Eric Wong [Fri, 5 Apr 2019 20:04:27 +0000 (20:04 +0000)]
config: support more cgit directives for project lists
Hopefully this gets us closer to matching cgit upstream behavior
(which also lacks tests). We'll still need to support macro
expansion at some point for compatibility...
Eric Wong [Fri, 5 Apr 2019 20:04:26 +0000 (20:04 +0000)]
cgit: serve static css, logo, favicon directly
We can reduce the configuration needed to run cgit by reusing
the static file handling logic of the dumb git HTTP protocol.
I hate logos and icons, so don't expect public-inbox.org or
80x24.org to ever have those to waste users' bandwidth with :P
But I expect other users to find this useful.
Eric Wong [Fri, 5 Apr 2019 20:04:25 +0000 (20:04 +0000)]
config: support cgit scan-path and scan-hidden-path
project_list support still needs to be done
And tests need to be written... :<
Kyle Meyer [Sun, 7 Apr 2019 03:31:40 +0000 (23:31 -0400)]
doc: fix a few minor typos
Eric Wong [Tue, 2 Apr 2019 02:29:44 +0000 (02:29 +0000)]
www: fix missing cgit fallback after legacy redirects
We need to instate our cgit handler everywhere we use NewsWWW
to catch wildcard requests which our normal endpoints do not
handle.
Eric Wong [Tue, 12 Mar 2019 01:47:59 +0000 (01:47 +0000)]
spawn: require soft and hard entries in RLIMIT_* handling
Our high-level config already treats single limits as a
soft==hard limit for limiters; so stop handling that redundant
in the low-level spawn() sub.
Eric Wong [Tue, 12 Mar 2019 01:18:30 +0000 (01:18 +0000)]
cgit: use a dedicated named limiter
I mainly need this to enforce RLIMIT_CPU (and RLIMIT_CORE)
when requests come which generate giant, unrealistic diffs.
Per-coderepo limiters may be added in the future. But for now,
I need to prevent cgit from monopolizing resources on my dinky
server.
Eric Wong [Mon, 11 Mar 2019 23:53:02 +0000 (23:53 +0000)]
qspawn: wire up RLIMIT_* handling to limiters
This allows users to configure RLIMIT_{CORE,CPU,DATA} using
our "limiter" config directive when spawning external processes.
Eric Wong [Mon, 11 Mar 2019 05:32:07 +0000 (05:32 +0000)]
www: wire up cgit as a 404 handler if cgitrc is configured
Requests intended for cgit are unlikely to conflict with
requests to inboxes. So we can safely hand those requests
off to cgit.cgi.
Eric Wong [Sun, 10 Mar 2019 22:02:29 +0000 (22:02 +0000)]
cgit: support running cgit as a standalone CGI
We depend on git-http-backend for smart HTTP clone support,
however; since cgit does not support smart clones natively.
WWW.pm will be able to cascade down to this as a 404 handler in
the future.
Eric Wong [Sun, 10 Mar 2019 21:40:00 +0000 (21:40 +0000)]
spawn: support absolute paths
cgit (and most other CGI executables) is not typically installed
for use via $PATH, so we'll need to support absolute paths to
run it.
Eric Wong [Sun, 10 Mar 2019 10:32:33 +0000 (10:32 +0000)]
githttpbackend: check for other errors and relax CRLF check
Reads to git-http-backend(1) could fail or EOF prematurely,
so we must be ready for that case.
Furthermore, cgit (and possibly other CGI) uses LF instead
of CRLF, so support those programs, too.
Eric Wong [Sun, 10 Mar 2019 07:51:45 +0000 (07:51 +0000)]
githttpbackend: move more psgi.input handling into subroutine
This will be useful for other CGI wrappers we make.
This also fixes a bug with some PSGI servers which did not
present a real IO::Handle in the psgi.input env field.
Eric Wong [Fri, 8 Mar 2019 01:21:22 +0000 (01:21 +0000)]
support publicinbox.cgitrc directive
We can save admins the trouble of declaring [coderepo "..."]
sections in the public-inbox config by parsing the cgitrc
directly.
