about summary refs log tree commit homepage
DateCommit message (Collapse)
2019-01-20solver: add a TODO note about making this fully evented
Applying a 100+ patch series can be a pain and lead to a wayward client monopolizing the connection. On the other hand, we'll also need to be careful and limit the number of in-flight file descriptors and parallel git-apply processes when we move to an evented model, here.
2019-01-20solver: note the synchronous nature of index preparation
It's not likely to be worth our time to support a callback-driven model for something which happens once per patch series.
2019-01-20solver: break @todo loop into a callback
This will allow each patch search via Xapian to "yield" the current client in favor of another client in the PSGI web interface for fairness.
2019-01-20solver: simplify control flow for initial loop
We'll be breaking this up into several steps, too; since searching inboxes for patch blobs can take 10s of milliseconds for me.
2019-01-20solver: switch patch application to use a callback
A bit messy at the moment, but we need to break this up into smaller steps for fairness with other clients, as applying dozens of patches can take several hundred milliseconds.
2019-01-20solver: break up patch application steps
We want more fine-grained scheduling for PSGI use, as the patch application step can take hundreds of milliseconds on my modest hardware
2019-01-20solver: more verbose blob resolution
Help users find out where each step of the resolution came from. Also, we must clean abort the process if we have missing blobs. And refine the output to avoid unnecessary braces, too.
2019-01-20git: support 'ambiguous' result from --batch-check
David Turner's patch to return "ambiguous" seems like a reasonable patch for future versions of git: https://public-inbox.org/git/672a6fb9e480becbfcb5df23ae37193784811b6b.camel@novalis.org/
2019-01-20view: enable naming hints for raw blob downloads
Meaningful names in URLs are nice, and it can make life easier for supporting syntax-highlighting
2019-01-19solver: operate directly on git index
No need to incur extra I/O traffic with a working-tree and uncompressed files on the filesystem. git can handle patch application in memory and we rely on exact blob matching anyways, so no need for 3way patch application.
2019-01-19git: disable abbreviations with cat-file hints
Ambiguity is not worth it for internal usage with the solver.
2019-01-19view: wire up diff and vcs viewers with solver
2019-01-19solver: various bugfixes and cleanups
Remove the make_path dependency and call mkdir directly. Capture mode on new files, avoid referencing non-existent functions and enhance the debug output for users to read.
2019-01-19git: check saves error on disambiguation
This will be useful for disambiguating short OIDs in older emails when abbreviations were shorter. Tested against the following script with /path/to/git.git ==> t.perl <== use strict; use PublicInbox::Git; use Data::Dumper; my $dir = shift or die "Usage: $0 GIT_DIR # (of git.git)"; my $git = PublicInbox::Git->new($dir); my @res = $git->check('dead'); print Dumper({res => \@res, err=> $git->last_check_err}); @res = $git->check('5335669531d83d7d6c905bcfca9b5f8e182dc4d4'); print Dumper({res => \@res, err=> $git->last_check_err});
2019-01-19git: add git_quote
It'll be helpful for displaying progress in SolverGit output.
2019-01-19git: support multiple URL endpoints
For redundancy and centralization resistance.
2019-01-19solver: initial Perl implementation
This will lookup git blobs from associated git source code repositories. If the blobs can't be found, an attempt to "solve" them via patch application will be performed. Eventually, this may become the basis of a type-agnostic frontend similar to "git show"
2019-01-19t/perf-msgview: add test to check msg_html performance
This will be necessary to ensure we maintain reasonable performance when we add diff-highlighting support.
2019-01-19hval: force monospace for <form> elements, too
Same reasoning as commit 7b7885fc3be2719c068c0a2fc860d53f17a1d933, because GUI browsers have a tendency to use a different font-family (and thus different size) as the rest of the page.
2019-01-19view: disable bold in topic display
It seems pointless due to the indentation, and interacts badly with some CSS colouring.
2019-01-18t/git.t: do not pass "-b" to git-repack(1)
Allows t/git.t to run on older versions of git without "-b" and avoids incurring extra I/O traffic for bitmaps.
2019-01-18git: git_unquote handles double-quote and backslash
We need to work with 0x22 (double-quote) and 0x5c (backslash); even if they're oddball characters in filenames which wouldn't be used by projects I'd want to work on.
