx_ver}
We don't confuse human readers with the Xapian schema version.
We also want to make it obvious this is the version of the inbox
we're indexing, these are Search or SearchIdx objects, not Inbox
objects.
Eric Wong [Sat, 25 Jan 2020 20:57:57 +0000 (20:57 +0000)]
switch to sysseek + sysread for serving static files
The "perlio" layer doesn't do read(2) syscalls over 8192 bytes
at the moment, and binmode($fh, ':unix') leaks[1]. So use
sysseek and sysread for now, since I can't see retaining
compatibility with PerlIO::scalar being worth the trouble.
Eric Wong [Fri, 24 Jan 2020 09:43:52 +0000 (09:43 +0000)]
wwwstatic: wire up buffer bypass for -httpd
This prevents public-inbox-httpd from buffering ->getline
results from a static file into another temporary file when
writing to slow clients. Instead we inject the static file
ref with offsets and length directly into the {wbuf} queue.
It took me a while to decide to go this route, some
rejected ideas:
1. Using Plack::Util::set_io_path and having PublicInbox::HTTP
serve the result directly. This is compatible with what
some other PSGI servers do using sendfile. However, neither
Starman or Twiggy currently use sendfile for partial responses.
2. Parsing the Content-Range response header for offsets and
lengths to use with set_io_path for partial responses.
These rejected ideas required increasing the complexity of HTTP
response writing in PublicInbox::HTTP in the common, non-static
file cases. Instead, we made minor changes to the colder write
buffering path of PublicInbox::DS and leave the hot paths
untouched.
We still support generic PSGI servers via ->getline. However,
since we don't know the characteristics of other PSGI servers,
we no longer do a 64K initial read in an attempt to negotiate a
larger TCP window.
Eric Wong [Fri, 24 Jan 2020 09:43:51 +0000 (09:43 +0000)]
ds: tmpio: store offsets per-buffer
We want to be able to inject existing file handles + offsets and
even lengths into this in the future, without going through the
->getline interface[1]
We also switch to using a 64K buffer size since we can safely
discard whatever got truncated on write and full writes can help
negotiate a larger TCP window for high-latency, high-bandwidth
links.
While we're at it, make it obvious that we're using O_APPEND for
our tmpfile() interface so we can seek freely for reading while
the writer always prints to the end of the file.
[1] the getline interface for serving static files may result
in us buffering on-FS data into another temporary file,
which is a waste.
Eric Wong [Fri, 24 Jan 2020 09:43:50 +0000 (09:43 +0000)]
wwwstatic: offload error handling to PSGI server
The PSGI server needs to account for ->getline failing
due to disk failures or truncated files, anyways. So
just die() ourselves and let the PSGI server log and
drop the client.
Eric Wong [Fri, 24 Jan 2020 22:09:32 +0000 (22:09 +0000)]
website: omit technical/ and other subdirs
We don't need to clutter the website with unnecessary technical
information. Anybody who reads the technical/ directory should
be looking at our source code, anyways; and we also have cgit
and gitweb mirrors.
Eric Wong [Fri, 24 Jan 2020 22:09:29 +0000 (22:09 +0000)]
doc: avoid needless rebuilds of NEWS
Repeatedly rebuilding `NEWS' because the mtime of `NEWS'
is synched to the latest release .eml is a bit annoying,
but necessary to save bandwidth for the website.
So we'll also update the mtime of the source .eml file when
reading them. It's kinda gross to be setting mtimes of source
.eml files in Documentation/RelNotes/, but I can't think of
anything better at the moment...