Macro expansion (e.g. $HTTP_HOST) expansion is not supported,
yet; but may be in the future.
Eric Wong [Mon, 25 Feb 2019 05:14:10 +0000 (05:14 +0000)]
spawn: support RLIMIT_CPU, RLIMIT_DATA and RLIMIT_CORE
We'll be spawning cgit and git-diff, which can take gigantic
amounts of CPU time and/or heap given the right (ermm... wrong)
input. Limit the damage that large/expensive diffs can cause.
Eric Wong [Sun, 24 Feb 2019 02:55:40 +0000 (02:55 +0000)]
viewvcs: match 8000-byte lookup for git
No need to scan the entire string, but prefer to match git
behavior. This might be faster if/when Perl can create
substrings efficiently using CoW.
Fix a 80-column violation while we're at it.
Eric Wong [Thu, 14 Feb 2019 02:29:10 +0000 (02:29 +0000)]
viewvcs: preliminary support for showing non-blobs
Eventually, we'll have special displays for various git objects
(commit, tree, tag). But for now, we'll just use git-show
to spew whatever comes from git.
Eric Wong [Sun, 17 Feb 2019 05:09:50 +0000 (05:09 +0000)]
git: add "commit_title" method
This will be useful for extracting titles/subjects from
commit objects when displaying commits.
Eric Wong [Mon, 1 Apr 2019 23:54:40 +0000 (23:54 +0000)]
githttpbackend: serve $GIT_DIR/info/attributes
This will be useful for reproducibility when mirroring
coderepos and generating diffs.
Eric Wong [Tue, 2 Apr 2019 00:45:48 +0000 (00:45 +0000)]
viewvcs: check for premature EOF from git-cat-file
Not entirely sure what is causing this, but it appears to
be causing infinite loops when attempting to display certain
blobs.
Fortunately, the fair scheduling of public-inbox-httpd prevented
this from becoming a real problem aside from increasing CPU
usage.
Eric Wong [Fri, 8 Mar 2019 22:11:02 +0000 (22:11 +0000)]
config: ignore missing config files
There's no reason for us to have git-config(1) warn users when a
config file is entirely missing.
Eric Wong [Fri, 8 Mar 2019 22:52:17 +0000 (22:52 +0000)]
v2writable: index_sync adds new epochs to alternates
Newly-cloned epochs need to be in alternates file of
all.git for the web and NNTP interfaces to work. So
allow invocations of "public-inbox-index" to idempotently
ensure the epoch is visible from the all.git repo.
Eric Wong [Fri, 8 Mar 2019 22:52:16 +0000 (22:52 +0000)]
v2writable: hoist out fill_alternates
We'll be using this sub to fill $GIT_DIR/objects/info/alternates
if somebody uses clone --mirror, too
Eric Wong [Wed, 27 Feb 2019 20:25:36 +0000 (20:25 +0000)]
v2writable: fix epoch rollover on incremental imports
All of our internal epoch rollover calculations are done using
the estimated unpacked (and uncompressed) size of the repo. The
importer instance needs to check that unpacked size before
selecting an epoch when an epoch already has packed data.
This bug did not impact the initial mass imports since we only
initialize the Import instance once-per-epoch and did not need
to take existing epochs into account.
Tested manually with -mda on a local clone of LKML
Reported-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
Eric Wong [Sat, 23 Feb 2019 10:56:38 +0000 (10:56 +0000)]
www: prevent '!important' in BOFH-specified CSS
CSS specified by the BOFH must never take precedence over
what a user sets in userContent.css.
Eric Wong [Sat, 23 Feb 2019 10:56:37 +0000 (10:56 +0000)]
usercontent.pm: set "!important" to override BOFH prefs
The sample userContent.css needs a higher priority than what
the BOFH specifies. In other words, user preference must
ALWAYS take precedence.
