]> Sergey Matveev's repositories - public-inbox.git/log
public-inbox.git
4 years agoMakefile.PL: add test scripts to syntax checks
Eric Wong [Fri, 14 Jun 2019 16:47:44 +0000 (16:47 +0000)]
Makefile.PL: add test scripts to syntax checks

I make syntax errors all the time :x

4 years agot/www_listing: favor HTTP::Tiny over Net::HTTP
Eric Wong [Fri, 14 Jun 2019 16:42:26 +0000 (16:42 +0000)]
t/www_listing: favor HTTP::Tiny over Net::HTTP

More testers are likely to have HTTP::Tiny than Net::HTTP, since
HTTP::Tiny is a dual-life module and distributed with Perl since
Perl 5.14 (2011-05-14), whereas Net::HTTP will likely live in
a separate package forever.

4 years agoedit: fix portability of editor invocation
Eric Wong [Mon, 10 Jun 2019 23:53:46 +0000 (23:53 +0000)]
edit: fix portability of editor invocation

The eval was unnecessary, and $0 can't be "--".
Tested with /bin/sh on FreeBSD 11.2

4 years agoMerge remote-tracking branch 'origin/reshard' into next
Eric Wong [Fri, 14 Jun 2019 16:25:39 +0000 (16:25 +0000)]
Merge remote-tracking branch 'origin/reshard' into next

* origin/reshard:
  xcpdb: support resharding v2 repos
  xcpdb: use destination shard as progress prefix
  xapcmd: preserve indexlevel based on the destination
  v2writable: use a smaller default for Xapian partitions

4 years agoMerge remote-tracking branch 'origin/manifest' into next
Eric Wong [Fri, 14 Jun 2019 16:23:13 +0000 (16:23 +0000)]
Merge remote-tracking branch 'origin/manifest' into next

* origin/manifest:
  git: ensure ->modified returns an integer
  www: support $INBOX/git/$EPOCH.git for v2 cloning
  www: wire up /$INBOX/manifest.js.gz, too
  wwwlisting: generate grokmirror-compatible manifest.js.gz
  wwwlisting: allow hiding entries from manifest

4 years agoMerge remote-tracking branch 'origin/edit' into next
Eric Wong [Fri, 14 Jun 2019 16:23:00 +0000 (16:23 +0000)]
Merge remote-tracking branch 'origin/edit' into next

* origin/edit:
  edit: unlink temporary file when done
  v2writable: replace: kill git processes before reindexing
  edit: drop unwanted headers before noop check
  edit|purge: improve output on rewrites
  edit: new tool to perform edits
  doc: document the --prune option for -index
  admin: expose ->config
  AdminEdit: move editability checks from -purge
  admin: beef up resolve_inboxes to handle purge options
  purge: start moving common options to AdminEdit module
  admin: remove warning arg for unconfigured inboxes
  v2writable: implement ->replace call
  import: switch to "replace_oids" interface for purge
  import: extract_author_info becomes extract_commit_info
  v2writable: consolidate overview and indexing call

4 years agoxcpdb: support resharding v2 repos
Eric Wong [Thu, 13 Jun 2019 00:29:37 +0000 (00:29 +0000)]
xcpdb: support resharding v2 repos

v2 repos are sometimes created on machines where CPU
parallelization exceeds the capability of the storage devices.

In that case, users may reshard the Xapian DB to any smaller,
positive integer to avoid excessive overhead and contention when
bottlenecked by slow storage.

Resharding can also be used to increase shard count after
hardware upgrades.

4 years agoxcpdb: use destination shard as progress prefix
Eric Wong [Wed, 12 Jun 2019 08:10:04 +0000 (08:10 +0000)]
xcpdb: use destination shard as progress prefix

For M:N resharding, we'll want to display the number from
the new shard number.

4 years agoxapcmd: preserve indexlevel based on the destination
Eric Wong [Wed, 12 Jun 2019 08:05:33 +0000 (08:05 +0000)]
xapcmd: preserve indexlevel based on the destination

To support M:N resharding, we need to ensure we store the
indexlevel in the destination shard, rather than the
originating one.

4 years agov2writable: use a smaller default for Xapian partitions
Eric Wong [Wed, 12 Jun 2019 00:35:32 +0000 (00:35 +0000)]
v2writable: use a smaller default for Xapian partitions

Apparently 16 CPUs (probably HT) and SATA storage is common
these days.  Having excessive Xapian partitions leads to
contention and excessive FD/space use.  So set a smaller
default but continue allowing user-specified values to bump
this up.

