| Date | Commit message (Collapse) |
|---|
|
I didn't wait until September to do it, this year!
|
|
It seems to make sense to the target audience that any of
the URLs displayed could work.
|
|
While both can be correct, the former seems more common,
is shorter, and is also consistent with the spelling found
in the AGPL-3.0 text.
|
|
This avoids uninitialized variable warnings when viewing
newly-created files.
|
|
We're often iterating through messages while writing to another
buffer in our WWW interface, causing memory usage to multiply.
Since we know we won't need to keep the MIME object around in
some cases, and can tell msg_iter to clobber the on-stack
variable while it operates on subparts of multipart messages.
With xt/mem-msgview.t switched to multipart from the previous
commit, this shows a 13 MB memory reduction on that test.
|
|
There's a bunch of leftover "require" and "use" statements we no
longer need and can get rid of, along with some excessive
imports via "use".
IO::Handle usage isn't always obvious, so add comments
describing why a package loads it. Along the same lines,
document the tmpdir support as the reason we depend on
File::Temp 0.19, even though every Perl 5.10.1+ user has it.
While we're at it, favor "use" over "require", since it it gives
us extra compile-time checking.
|
|
While filenames are escaped, the actual diff contents may
contain an unescaped "\r" carriage return byte not in front
of the "\n" line feed. So just allow "\r" to appear in the
middle of a line.
|
|
Initialize the $di hashref at use to make it more obvious it's
a local variable. We can also use the :utf8 IO layer via
open+print to save ourselves the trouble of converting the UTF-8
patch to an octet stream.
|
|
This is needed to work with patches with many renames,
such as what makes "git/eebf7a8/s/?b=t%2Ftest-lib.sh"
|
|
solver can spawn multiple processes per HTTP request, but
"git apply" failures are needlessly noisy due to corrupt
patches. We also don't want to silence "git ls-files"
or "git update-index" errors using $env->{'qspawn.quiet'},
either, so this granularity is needed.
Admins can check for 500 errors in access logs to detect
(and reproduce) solver failures, anyways, so there's no
need to log every time "git apply" rejects a corrupt patch.
|
|
Rewrite the patch extraction loop using a single regexp which
accounts for missing "diff --git ..." lines and is capable of
extracting pathnames off the "+++ b/foo" line.
This fixes the solving of blob "96f1c7f" off
<2841d2de-32ad-eae8-6039-9251a40bb00e@tngtech.com>
in git@vger archives.
v2:
* Fix regressions in git@vger archives:
- git/776fa90f7f/s/?b=contrib/git-jump/git-jump
(fallback to "old mode" properly)
- git/5cd8845/s/?b=submodule.c
(no leading space in context)
* use "state" in a Perl <5.28.0-compatible way
|
|
Sometimes a patch is corrupted and resent to create the same
OID. We need to account for that case and actually move onto
the next patch instead of blindly trying "git ls-files" to get
nothing out of it.
|
|
This simplifies our admin module a bit and allows solver to be
used with v1 inboxes using git versions prior to v1.8.5 (but
still >= git v1.8.0).
|
|
We can save callers the trouble of {-hold} and {-dev_null}
refs as well as the trouble of calling fileno().
|
|
This allows us to get rid of the requirement to capture
on-stack variables with an anonymous sub, as illustrated
with the update to viewvcs to take advantage of this.
v2: fix error handling for missing OIDs
|
|
And remove the last anonymous sub in SolverGit itself.
|
|
By passing a user-supplied arg to $qx_cb, we can eliminate the
callers' need to capture on-stack variables with a closure.
This saves several kilobytes of memory allocation at the expense
of some extra hash table lookups in user-supplied callbacks. It
also reduces the risk of memory leaks by eliminating a common
source of circular references.
|
|
Another step towards removing anonymous subs to eliminate
a possible source of memory leaks and high memory use.
|
|
This is distributed with Perl 5.10.1 and onwards, so it should
not be an installation burden for any users. I'm planning to
move away from tempdir() entirely and use File::Temp->newdir to
remove dependencies on END{} blocks.
|
|
I sometimes post context-free documentation patches generated
with "-U0" to reduce size and bandwidth overhead when replacing
URLs or updating copyright notices. git-apply(1) needs the
--unidiff-zero switch to work properly with context-free
patches.
Given our search looks for blob OIDs, and we're never going
to be running the code we regenerate, "--unidiff-zero" ought
to be safe.
|
|
When solving for blob 81c1164ae5 in https://public-inbox.org/git/,
at least two messages get indexed with the dfpost result for
that blob (after fixing MsgIter to decode all text/* parts):
1. https://public-inbox.org/git/b9fb52b8-8168-6bf0-9a72-1e6c44a281a5@oracle.com/
2. https://public-inbox.org/git/56664222-6c29-09dc-ef78-7b380b113c4a@oracle.com/
However, only the first message contains a usable patch. So
we must adjust SolverGit to account for multiple messages
hitting the same "dfpost:" search result and attempt
"git apply" on all results, not just the first.
In the future, changes to SearchIdx.pm may rid us of invalid
search results and speed up performance (at the expense of
developer/indexing time); but we need to account for old search
indices, first.
|
|
Although we always unlink temporary files, give them a
meaningful name so that we can we can still make sense
of the pre-unlink name when using lsof(8) or similar
tools on Linux.
|
|
It's possible for Qspawn callers to be deferred, in which case
we must ensure we don't cause the temporary file used for
stdin to become unref-ed and closed.
This can be a problem when we exceed the default Qspawn
limiter of 32 concurrent processes for "git update-index".
|
|
We're using Qspawn, now
|
|
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.
|
|
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.
|
|
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.
|
|
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)
|
|
We can rely on git to disambiguate, here; because sometimes
shorter OIDs can be unambiguous even if we only resolved the
longer one.
|
|
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.
|
|
Xapian will interpret ".." as ranges, even quoted phrases.
So break up words on ".." since punctuation (AFAIK) is not
searchable, anyways.
|
|
Using git worktrees was causing t/solver_git.t to fail on me.
|
|
At least, without extra directory levels, since
git-diff supports --src-prefix and --dst-prefix,
and /git/6aa8857a11/s/ uses it...
|
|
grep() won't set $1, so use "=~", instead.
|
|
Just quiet Perl down, since we don't know or care about the
encoding of the patch we hand off 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.
|
|
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.
|
|
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.
|
|
"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.
|
|
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.
|
|
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.
|
|
It was harmless, besides wasting space and memory.
|
|
..if the Email::MIME ->crlf is LF.
Email::MIME::Encodings forces everything to CRLF on
quoted-printable messages for RFC-compliance; and
git-apply --ignore-whitespace seems to miss a context
line which is just "\r\n" (w/o leading space).
|
|
Apparently Email::MIME returns quoted-printable text
with CRLF. So use --ignore-whitespace with git-apply(1)
and ensure we don't capture '\r' in pathnames from
those emails.
And restore "$@" dumping when we die while solving.
|
|
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.
|
|
It's not likely to be worth our time to support
a callback-driven model for something which happens
once per patch series.
|
|
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.
|
|
We'll be breaking this up into several steps, too; since
searching inboxes for patch blobs can take 10s of milliseconds
for me.
|
|
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.
|
|
We want more fine-grained scheduling for PSGI use, as
the patch application step can take hundreds of milliseconds
on my modest hardware
|