Reported-by: Dmitry Alexandrov <321942@gmail.com>
cf. https://public-inbox.org/meta/87mumn4kx8.fsf@gmail.com/
Eric Wong [Wed, 13 Feb 2019 22:44:42 +0000 (22:44 +0000)]
ensure bytes::length is available to callers
We were relying on Danga::Socket using the "bytes" pragma,
previously. Nowadays, the "bytes" pragma is not recommended in
general, but bytes::length remains acceptable for getting the
byte-size of a scalar.
Eric Wong [Tue, 12 Feb 2019 23:01:52 +0000 (23:01 +0000)]
MANIFEST: add newswww.psgi
Fixes: 285b9b4d7de53b0d ("examples/newswww.psgi: demonstrate standalone NewsWWW usage")
Eric Wong [Wed, 6 Feb 2019 23:22:56 +0000 (23:22 +0000)]
t/perf-msgview.t: fix broken performance test
WwwStream started depending on the WWW::style method
for configurable CSS, so mock ::style so the benchmark
runs properly.
Fixes: f026dbdd392c9dd5 ('www: admin-configurable CSS via "publicinbox.css"')
Eric Wong [Wed, 6 Feb 2019 23:12:10 +0000 (23:12 +0000)]
t/perf-msgview: don't warn about --unordered if skipping
No point in making noise about something that isn't used.
Eric Wong [Mon, 4 Feb 2019 11:00:26 +0000 (11:00 +0000)]
examples/newswww.psgi: demonstrate standalone NewsWWW usage
Plack::Builder allows "mounting" on with hostnames as well as
path names to enable virtual hosting. This example demonstrates
how port 80/443 for "news.example.com" can redirect browser
requests when somebody attempts to use a "nntp://" URL and
the software assumes "http://"
Eric Wong [Wed, 6 Feb 2019 01:15:55 +0000 (01:15 +0000)]
nntp: get rid of long_response_limit constant
Unused since commit
b8c41362f2a5c8fcc6b1846a79c72bfa77565297
("nntp: simplify the long_response API")
Eric Wong [Wed, 6 Feb 2019 10:33:21 +0000 (10:33 +0000)]
http: cleanup partial-write handling on readonly values
Don't bother assigning to $_[1]; just let Danga::Socket
do its thing since $_[1] should be out-of-scope soon.
Eric Wong [Wed, 6 Feb 2019 11:07:41 +0000 (11:07 +0000)]
daemon: favor Socket.pm in Perl 5.14+ for IPv6
Users on Perl 5.14+ are common, so we can try the bundled Socket
(not "Socket6") module before attempting Socket6 for IPv6.
Eric Wong [Wed, 6 Feb 2019 11:07:40 +0000 (11:07 +0000)]
daemon: die properly when Socket6 is missing
We need to keep Unix-socket-only httpd instances working
without Socket6. This fixes t/httpd-unix.t with Socket6
uninstalled.
Eric Wong [Tue, 5 Feb 2019 22:24:19 +0000 (22:24 +0000)]
Merge remote-tracking branch 'origin/help-color'
* origin/help-color:
wwwtext: inline sample CSS and use highlight
hlmod: support "```$LANG" blocks in text
hlmod: do_hl* performs src_escape immediately
hlmod: make into a singleton
hlmod: hoist out do_hl_lang sub
viewvcs: cleanup utf8 handling
Eric Wong [Tue, 5 Feb 2019 21:33:16 +0000 (21:33 +0000)]
solvergit: include the $oid_want tmpdir name
This can help admins diagnose problems with SolverGit, since
qspawn logs the failed "git apply" command-line in stderr.
(or it can waste admins' time because sometimes there's crap
mail clients which mangle patches)
Eric Wong [Tue, 5 Feb 2019 11:25:11 +0000 (11:25 +0000)]
TODO: add item for "scraper" importers
The git-users mailing list is on Google Groups with obfuscated
addresses and censored archives. We should allow users to
import them soon, as obfuscated/censored archives are better
than not having archives at all when Google decides to shut down
yet-another-service.