2019-01-18t/git.t: avoid passing read-only value to git_unquote
Older versions of Perl (tested 5.14.2 on Debian wheezy(*), reported by Konstantin on Perl 5.16.3) considered the result of concatenating two string literals to be a constant value. (*) not that other stuff works on wheezy, but t/git.t should. Reported-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
2019-01-16nntp: header responses use CRLF consistently
Alpine is apparently stricter than other clients I've tried w.r.t. using CRLF for headers. So do the same thing we do for bodies to ensure we only emit CRLFs and no bare LFs. Reported-by: Wang Kang <i@scateu.me> https://public-inbox.org/meta/alpine.DEB.2.21.99.1901161043430.29788@la.scateu.me/
2019-01-15config: inbox name checking matches git.git more closely
Actually, it turns out git.git/remote.c::valid_remote_nick rules alone are insufficient. More checking is performed as part of the refname in the git.git/refs.c::check_refname_component I also considered rejecting URL-unfriendly inbox names entirely, but realized some users may intentionally configure names not handled by our WWW endpoint for archives they don't want accessible over HTTP.
2019-01-15git_unquote: perform modifications in-place
This function doesn't have a lot of callers at the moment so none of them are affected by this change. But the plan is to use this in our WWW code for things, so do it now before we call it in more places. Results from a Thinkpad X200 with a Core2Duo P8600 @ 2.4GHz: Benchmark: timing 10 iterations of cp, ip... cp: 12.868 wallclock secs (12.86 usr + 0.00 sys = 12.86 CPU) @ 0.78/s (n=10) ip: 10.9137 wallclock secs (10.91 usr + 0.00 sys = 10.91 CPU) @ 0.92/s (n=10) Note: I mainly care about unquoted performance because that's the common case for the target audience of public-inbox. Script used to get benchmark results against the Linux source tree: ==> bench_unquote.perl <== use strict; use warnings; use Benchmark ':hireswallclock'; my $nr = 50; my %GIT_ESC = ( a => "\a", b => "\b", f => "\f", n => "\n", r => "\r", t => "\t", v => "\013", ); sub git_unquote_ip ($) { return $_[0] unless ($_[0] =~ /\A"(.*)"\z/); $_[0] = $1; $_[0] =~ s/\\([abfnrtv])/$GIT_ESC{$1}/g; $_[0] =~ s/\\([0-7]{1,3})/chr(oct($1))/ge; $_[0]; } sub git_unquote_cp ($) { my ($s) = @_; return $s unless ($s =~ /\A"(.*)"\z/); $s = $1; $s =~ s/\\([abfnrtv])/$GIT_ESC{$1}/g; $s =~ s/\\([0-7]{1,3})/chr(oct($1))/ge; $s; } chomp(my @files = `git -C ~/linux ls-tree --name-only -r v4.19.13`); timethese(10, { cp => sub { for (0..$nr) { git_unquote_cp($_) for @files } }, ip => sub { for (0..$nr) { git_unquote_ip($_) for @files } }, });
2019-01-15searchidx: move git_unquote to PublicInbox::Git
We'll be using it outside of searchidx...
2019-01-15index: allow working on unconfigured inboxes, again
2019-01-11HACKING: update Debian version information
It's been a few years since this was updated...
2019-01-10Merge commit 'mem'
* commit 'mem': view: more culling for search threads over: cull unneeded fields for get_thread searchmsg: remove unused fields for PSGI in Xapian results searchview: drop unused {seen} hashref searchmsg: remove Xapian::Document field searchmsg: get rid of termlist scanning for mid httpd: remove psgix.harakiri reference
2019-01-10t/v2writable.t: force more consistent "git log" output
This should probably use lower-level git plumbing, but until then, consistently add a bunch of --no-* options to "git log" to get more consistent output. Noticed-by: Johannes Berg https://public-inbox.org/meta/1538164205.14416.76.camel@sipsolutions.net/
2019-01-10check git version requirements
This allows v1 tests to continue working on git 1.8.0 for now. This allows git 2.1.4 packaged with Debian 8 ("jessie") to run old tests, at least. I suppose it's safe to drop Debian 7 ("wheezy") due to our dependency on git 1.8.0 for "merge-base --is-ancestor". Writing V2 repositories requires git 2.6 for "get-mark" support, so mask out tests for older gits.