4 years agogit: remove cat_file sub callback interface
Eric Wong [Thu, 13 Jun 2019 08:10:02 +0000 (08:10 +0000)]
git: remove cat_file sub callback interface

We weren't using it, and in retrospect, it makes no sense to use
this API cat_file for giant responses which can't read quickly
with minimal context-switching (or sanely fit into memory for
Email::Simple/Email::MIME).

For giant blobs which we don't want slurped in memory, we'll
spawn a short-lived git-cat-file process like we do in ViewVCS.

Otherwise, monopolizing a git-cat-file process for a giant
blob is harmful to other PSGI/NNTP users.

A better interface is coming which will be more suitable for
for batch processing of "small" objects such as commits and
email blobs.

4 years agonntp: filter out duplicate Message-IDs for leafnode
Eric Wong [Fri, 14 Jun 2019 00:27:31 +0000 (00:27 +0000)]
nntp: filter out duplicate Message-IDs for leafnode

It's the unfortunate reality that there are some clients which
reuse Message-IDs (in which we generate + use another) or set
multiple Message-IDs on their own.  While the v2 format
addresses that, NNTP clients such as leafnode are not always
prepared to deal with that case.

So, ensure NNTP clients only see a single Message-ID, and
show the others as 'X-Alt-Message-ID'.

4 years agonntp: ensure Message-ID is not folded for leafnode
Eric Wong [Thu, 13 Jun 2019 20:46:52 +0000 (20:46 +0000)]
nntp: ensure Message-ID is not folded for leafnode

Leafnode cannot handle Message-ID headers which are too long and
require folding via Email::Simple::Header.  Since there are
already many of these messages in git with the header already
folded, we need to handle the unfolding when emitting the
message via NNTP.

As far as we know, Leafnode is the only client software
incapable of handling this case.

4 years agonntp: add Path: header for leafnode
Eric Wong [Thu, 13 Jun 2019 06:54:13 +0000 (06:54 +0000)]
nntp: add Path: header for leafnode

Apparently leafnode just needs any junk in the Path: header.
Lets not waste bandwidth and just use a single byte to keep
leafnode happy.

Cc: Dave Taht <dave@taht.net>
4 years agoMakefile.PL: "dsyn" target to check syntax of changed files
Eric Wong [Thu, 13 Jun 2019 03:44:41 +0000 (03:44 +0000)]
Makefile.PL: "dsyn" target to check syntax of changed files

We have lots of files and syntax-checking every single one of
them is slow.  Enable "perl -w" in the existing syntax check
while we're at it.

4 years agodoc: update dependencies for CentOS 7.x
Eric Wong [Thu, 13 Jun 2019 02:26:30 +0000 (02:26 +0000)]
doc: update dependencies for CentOS 7.x

Digest::SHA is the most notable missing package at runtime
for a minimal system.

Tests don't run at all without Test::Simple (or Test::More).
Plack::Test is also a separate package, too...

Also, the package for IO::Compress::Gzip should be IO::Compress;
as perl-PerlIO-gzip is a different thing entirely which is not
relevant to our needs.

Test::HTTP::Server::Simple doesn't seem required at all for Plack
tests.

ExtUtils::MakeMaker needs to be documented as a install dependency
for people installing this, too; since AFAIK public-inbox is not
yet in any distros.

4 years agot/common.perl: fix error message for git requirements
Eric Wong [Thu, 13 Jun 2019 01:53:18 +0000 (01:53 +0000)]
t/common.perl: fix error message for git requirements

And enable strict + warnings in the scope of t/common.perl, too.

4 years agosearchmsg: remove unused ->get subroutine
Eric Wong [Thu, 13 Jun 2019 01:49:11 +0000 (01:49 +0000)]
searchmsg: remove unused ->get subroutine

It's obsolete and unusable since our search schema version 15;
which made the Xapian document ID correspond to the NNTP article
number.

4 years agosearchidx: improve error message when Xapian fails
Eric Wong [Wed, 12 Jun 2019 00:18:01 +0000 (00:18 +0000)]
searchidx: improve error message when Xapian fails

Make it easier to detect if a partition is corrupt.