There's also some mangling that (most) instances of Mailman do
(e.g. cgit), but being able to follow such groups over NNTP
and use our search functionality is still useful and better
than what typical Mailman installations provide.
Eric Wong [Tue, 5 Feb 2019 05:28:03 +0000 (05:28 +0000)]
wwwtext: inline sample CSS and use highlight
For user documentation regarding CSS; showing users the sample
CSS with comments is probably more helpful than having
standalone documentation on CSS classes.
Eric Wong [Tue, 5 Feb 2019 05:24:39 +0000 (05:24 +0000)]
hlmod: support "```$LANG" blocks in text
This is compatible with Markdown; but we still keep the WYSIWYG
nature of plain-text with this. This is only intended for use
with our documentation. Enabling any type of Markdown support
for emails can lead to incompatibilities or interopability
problems with alternative implementations.
Eric Wong [Sat, 2 Feb 2019 16:46:47 +0000 (16:46 +0000)]
hlmod: do_hl* performs src_escape immediately
We want to be able to take advantage of this in other modules
Eric Wong [Sat, 2 Feb 2019 07:35:06 +0000 (07:35 +0000)]
hlmod: make into a singleton
It turns out there's no point in having multiple instances of
this or having to worry about destruction or destruction
ordering.
This will make it easier to reuse the one instance we have
across different modules.
Eric Wong [Sat, 2 Feb 2019 06:02:41 +0000 (06:02 +0000)]
hlmod: hoist out do_hl_lang sub
We'll want to use to support highlighting syntax used by
Markdown and possibly other markup languages (while retaining
the raw plain-text layout and formatting).
Eric Wong [Fri, 1 Feb 2019 22:12:52 +0000 (22:12 +0000)]
viewvcs: cleanup utf8 handling
Favor in-place utf8::decode since it's a bit faster without
method dispatch overhead; and don't care about validity just
yet.
HlMod->do_hl itself should return "utf8" strings, since other
parts of our code can use it, so it's not the job of ViewVCS to
post-process HlMod output.
Eric Wong [Mon, 4 Feb 2019 22:37:32 +0000 (22:37 +0000)]
TODO: support migration to v2 without breaking v1 "git fetch"
I'd like to move https://public-inbox.org/git/ to v2; but
cronjobs using "git fetch" to following should not break.
Eric Wong [Mon, 4 Feb 2019 11:40:13 +0000 (11:40 +0000)]
viewdiff: group path match to not be confused by "/dev/null"
Leaving out parentheses caused transitions to state="del" or
state="add" to be misidentified.
cf. https://public-inbox.org/meta/
20190204105454.GG10587@szeder.dev/
Reported-by: SZEDER Gábor <szeder.dev@gmail.com>
Eric Wong [Wed, 9 Jan 2019 11:43:27 +0000 (11:43 +0000)]
newswww: add /$MESSAGE_ID global redirector endpoint
This is the fallback for the normal WWW endpoint.
Adding this to the top-level seems to be alright, since lynx and
w3m both understand nntp://<HOSTNAME>/<Message-ID> anyways.
If newsgroup and inbox names conflict, then consider it the
fault of the original sender.
Since NewsWWW is intended to support buggy linkifiers in mail clients,
they can interpret nntp:// URLs as http://<HOSTNAME>/<Message-ID>
Inbox ordering from the config file is preserved since
commit
cfa8ff7c256e20f3240aed5f98d155c019788e3b
("config: each_inbox iteration preserves config order"),
so admins can rely on that to configure how scanning
works.
Requested-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
cf. https://public-inbox.org/meta/
20190107190719.GE9442@pure.paranoia.local/
nntp://news.public-inbox.org/
20190107190719.GE9442@pure.paranoia.local
Eric Wong [Fri, 1 Feb 2019 07:36:02 +0000 (07:36 +0000)]
viewdiff: support renames and long paths in diffstat anchors
This is best-effort, but works well-enough in practice for
projects which use shell-friendly filenames as well as the
long path names for some Linux kernel selftests.
