revlog: automatically read from opened file handles
The revlog reading code commonly opens a new file handle for
reading on demand. There is support for passing a file handle
to revlog.revision(). But it is marked as an internal argument.
When revlogs are written, we write() data as it is available. But
we don't flush() data until all revisions are written.
Putting these two traits together, it is possible for an in-process
revlog reader during active writes to trigger the opening of a new
file handle on a file with unflushed writes. The reader won't have
access to all "available" revlog data (as it hasn't been flushed).
And with the introduction of the previous patch, this can lead to
the revlog raising an error due to a partial read.
I witnessed this behavior when applying changegroup data (via
`hg pull`) before issue6006 was fixed via different means. Having
this and the previous patch in play would have helped cause errors
earlier rather than manifesting as hash verification failures.
While this has been a long-standing issue, I believe the relatively
new delta computation code has tickled it into being more common.
This is because the new delta computation code will compute deltas
in more scenarios. This can lead to revlog reading. While the delta
computation code is probably supposed to reuse file handles, it
appears it isn't doing so in all circumstances.
But the issue runs deeper than that. Theoretically, any code can
access revision data during revlog writes. It appears we were just
getting lucky that it wasn't. (The "add revision callback" passed to
addgroup() provides an avenue to do this.)
If I changed the revlog's behavior to not cache the full revision
text or to clear caches after revision insertion during addgroup(),
I was able to produce crashes 100% of the time when writing changelog
revisions. This is because changelog's add revision callback attempts
to resolve the revision data to access the changed files list. And
without the revision's fulltext being cached, we performed a revlog
read, which required opening a new file handle. This attempted to read
unflushed data, leading to a partial read and a crash.
This commit teaches the revlog to store the file handles used for
writing multiple revisions during addgroup(). It also teaches the
code for resolving a file handle when reading to use these handles,
if available. This ensures that *any* reads (regardless of their
source) use the active writing file handles, if available. These
file handles have access to the unflushed data because they wrote it.
This allows reads to complete without issue.
Differential Revision: https://phab.mercurial-scm.org/D5267
$ . "$TESTDIR/remotefilelog-library.sh"
$ hg init master
$ cd master
$ cat >> .hg/hgrc <<EOF
> [remotefilelog]
> server=True
> EOF
$ echo x > x
$ hg commit -qAm x
$ cd ..
# shallow clone from full
$ hgcloneshallow ssh://user@dummy/master shallow --noupdate
streaming all changes
2 files to transfer, 227 bytes of data
transferred 227 bytes in * seconds (*/sec) (glob)
searching for changes
no changes found
$ cd shallow
$ cat .hg/requires
dotencode
exp-remotefilelog-repo-req-1
fncache
generaldelta
revlogv1
store
$ hg update
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
1 files fetched over 1 fetches - (1 misses, 0.00% hit ratio) over *s (glob)
$ cat x
x
$ ls .hg/store/data
$ echo foo > f
$ hg add f
$ hg ci -m 'local content'
$ ls .hg/store/data
4a0a19218e082a343a1b17e5333409af9d98f0f5
$ cd ..
# shallow clone from shallow
$ hgcloneshallow ssh://user@dummy/shallow shallow2 --noupdate
streaming all changes
3 files to transfer, 564 bytes of data
transferred 564 bytes in * seconds (*/sec) (glob)
searching for changes
no changes found
$ cd shallow2
$ cat .hg/requires
dotencode
exp-remotefilelog-repo-req-1
fncache
generaldelta
revlogv1
store
$ ls .hg/store/data
4a0a19218e082a343a1b17e5333409af9d98f0f5
$ hg update
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat x
x
$ cd ..
# full clone from shallow
Note: the output to STDERR comes from a different process to the output on
STDOUT and their relative ordering is not deterministic. As a result, the test
was failing sporadically. To avoid this, we capture STDERR to a file and
check its contents separately.
$ TEMP_STDERR=full-clone-from-shallow.stderr.tmp
$ hg clone --noupdate ssh://user@dummy/shallow full 2>$TEMP_STDERR
streaming all changes
remote: abort: Cannot clone from a shallow repo to a full repo.
[255]
$ cat $TEMP_STDERR
abort: pull failed on remote
$ rm $TEMP_STDERR
# getbundle full clone
$ printf '[server]\npreferuncompressed=False\n' >> master/.hg/hgrc
$ hgcloneshallow ssh://user@dummy/master shallow3
requesting all changes
adding changesets
adding manifests
adding file changes
added 1 changesets with 0 changes to 0 files
new changesets b292c1e3311f
updating to branch default
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ ls shallow3/.hg/store/data
$ cat shallow3/.hg/requires
dotencode
exp-remotefilelog-repo-req-1
fncache
generaldelta
revlogv1
store