4 years agoedit: unlink temporary file when done
Eric Wong (Contractor, The Linux Foundation) [Tue, 11 Jun 2019 18:34:58 +0000 (18:34 +0000)]
edit: unlink temporary file when done

We don't need to leave temporary files lying around.

4 years agov2writable: replace: kill git processes before reindexing
Eric Wong (Contractor, The Linux Foundation) [Mon, 10 Jun 2019 21:43:52 +0000 (21:43 +0000)]
v2writable: replace: kill git processes before reindexing

Xapian on Linux <3.15 has trouble with coprocesses since it used
fork() for locking and would hold onto pipes used for git
unnecessarily.

4 years agoedit: drop unwanted headers before noop check
Eric Wong (Contractor, The Linux Foundation) [Mon, 10 Jun 2019 18:09:27 +0000 (18:09 +0000)]
edit: drop unwanted headers before noop check

mutt will set Content-Length, Lines, and Status headers
unconditionally, so we need to account for that before
doing header comparisons to avoid making expensive changes
when noop edits are made.

4 years agoedit|purge: improve output on rewrites
Eric Wong (Contractor, The Linux Foundation) [Mon, 10 Jun 2019 16:58:55 +0000 (16:58 +0000)]
edit|purge: improve output on rewrites

Fill in undef as "(unchanged)" when displaying commits
and prefix the epoch name.

4 years agogit: ensure ->modified returns an integer
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 00:53:29 +0000 (00:53 +0000)]
git: ensure ->modified returns an integer

We don't want to serialize timestamps as strings to JSON.
I only noticed this bug on a 32-bit system.

4 years agodoc: index.pod: fix mismatched =back
Eric Wong [Mon, 10 Jun 2019 02:44:07 +0000 (02:44 +0000)]
doc: index.pod: fix mismatched =back

4 years agodoc: generate manpage and HTML from -convert POD
Eric Wong [Sun, 9 Jun 2019 09:10:30 +0000 (09:10 +0000)]
doc: generate manpage and HTML from -convert POD

It's been written for over a year, but I forgot to include
it in the build so it did not get installed or put on the site.

4 years agodoc: clean-doc target removes *.8 manpages
Eric Wong [Sun, 9 Jun 2019 09:10:26 +0000 (09:10 +0000)]
doc: clean-doc target removes *.8 manpages

4 years agoedit: new tool to perform edits
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:47 +0000 (02:51 +0000)]
edit: new tool to perform edits

This wrapper around V2Writable->replace provides a user-interface
for editing messages as single-message mboxes (or the raw text
via $EDITOR).

4 years agodoc: document the --prune option for -index
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:46 +0000 (02:51 +0000)]
doc: document the --prune option for -index

We've had it around for a while, but I forgot to document it :x

4 years agoadmin: expose ->config
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:45 +0000 (02:51 +0000)]
admin: expose ->config

No point in forcing admin programs to reparse the config
themselves; and we won't support multiple instances of it;
unlike the WWW code.

4 years agoAdminEdit: move editability checks from -purge
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:44 +0000 (02:51 +0000)]
AdminEdit: move editability checks from -purge

We'll be reusing the same logic for -edit

4 years agoadmin: beef up resolve_inboxes to handle purge options
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:43 +0000 (02:51 +0000)]
admin: beef up resolve_inboxes to handle purge options

We'll be using this in -edit, and maybe other admin-oriented
tools for UI-consistency.

4 years agopurge: start moving common options to AdminEdit module
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:42 +0000 (02:51 +0000)]
purge: start moving common options to AdminEdit module

Editing and purging are similar operations involving history
rewrites, so there'll be common options and code between them.

4 years agoadmin: remove warning arg for unconfigured inboxes
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:41 +0000 (02:51 +0000)]
admin: remove warning arg for unconfigured inboxes

We no longer make -index warn on it, no other code uses it;
and working on unconfigured inboxes is totally reasonable
for admins who are setting things up.

4 years agov2writable: implement ->replace call
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:40 +0000 (02:51 +0000)]
v2writable: implement ->replace call

Much of the existing purge code is repurposed to a general
"replace" functionality.

->purge is simpler because it can just drop the information.
Unlike ->purge, ->replace needs to edit existing git commits (in
case of From: and Subject: headers) and reindex the modified
message.