Eric Wong [Fri, 1 Feb 2019 06:30:17 +0000 (06:30 +0000)]
viewdiff: escape HTML ampersand for renames
For URLs we generate, we need to escape '&' in query parameters
for correctness.
Eric Wong [Fri, 1 Feb 2019 06:06:04 +0000 (06:06 +0000)]
viewvcs: allow '0' as a valid filename for blob downloads
Only to be pedantic...
Eric Wong [Fri, 1 Feb 2019 05:27:52 +0000 (05:27 +0000)]
linkify: support proto://hostname without trailing slash
Sometimes users will write "http://example.com" without the
trailing slash, which every browser and tool I've tested seems
to understand.
Eric Wong [Fri, 1 Feb 2019 03:06:47 +0000 (03:06 +0000)]
view: simplify quote splitting
Perl "split" can capture and group in the regexp itself,
so rely on that to shorten our code.
Comparing the /T/ HTML output of a thread from hell (on LKML with
1356 messages) reveals no difference in the rendered result.
Only the HTML source differs in newline placement before/after
the closing </span>
This allows a minor speedup on my X32 Thinkpad @ 1.6GHz with
the aforementioned LKML thread from hell:
before: 3.67s
after: 3.55s
Eric Wong [Fri, 1 Feb 2019 01:47:31 +0000 (01:47 +0000)]
view: fix broken hunk header hrefs in Atom feeds
We use absolute URLs in the Atom feeds (to ease
syndication/mirroring), so hunk headers need to point to the
solver URLs.
Eric Wong [Thu, 31 Jan 2019 11:33:49 +0000 (11:33 +0000)]
view: diffstat anchors for multi-message/attachment views
diffstat <-> ^diff anchors work within the same attachment or
message while in HTML views which display multiple messages.
Eric Wong [Thu, 31 Jan 2019 10:45:55 +0000 (10:45 +0000)]
viewdiff: diffstat links to diff anchors
This can be helpful for reviewing larger patches which span
across several files on the permalink (/$MESSAGE_ID/) HTML
page.
More work will be needed to get this working for the /T/ and /t/
pages which show multiple emails, as the filename-based anchors
will conflict at the moment.
Eric Wong [Mon, 18 Jan 2016 21:16:14 +0000 (21:16 +0000)]
hval: routines for attribute escaping
We'll use HTML attributes + anchor links to link to filenames
in coming commits.
Eric Wong [Thu, 31 Jan 2019 21:08:48 +0000 (21:08 +0000)]
Merge remote-tracking branch 'origin/purge'
* origin/purge:
implement public-inbox-purge tool
v2writable: read epoch on purge
v2writable: cleanup processes when done
v2writable: purge ignores non-existent git epoch directories
v2writable: ->purge returns undef on no-op
import: purge: reap fast-export process
hoist out resolve_repo_dir from -index
Eric Wong [Thu, 31 Jan 2019 09:04:12 +0000 (09:04 +0000)]
doc: remove completed TODO items
Eric Wong [Thu, 31 Jan 2019 09:17:45 +0000 (09:17 +0000)]
doc/config: document "replyto" configuration knob
I hate it, but it's necessary to support some mirrors.
Eric Wong [Thu, 31 Jan 2019 08:52:49 +0000 (08:52 +0000)]
doc/config: user documentation for limiters
I've relied on this feature to keep the VPS behind
https://public-inbox.org/git/ from OOM-ing since 2016,
so document it to ensure others can make use of low-end
servers like I do.
More limiters may become configurable for viewvcs and
solver functionality (or we continue using the default
one).
Eric Wong [Thu, 31 Jan 2019 08:28:41 +0000 (08:28 +0000)]
config: tiny cleanup to use _array() sub
Eric Wong [Thu, 31 Jan 2019 08:17:15 +0000 (08:17 +0000)]
qspawn: documentation updates
This will become critical for future changes to display
git commits, diffs, and trees.
Use "qspawn.wcb" instead of "qspawn.response" to enhance
readability.