2019-01-10INSTALL: Net::Server is not needed for systemd use
Also, move Socket6 down since it's usually pulled in as a dependency of Net::Server or SpamAssassin; and we can fail gracefully without it.
2019-01-10daemon: make Socket6 optional, note about Net::Socket::IP
It looks like Net::Socket::IP comes with Perl 5.20 and later; so we won't have to hassle users with another package to install.
2019-01-09doc: various overview-level module comments
Hopefully this helps people familiarize themselves with the source code.
2019-01-09INSTALL: fix Date::Parse dependency for Debian
So yes, both "libdatetime-perl" and "libtimedate-perl" exist in Debian. We want the latter for the Date::Parse module. (And the former pulls in THIRTY-SEVEN dependencies on a fresh sid chroot, ohg yrsgcnq vf abg bar bs gurz).
2019-01-08view: more culling for search threads
{mapping} overhead is now down to ~1.3M at the end of a giant thread from hell.
2019-01-08over: cull unneeded fields for get_thread
On a certain ugly /$INBOX/$MESSAGE_ID/T/ endpoint with 1000 messages in the thread, this cuts memory usage from 2.5M to 1.9M (which still isn't great, but it's a start).
2019-01-08searchmsg: remove unused fields for PSGI in Xapian results
These fields are only necessary in NNTP and not even stored in Xapian; so keeping them around for the PSGI web UI search results wastes nearly 80K when loading large result sets.
2019-01-08searchview: drop unused {seen} hashref
Unused since commit 5f09452bb7e6cf49fb6eb7e6cf166a7c3cdc5433 ("view: cull redundant phrases in subjects")
2019-01-08searchmsg: remove Xapian::Document field
We don't need to be carrying this around with the many SearchMsg objects we have. This saves about 20K from a large SearchView "&x=t" response.
2019-01-08searchmsg: get rid of termlist scanning for mid
It doesn't seem to be used anywhere
2019-01-08httpd: remove psgix.harakiri reference
We don't need to set "psgix." extension fields for things we don't support. This saves 138 bytes per-client in $env as measured by Devel::Size::total_size
2019-01-08view: fix wrong date for non-Xapian/SQLite v1 users
We need to parse the MIME object in order to get the datestamp for those sites. Fixes: 7d02b9e64455 ("view: stop storing all MIME objects on large threads")
2019-01-08nntp: fix uninitialized variable in event_read
do_write must return 0 or 1.
2019-01-08view: stop storing all MIME objects on large threads
While we try to discard the $smsg (SearchMsg) objects quickly, they remain referenced via $node (SearchThread::Msg) objects, which are stored forever in $ctx->{mapping} to cull redundant words out of subjects in the thread skeleton. This significantly cuts memory bloat with large search results with '&x=t'. Now, the search results overhead of SearchThread::Msg and linked objects are stable at around 350K instead of ~7M per response in a rough test (there's more savings to be had in the same areas). Several hundred kilobytes is still huge and a large per-client cost; but it's far better than MEGABYTES per-client.
2019-01-07t/mda_filter_rubylang.t: set PI_EMERGENCY for -mda
Tests should not write to the default ~/.public-inbox/emergency
2019-01-05shrink low-bandwidth pipes under Linux
I've hit /proc/sys/fs/pipe-user-pages-* limits on some systems. So stop hogging resources on pipes which don't benefit from giant sizes. Some of these can use eventfd in the future to further reduce resource use.
2019-01-05index: quiet down git-log error messages on new inboxes
The new t/*filter_rubylang.t tests call -index immediately after -init, which causes confusing messages to show up to the end user. Check the validity of the ref before calling "git-log".
2019-01-05filter/rubylang: fix SQLite DB lifetime problems
Clearly the AltId stuff was never tested for v2. Ensure this tricky filter (which reuses Msgmap to avoid introducing new serial numbers) doesn't trigger deadlocks SQLite due to opening a DB for writing multiple times. I went through several iterations of this change before going with this one, which is the least intrusive I could fine.