We currently disallow editing of References:, In-Reply-To: and
Message-ID headers because it can cause bad side effects with
our threading (and our lack of rethreading support to deal with
excessive matching from incorrect/invalid References).

4 years agoimport: switch to "replace_oids" interface for purge
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:39 +0000 (02:51 +0000)]
import: switch to "replace_oids" interface for purge

Continuing the work by Eric Biederman in commit a118d58a402bd31b
("Import.pm: When purging replace a purged file with a zero length file"),
we can use a generic OID replacement mechanism to implement
purge.

4 years agoimport: extract_author_info becomes extract_commit_info
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:38 +0000 (02:51 +0000)]
import: extract_author_info becomes extract_commit_info

We will be reusing the same logic for extracting all
the authorship and commit title logic for edits; so
put it all into one sub.

4 years agov2writable: consolidate overview and indexing call
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 02:51:37 +0000 (02:51 +0000)]
v2writable: consolidate overview and indexing call

It's one ugly sub with lots of parameters, but it's better
than calling a bunch of ugly subs with lots of parameters;
as we'll be needing to call it again when reindexing for
message replacements.

4 years agowww: support $INBOX/git/$EPOCH.git for v2 cloning
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 04:31:05 +0000 (04:31 +0000)]
www: support $INBOX/git/$EPOCH.git for v2 cloning

And use it in manifest.js.

To ease maintaining mirrors with grokmirror(1), we can accept
a "git/" directory prefix before the epoch, and ".git" suffix
after the epoch number.

We maintain compatibility with "$INBOX/$EPOCH" cloning, of
course, and it's still easier-to-type on the command-line.

4 years agowww: wire up /$INBOX/manifest.js.gz, too
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 04:31:04 +0000 (04:31 +0000)]
www: wire up /$INBOX/manifest.js.gz, too

I can imagine myself just wanting to clone a single v2 inbox
and all its epochs without thinking about include/exclude
rules in a grokmirror config file.

4 years agowwwlisting: generate grokmirror-compatible manifest.js.gz
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 04:31:03 +0000 (04:31 +0000)]
wwwlisting: generate grokmirror-compatible manifest.js.gz

Support on-demand generation of "/manifest.js.gz" for inboxes.
By default, this matches inboxes with URLs matching the given
request hostname by default.

This makes it easier to create full mirrors of several inboxes
without needing to configure static file serving.

cf. https://git.kernel.org/pub/scm/utils/grokmirror/grokmirror.git

4 years agowwwlisting: allow hiding entries from manifest
Eric Wong (Contractor, The Linux Foundation) [Sun, 9 Jun 2019 04:31:02 +0000 (04:31 +0000)]
wwwlisting: allow hiding entries from manifest

Since we already have a mechanism for hiding repositories from
the WWW listing, we might as well support another one for hiding
repositories from the upcoming manifest.js.gz generation.

4 years agoscripts: add README to describe its purpose
Eric Wong [Wed, 5 Jun 2019 02:20:05 +0000 (02:20 +0000)]
scripts: add README to describe its purpose

Well, it could probably be moved to contrib...

4 years agoscripts: require ASCII digits in a few places
Eric Wong [Wed, 5 Jun 2019 02:01:29 +0000 (02:01 +0000)]
scripts: require ASCII digits in a few places

I haven't touched most these scripts in ages, but we might as well
purge \d usage from here, as well.

4 years agotighten up digit matches to ASCII for git output
Eric Wong [Wed, 5 Jun 2019 01:26:55 +0000 (01:26 +0000)]
tighten up digit matches to ASCII for git output

While I don't expect git to suddenly start spewing non-ASCII
digits in places I'd expect ASCII, this would make things easier
for future hackers and reviewers.

4 years agot: avoid "subtest" for Perl 5.10.1 compatibility
Eric Wong [Tue, 4 Jun 2019 11:29:39 +0000 (11:29 +0000)]
t: avoid "subtest" for Perl 5.10.1 compatibility

The version of Test::More from Perl 5.10.1 did not support
"subtest", and the earliest version which did is Perl 5.12.0

The good news is this gives me an excuse to parallelize
the indexlevels-mirror test by splitting it into two.
(it could be further split, even).

Update t/nntpd. to use PI_TEST_VERSION consistently while
we're at it.