Eric Wong [Thu, 31 Jan 2019 06:00:22 +0000 (06:00 +0000)]
inbox: drop psgi.url_scheme requirement from base_url
This will make it easier to make command-line tools
from SolverGit.
Eric Wong [Thu, 31 Jan 2019 04:10:29 +0000 (04:10 +0000)]
viewvcs: support streaming large blobs
Forking off git-cat-file here for streaming large blobs is
reasonably efficient, at least no worse than using
git-http-backend for serving clones. So let our limiter
framework deal with it.
git itself isn't great for large files, and AFAIK there's no
stable/widely-available mechanisms for reading smaller chunks
of giant blobs in git itself.
Tested with some giant GPU headers in the Linux kernel.
Eric Wong [Thu, 31 Jan 2019 02:46:13 +0000 (02:46 +0000)]
solvergit: allow shorter-than-necessary OIDs from user
We can rely on git to disambiguate, here; because sometimes
shorter OIDs can be unambiguous even if we only resolved the
longer one.
Eric Wong [Thu, 31 Jan 2019 01:41:15 +0000 (01:41 +0000)]
solvergit: allow searching on longer-than-needed OIDs
public-inbox can only index the abbreviated object_ids in
emails, not the full or even longer-than-necessary object_ids.
So retry failed object_ids if they're longer than 7 hex
characters.
Eric Wong [Thu, 31 Jan 2019 00:32:42 +0000 (00:32 +0000)]
inbox: perform cleanup of Git objects for coderepos
Otherwise, long-running but idle git processes may keep unlinked
packs around indefinitely and waste disk space.
Eric Wong [Wed, 30 Jan 2019 20:48:53 +0000 (20:48 +0000)]
t/config.t: test PublicInbox::Git sharing between inboxes
We need to ensure we don't introduce unnecessary processes
and memory usage for mapping multiple inboxes to the same
code repos.
Eric Wong [Wed, 30 Jan 2019 11:38:52 +0000 (11:38 +0000)]
doc/config: document coderepo and css bits
New features ought to be documented
Eric Wong [Wed, 30 Jan 2019 10:33:55 +0000 (10:33 +0000)]
solvergit: don't confuse Xapian with ".." in filenames
Xapian will interpret ".." as ranges, even quoted phrases.
So break up words on ".." since punctuation (AFAIK) is not
searchable, anyways.
Eric Wong [Wed, 30 Jan 2019 08:47:43 +0000 (08:47 +0000)]
git: use "git rev-parse --git-path"
Using git worktrees was causing t/solver_git.t to fail on me.
Eric Wong [Wed, 30 Jan 2019 07:40:35 +0000 (07:40 +0000)]
Merge remote-tracking branch 'origin/viewvcs' into master
* origin/viewvcs: (66 commits)
solvergit: deal with alternative diff prefixes
solvergit: extract mode from diff headers properly
solvergit: avoid "Wide character" warnings
solvergit: do not show full path names to "git apply"
css/216dark: add comments and tweak highlight colors
viewvcs: avoid segfault with highlight.pm at shutdown
solvergit: do not solve blobs twice
t/check-www-inbox: disable history
t/check-www-inbox: don't follow mboxes
t/check-www-inbox: replace IPC::Run with PublicInbox::Spawn
hval: add src_escape for highlight post-processing
viewvcs: wire up syntax-highlighting for blobs
hlmod: disable enclosing <pre> tag
t/hl_mod: extra check to ensure we escape HTML
wwwhighlight: read_in_full returns undef on errors
solver: crank up max patches to 9999
viewvcs: do not show final error message twice
qspawn: decode $? for user-friendliness
solver: reduce "git apply" invocations
solver: hold patches in temporary directory
...
Eric Wong [Wed, 30 Jan 2019 07:38:51 +0000 (07:38 +0000)]
view: remove unused _msg_date sub
Not needed since commit
956abe9ad5f13a0d1755262be412d6a54fda72e9
("view: depend on SearchMsg for Message-ID")
Eric Wong [Wed, 30 Jan 2019 07:39:02 +0000 (07:39 +0000)]
httpd: a few comments about some fields we set
Removing 'psgix.input.buffered' could be a possibility in
the future.
Eric Wong [Wed, 30 Jan 2019 04:38:26 +0000 (04:38 +0000)]
solvergit: deal with alternative diff prefixes
At least, without extra directory levels, since
git-diff supports --src-prefix and --dst-prefix,
and /git/
6aa8857a11/s/ uses it...
Eric Wong [Wed, 30 Jan 2019 04:29:24 +0000 (04:29 +0000)]
solvergit: extract mode from diff headers properly
grep() won't set $1, so use "=~", instead.
Eric Wong [Wed, 30 Jan 2019 03:31:12 +0000 (03:31 +0000)]
solvergit: avoid "Wide character" warnings
Just quiet Perl down, since we don't know or care about the
encoding of the patch we hand off to git-apply.
Eric Wong [Wed, 30 Jan 2019 02:32:29 +0000 (02:32 +0000)]
solvergit: do not show full path names to "git apply"
"git apply" will warn about whitespace with the full path of the
patch, which will expose the $TMPDIR environment to users over
HTTP(S).
This change breaks compatibility with git pre-1.8.5, again;
but that was released in late-2013; so hopefully everybody
is on newer versions.
Eric Wong [Tue, 29 Jan 2019 22:31:26 +0000 (22:31 +0000)]
css/216dark: add comments and tweak highlight colors
Overkill, but "highlight" supports single-line comments (slc)
independently of multi-line comments (com); but we'll use the
same color for that.
We'll also use #0f0 instead of #0ff for "kwb" (keyword class "b")
since blue shades are prevalent in <a> links and comments, while
green was unused.
Eric Wong [Tue, 29 Jan 2019 21:35:25 +0000 (21:35 +0000)]
viewvcs: avoid segfault with highlight.pm at shutdown
Proper ordering of destruction seems required to avoid segfaults
at shutdown.
Eric Wong [Tue, 29 Jan 2019 04:01:21 +0000 (04:01 +0000)]
solvergit: do not solve blobs twice
In some cases, a file may ping-pong between blob IDs in the same
message when reverts occur. So break out of this early.
This doesn't account for different abbreviations, but the
limited variations of abbreviations should alleviate the
problem.
Eric Wong [Tue, 29 Jan 2019 03:42:48 +0000 (03:42 +0000)]
t/check-www-inbox: disable history
WWW::Mechanize keeps an infinitely large stack, which was
leading to OOM errors on my system.
Eric Wong [Tue, 29 Jan 2019 03:37:24 +0000 (03:37 +0000)]
t/check-www-inbox: don't follow mboxes
They can be extremely large with no limit, so can lead to OOM
errors.
Eric Wong [Mon, 28 Jan 2019 09:02:42 +0000 (09:02 +0000)]
t/check-www-inbox: replace IPC::Run with PublicInbox::Spawn
Because WWW::Mechanize uses truckload of memory, fork
needs to prepare all that memory for CoW, which ends up
bailing with ENOMEM.
Eric Wong [Tue, 29 Jan 2019 07:44:28 +0000 (07:44 +0000)]
mid: filter out 'y', 'n', and email addresses from references()
Looking at git@vger history, several emails had broken
References/In-Reply-To pointing to <y>, <n> and email
addresses as Message-IDs in References and In-Reply-To
headers.
This was causing too many unrelated messages to be linked
together in the same thread.
Eric Wong [Sun, 27 Jan 2019 11:43:56 +0000 (11:43 +0000)]
hval: add src_escape for highlight post-processing
We need to post-process "highlight" output to ensure it doesn't
contain odd bytes which cause "wide character" warnings or
require odd glyphs in source form.
Eric Wong [Sun, 27 Jan 2019 11:38:42 +0000 (11:38 +0000)]
viewvcs: wire up syntax-highlighting for blobs
And update 216dark.css to match a color scheme I'm used to;
which is fairly minimal and doesn't use all the classes
"highlight" provides.