4 years agoMerge branch 'charclass'
Eric Wong [Tue, 4 Jun 2019 10:38:20 +0000 (10:38 +0000)]
Merge branch 'charclass'

* charclass: (24 commits)
  www: require ASCII word characters for CSS filenames
  www: require ASCII range for mbox downloads
  githttpbackend: require ASCII in path
  require ASCII digits for local FS items
  www: require ASCII digit for git epoch
  solver|viewdiff: restrict digit matches to ASCII
  inbox: require ASCII digits for feedmax var
  filter/rubylang: require ASCII digit for mailcount
  msgtime: require ASCII digits for parsing dates
  searchview: do not allow non-ASCII offsets and limits
  githttpbackend: require Range:, Status: to be ASCII digits
  view: require YYYYmmDD(HHMMSS) timestamps to be ASCII
  newswww: only accept ASCII digits as article numbers
  config: do not accept non-ASCII digits in cgitrc params
  www: require ASCII filenames in git blob downloads
  www: only emit ASCII chars in attachment filenames
  wwwattach: only pass the charset through if ASCII
  wwwlisting: require ASCII digit for port number
  http: require SERVER_PORT to be ASCII digit
  feed: only accept ASCII digits for ref~$N
  ...

4 years agoMANIFEST: add examples/nginx_proxy
Eric Wong [Tue, 4 Jun 2019 10:37:52 +0000 (10:37 +0000)]
MANIFEST: add examples/nginx_proxy

4 years agoexamples: add sample nginx configuration
Ali Alnubani [Tue, 4 Jun 2019 09:22:27 +0000 (09:22 +0000)]
examples: add sample nginx configuration

The sample configuration can be used to proxy-pass requests
to public-inbox-httpd or to a standalone PSGI/Plack server.

4 years agowww: require ASCII word characters for CSS filenames
Eric Wong [Tue, 4 Jun 2019 10:19:34 +0000 (10:19 +0000)]
www: require ASCII word characters for CSS filenames

Allowing admins to set non-ASCII CSS filenames could
cause unnecessary problems for client and proxies.

4 years agowww: require ASCII range for mbox downloads
Eric Wong [Tue, 4 Jun 2019 09:05:51 +0000 (09:05 +0000)]
www: require ASCII range for mbox downloads

We do not support many mboxrd download range specifications at
the moment; but parsing non-ASCII characters isn't planned.
This makes no difference aside from being able to return 404
slightly earlier than we would've in the past.

4 years agogithttpbackend: require ASCII in path
Eric Wong [Tue, 4 Jun 2019 08:58:32 +0000 (08:58 +0000)]
githttpbackend: require ASCII in path

We mainly support git-upload-pack; and maybe somebody uses
git-receive-pack with this.  Perhaps other (experimental)
command names are acceptable.  But it's unlikely anybody will
want Unicode command names for git services.

4 years agorequire ASCII digits for local FS items
Eric Wong [Tue, 4 Jun 2019 08:40:34 +0000 (08:40 +0000)]
require ASCII digits for local FS items

In case some BOFH decides to randomly create directories
using non-ASCII digits all over the place.

4 years agowww: require ASCII digit for git epoch
Eric Wong [Tue, 4 Jun 2019 08:36:18 +0000 (08:36 +0000)]
www: require ASCII digit for git epoch

Don't inadvertantly serve git repos containing non-ASCII
digit characters.

4 years agosolver|viewdiff: restrict digit matches to ASCII
Eric Wong [Tue, 4 Jun 2019 08:30:55 +0000 (08:30 +0000)]
solver|viewdiff: restrict digit matches to ASCII

git would not generate non-ASCII digits to describe
hunk offsets, so don't waste more time than necessary
to make sense of non-ASCII digit chars for line offsets.

4 years agoinbox: require ASCII digits for feedmax var
Eric Wong [Tue, 4 Jun 2019 08:29:36 +0000 (08:29 +0000)]
inbox: require ASCII digits for feedmax var

Don't waste more cycles than necessary if somebody decides to
put non-ASCII digits in their ~/.public-inbox/config

4 years agofilter/rubylang: require ASCII digit for mailcount
Eric Wong [Tue, 4 Jun 2019 08:25:10 +0000 (08:25 +0000)]
filter/rubylang: require ASCII digit for mailcount

Unlikely to matter, but who knows...