Eric Wong [Sun, 27 Jan 2019 11:36:22 +0000 (11:36 +0000)]
hlmod: disable enclosing <pre> tag
We already have a <pre> tag in ViewVCS, and nesting <pre>
inside the pre-existing <pre> overrides the "white-space:pre"
we use to align line numbers.
Eric Wong [Sun, 27 Jan 2019 11:35:29 +0000 (11:35 +0000)]
t/hl_mod: extra check to ensure we escape HTML
Otherwise, it's open season on our users :<
Eric Wong [Sun, 27 Jan 2019 05:09:22 +0000 (05:09 +0000)]
wwwhighlight: read_in_full returns undef on errors
The return value of "print" is not undef for Perl IO::Handle.
Eric Wong [Sun, 27 Jan 2019 03:38:41 +0000 (03:38 +0000)]
solver: crank up max patches to 9999
Might as well, since the only constraint is filesystem space
for temporary files for public-inbox-httpd users.
-httpd can fairly share work across clients with our use of
psgi_qx; and there's a recent patch series in git@vger with 64
patches in sequence.
Eric Wong [Sun, 27 Jan 2019 03:32:01 +0000 (03:32 +0000)]
viewvcs: do not show final error message twice
SolverGit::ERR already writes the exception to the debug
log before calling {user_cb}, so there's no need for viewvcs
to append it.
Eric Wong [Sun, 27 Jan 2019 03:30:30 +0000 (03:30 +0000)]
qspawn: decode $? for user-friendliness
The raw value of $? isn't very useful, generally.
Eric Wong [Sun, 27 Jan 2019 00:34:35 +0000 (00:34 +0000)]
solver: reduce "git apply" invocations
"git apply" is capable of applying multiple patches in one
invocation, so give it multiple patches on the command-line
now that we no longer rely on anonymous file handles to hold
patches.
This cuts down a 64-patch series on git@vger from ~1s to ~800ms
with vfork spawn enabled using Inline::C.
Eric Wong [Sun, 27 Jan 2019 00:21:51 +0000 (00:21 +0000)]
solver: hold patches in temporary directory
We can avoid bumping up RLIMIT_NOFILE too much by storing
patches in a temporary directory. And we can share this
top-level directory with our temporary git repository.
Since we no longer rely on a working-tree for git, we are free
to rearrange the layout and avoid relying on the ".git"
convention and relying on "git -C" for chdir.
This may also ease porting public-inbox to older systems
where git does not support "-C" for chdir.
Eric Wong [Tue, 22 Jan 2019 02:10:13 +0000 (02:10 +0000)]
solver: rewrite to use Qspawn->psgi_qx and pi-httpd.async
The psgi_qx routine in the now-abandoned "repobrowse" branch
allows us to break down blob-solving at each process execution
point. It reuses the Qspawn facility for git-http-backend(1),
allowing us to limit parallel subprocesses independently of Perl
worker count.
This is actually a 2-3% slower a fully-synchronous execution;
but it is fair to other clients as it won't monopolize the server
for hundreds of milliseconds (or even seconds) at a time.
Eric Wong [Sat, 26 Jan 2019 20:05:56 +0000 (20:05 +0000)]
view: swap CRLF for LF in HTML output
It makes no difference to browsers aside from saving a few
bytes; and this means we won't have to worry about extra
'%0D' showing up in links to solver.
Eric Wong [Tue, 22 Jan 2019 07:57:06 +0000 (07:57 +0000)]
t/qspawn.t: psgi_qx stderr test
Eric Wong [Wed, 18 Jan 2017 07:27:03 +0000 (07:27 +0000)]
qspawn: implement psgi_qx
This new asynchronous API, will allow us to take
advantage of non-blocking I/O from even small commands;
as those may still need to wait for slow operations.
Eric Wong [Fri, 13 Jan 2017 23:10:25 +0000 (23:10 +0000)]
httpd/async: stop running command if client disconnects
If an HTTP client disconnects while we're piping the output of a
process to them, break the pipe of the process to reclaim
resources as soon as possible.