4 years agomsgtime: require ASCII digits for parsing dates
Eric Wong [Tue, 4 Jun 2019 02:04:24 +0000 (02:04 +0000)]
msgtime: require ASCII digits for parsing dates

User input contains the darndest things.  Don't waste more time
than necessary trying to parse dates out of non-ASCII digits.

4 years agosearchview: do not allow non-ASCII offsets and limits
Eric Wong [Tue, 4 Jun 2019 02:04:33 +0000 (02:04 +0000)]
searchview: do not allow non-ASCII offsets and limits

Non-ASCII digits would be interpreted as zero when used as integers.

4 years agogithttpbackend: require Range:, Status: to be ASCII digits
Eric Wong [Tue, 4 Jun 2019 02:04:32 +0000 (02:04 +0000)]
githttpbackend: require Range:, Status: to be ASCII digits

Non-ASCII digits would be interpreted as a zeroes as integers.

While we're at it, ensure the Status: code is an ASCII digit,
too; though I would not expect git-http-backend(1) or cgit(1)
start spewing non-ASCII digits at us.

4 years agoview: require YYYYmmDD(HHMMSS) timestamps to be ASCII
Eric Wong [Tue, 4 Jun 2019 02:04:31 +0000 (02:04 +0000)]
view: require YYYYmmDD(HHMMSS) timestamps to be ASCII

Passing digits to `timegm' which it does not understand would
be a waste of time.

4 years agonewswww: only accept ASCII digits as article numbers
Eric Wong [Tue, 4 Jun 2019 02:04:30 +0000 (02:04 +0000)]
newswww: only accept ASCII digits as article numbers

Non-ASCII digits aren't specified in RFC3977 for article numbers;
so don't waste a trip to SQLite only to turn up empty.

4 years agoconfig: do not accept non-ASCII digits in cgitrc params
Eric Wong [Tue, 4 Jun 2019 02:04:29 +0000 (02:04 +0000)]
config: do not accept non-ASCII digits in cgitrc params

cgit uses atoi(3), and now we can retain compatibility.

4 years agowww: require ASCII filenames in git blob downloads
Eric Wong [Tue, 4 Jun 2019 09:02:01 +0000 (09:02 +0000)]
www: require ASCII filenames in git blob downloads

Our Hval::to_filename sub has always been strict about emitting
ASCII-only characters for ViewVCS "raw" links.

However, somebody could manually generate a filename with
non-ASCII words for somebody else to download (we have no
cheap and fast way of mapping filenames back to blobs for
validation).

4 years agowww: only emit ASCII chars in attachment filenames
Eric Wong [Tue, 4 Jun 2019 02:04:28 +0000 (02:04 +0000)]
www: only emit ASCII chars in attachment filenames

We don't want to emit funky URLs which can be lost in
translation or cause problems with non-Unicode-aware
clients.

Then, don't accept non-ASCII filenames in URLs, since
a manually-generated URL/filename in attachment downloads
could be used for Unicode homographs to confuse folks who
down the attachment.

4 years agowwwattach: only pass the charset through if ASCII
Eric Wong [Tue, 4 Jun 2019 02:04:27 +0000 (02:04 +0000)]
wwwattach: only pass the charset through if ASCII

AFAIK all names of charsets are ASCII, so passing non-ASCII
characters from emails to clients would probably confuse clients.

4 years agowwwlisting: require ASCII digit for port number
Eric Wong [Tue, 4 Jun 2019 08:32:27 +0000 (08:32 +0000)]
wwwlisting: require ASCII digit for port number

We only care about the hostname portion for matching,
so this change is probably inconsequential.

4 years agohttp: require SERVER_PORT to be ASCII digit
Eric Wong [Tue, 4 Jun 2019 02:04:26 +0000 (02:04 +0000)]
http: require SERVER_PORT to be ASCII digit

I'm not sure what middlewares care for for SERVER_PORT; but
allowing non-ASCII digits seems non-sensical, here.

4 years agofeed: only accept ASCII digits for ref~$N
Eric Wong [Tue, 4 Jun 2019 02:04:22 +0000 (02:04 +0000)]
feed: only accept ASCII digits for ref~$N

We don't want to waste cycles passing non-ASCII characters
to git.

4 years agomid: id_compress requires ASCII-clean words
Eric Wong [Tue, 4 Jun 2019 02:04:23 +0000 (02:04 +0000)]
mid: id_compress requires ASCII-clean words

Its result is used for HTML anchors and such.

4 years agonntp: ensure we only handle ASCII whitespace
Eric Wong [Tue, 4 Jun 2019 08:20:40 +0000 (08:20 +0000)]
nntp: ensure we only handle ASCII whitespace

RFC3977 does not have provisions for whitespace beyond ASCII
TAB, SP, CR and LF.  I doubt there's any NNTP clients broken
enough to be sending non-ASCII whitespace delimiters.

We're probably excessively liberal regarding TAB acceptance,
even; but it's probably too late to change at this point...

4 years agonntp: be explicit about ASCII digit matches
Eric Wong [Tue, 4 Jun 2019 02:04:25 +0000 (02:04 +0000)]
nntp: be explicit about ASCII digit matches

We aren't able to make sense of non-ASCII digits

cf. perlrecharclass(1) / "Digits" section

4 years agolinkify: support Internationalized Domain Names in URLs
Eric Wong [Tue, 4 Jun 2019 02:04:21 +0000 (02:04 +0000)]
linkify: support Internationalized Domain Names in URLs

The "\w" character class in Perl matches any word characters
in the Unicode database, not just ASCII characters.  So we
must be prepared for that and generate links to IDNs.

4 years agoMerge remote-tracking branch 'origin/git-cleanup'
Eric Wong [Mon, 3 Jun 2019 10:06:04 +0000 (10:06 +0000)]
Merge remote-tracking branch 'origin/git-cleanup'

* origin/git-cleanup:
  git: drop the deleted err_c file
  git: unconditional expiry

4 years agoMerge remote-tracking branch 'origin/ds'
Eric Wong [Mon, 3 Jun 2019 09:04:01 +0000 (09:04 +0000)]
Merge remote-tracking branch 'origin/ds'

* origin/ds:
  ds: remove PLCMap and per-socket PostLoopCallback
  ds: drop write_set_watch field
  ds: drop unused EVENT: label in epoll code path
  ds: drop checks for invalid descriptors
  ds: drop set_writer_func support
  ds: add a note about planned future changes
  ds: drop more unused subs

4 years agods: remove PLCMap and per-socket PostLoopCallback
Eric Wong [Mon, 3 Jun 2019 09:00:28 +0000 (09:00 +0000)]
ds: remove PLCMap and per-socket PostLoopCallback

We don't need and won't be needing per-socket PostLoopCallbacks.

4 years agot/psgi_search.t: require DBD::SQLite
Eric Wong [Mon, 3 Jun 2019 01:40:06 +0000 (01:40 +0000)]
t/psgi_search.t: require DBD::SQLite

In case we encounter an odd system which has Search::Xapian
but not DBD::SQLite.

4 years agods: drop write_set_watch field
Eric Wong [Sun, 2 Jun 2019 03:39:37 +0000 (03:39 +0000)]
ds: drop write_set_watch field

We never enable write watches ourselves for HTTP and NNTP,
and only enable the write watch with EvCleanup because it's
an "always on" watch.

4 years agods: drop unused EVENT: label in epoll code path
Eric Wong [Sun, 2 Jun 2019 03:39:36 +0000 (03:39 +0000)]
ds: drop unused EVENT: label in epoll code path

This was never used in Danga::Socket 1.61, either.

4 years agods: drop checks for invalid descriptors
Eric Wong [Sun, 2 Jun 2019 03:39:35 +0000 (03:39 +0000)]
ds: drop checks for invalid descriptors

I've used Danga::Socket for well over a decade in various
projects at this point and have never seen the need for it.

If such a bug ever happens; the process should fall over so
it gets fixed ASAP.

4 years agods: drop set_writer_func support
Eric Wong [Sun, 2 Jun 2019 03:39:34 +0000 (03:39 +0000)]
ds: drop set_writer_func support

This is not used by perlbal for OpenSSL support, either;
and it does not appear to be the right layer for doing
write translations anyways (IO::Socket::SSL uses `tie').

4 years agods: add a note about planned future changes
Eric Wong [Sun, 2 Jun 2019 03:39:33 +0000 (03:39 +0000)]
ds: add a note about planned future changes

Sometimes I get bored with the email part of this project and
need a distraction :P

4 years agods: drop more unused subs
Eric Wong [Sun, 2 Jun 2019 03:39:32 +0000 (03:39 +0000)]
ds: drop more unused subs

ToClose and HaveEpoll are of no use to us and I see no
future use for them, either.

4 years agods: fix and test for FD leaks with kqueue on ->Reset
Eric Wong [Sat, 1 Jun 2019 21:43:47 +0000 (21:43 +0000)]
ds: fix and test for FD leaks with kqueue on ->Reset

Even though we currently don't use it repeatedly, ->Reset
should close() kqueue FDs and not cause the process to run
out of descriptors.

Add a close-on-exec test while we're at it.

4 years agodoc/v2: note that Xapian is now optional in v2
Eric Wong [Sat, 1 Jun 2019 07:28:17 +0000 (07:28 +0000)]
doc/v2: note that Xapian is now optional in v2

This is true as of e220b8b2ee5cfd458167dc2c6c92726352c4c80e
("Merge remote-tracking branch 'origin/xap-optional' into master")

4 years agods: set close-on-exec flag on epoll descriptors
Eric Wong [Sat, 1 Jun 2019 00:23:22 +0000 (00:23 +0000)]
ds: set close-on-exec flag on epoll descriptors

We should not be leaking these FDs to git(1) processes,
in case git has a bug that causes it to access the wrong FD.

4 years agogit: drop the deleted err_c file
Eric Wong [Sat, 1 Jun 2019 03:27:55 +0000 (03:27 +0000)]
git: drop the deleted err_c file

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.

4 years agogit: unconditional expiry
Eric Wong [Sat, 1 Jun 2019 00:20:51 +0000 (00:20 +0000)]
git: unconditional expiry

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.

4 years agogit: drop the deleted err_c file
Eric Wong [Sat, 1 Jun 2019 03:27:55 +0000 (03:27 +0000)]
git: drop the deleted err_c file

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.

4 years agogit: unconditional expiry
Eric Wong [Sat, 1 Jun 2019 00:20:51 +0000 (00:20 +0000)]
git: unconditional expiry

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.

4 years agoTODO: add item for optional Cache::FastMmap
Eric Wong [Fri, 31 May 2019 16:58:46 +0000 (16:58 +0000)]
TODO: add item for optional Cache::FastMmap

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.

4 years agoviewdiff: avoid repeat variable expansion
Eric Wong [Fri, 31 May 2019 16:46:04 +0000 (16:46 +0000)]
viewdiff: avoid repeat variable expansion

This is worth a 1-2% speedup in t/perf-msgview.t rendering 2620
messages currently in https://public-inbox.org/meta/

4 years agoMerge remote-tracking branch 'origin/v2-noop-speedup'
Eric Wong [Thu, 30 May 2019 19:50:04 +0000 (19:50 +0000)]
Merge remote-tracking branch 'origin/v2-noop-speedup'

* 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

4 years agodoc/hosted: drop some links and clarify wording
Eric Wong [Thu, 30 May 2019 06:54:06 +0000 (06:54 +0000)]
doc/hosted: drop some links and clarify wording

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

4 years agov2writable: short-circuit is_ancestor check on equality
Eric Wong [Thu, 30 May 2019 06:35:05 +0000 (06:35 +0000)]
v2writable: short-circuit is_ancestor check on equality

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.

4 years agov2writable: avoid mm_tmp creation without regen
Eric Wong [Thu, 30 May 2019 03:59:40 +0000 (03:59 +0000)]
v2writable: avoid mm_tmp creation without regen

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/

4 years agov2writable: hoist out index_epoch sub
Eric Wong [Thu, 30 May 2019 03:59:39 +0000 (03:59 +0000)]
v2writable: hoist out index_epoch sub

This will make future changes easier-to-follow.

4 years agov2writable: split off unindex_range mapping
Eric Wong [Thu, 30 May 2019 03:59:38 +0000 (03:59 +0000)]
v2writable: split off unindex_range mapping

It'll make it easier to detect if we have anything to
unindex and run git-log on, at all.

4 years agosearchidx: store indexlevel=medium as metadata
Eric Wong [Wed, 29 May 2019 20:56:32 +0000 (20:56 +0000)]
searchidx: store indexlevel=medium as metadata

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).

4 years agoindex: remove warning on unconfigured inboxes
Eric Wong [Wed, 29 May 2019 20:56:31 +0000 (20:56 +0000)]
index: remove warning on unconfigured inboxes

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).