wireproto: add streams to frame-based protocol
Previously, the frame-based protocol was just a series of frames,
with each frame associated with a request ID.
In order to scale the protocol, we'll want to enable the use of
compression. While it is possible to enable compression at the
socket/pipe level, this has its disadvantages. The big one is it
undermines the point of frames being standalone, atomic units that
can be read and written: if you add compression above the framing
protocol, you are back to having a stream-based protocol as opposed
to something frame-based.
So in order to preserve frames, compression needs to occur at
the frame payload level.
Compressing each frame's payload individually will limit compression
ratios because the window size of the compressor will be limited
by the max frame size, which is 32-64kb as currently defined. It
will also add CPU overhead, as it is more efficient for compressors
to operate on fewer, larger blocks of data than more, smaller blocks.
So compressing each frame independently is out.
This means we need to compress each frame's payload as if it is part
of a larger stream.
The simplest approach is to have 1 stream per connection. This
could certainly work. However, it has disadvantages (documented below).
We could also have 1 stream per RPC/command invocation. (This is the
model HTTP/2 goes with.) This also has disadvantages.
The main disadvantage to one global stream is that it has the very
real potential to create CPU bottlenecks doing compression. Networks
are only getting faster and the performance of single CPU cores has
been relatively flat. Newer compression formats like zstandard offer
better CPU cycle efficiency than predecessors like zlib. But it still
all too common to saturate your CPU with compression overhead long
before you saturate the network pipe.
The main disadvantage with streams per request is that you can't
reap the benefits of the compression context for multiple requests.
For example, if you send 1000 RPC requests (or HTTP/2 requests for
that matter), the response to each would have its own compression
context. The overall size of the raw responses would be larger because
compression contexts wouldn't be able to reference data from another
request or response.
The approach for streams as implemented in this commit is to support
N streams per connection and for streams to potentially span requests
and responses. As explained by the added internals docs, this
facilitates servers and clients delegating independent streams and
compression to independent threads / CPU cores. This helps alleviate
the CPU bottleneck of compression. This design also allows compression
contexts to be reused across requests/responses. This can result in
improved compression ratios and less overhead for compressors and
decompressors having to build new contexts.
Another feature that was defined was the ability for individual frames
within a stream to declare whether that individual frame's payload
uses the content encoding (read: compression) defined by the stream.
The idea here is that some servers may serve data from a combination
of caches and dynamic resolution. Data coming from caches may be
pre-compressed. We want to facilitate servers being able to essentially
stream bytes from caches to the wire with minimal overhead. Being
able to mix and match with frames are compressed within a stream
enables these types of advanced server functionality.
This commit defines the new streams mechanism. Basic code for
supporting streams in frames has been added. But that code is
seriously lacking and doesn't fully conform to the defined protocol.
For example, we don't close any streams. And support for content
encoding within streams is not yet implemented. The change was
rather invasive and I didn't think it would be reasonable to implement
the entire feature in a single commit.
For the record, I would have loved to reuse an existing multiplexing
protocol to build the new wire protocol on top of. However, I couldn't
find a protocol that offers the performance and scaling characteristics
that I desired. Namely, it should support multiple compression
contexts to facilitate scaling out to multiple CPU cores and
compression contexts should be able to live longer than single RPC
requests. HTTP/2 *almost* fits the bill. But the semantics of HTTP
message exchange state that streams can only live for a single
request-response. We /could/ tunnel on top of HTTP/2 streams and
frames with HEADER and DATA frames. But there's no guarantee that
HTTP/2 libraries and proxies would allow us to use HTTP/2 streams
and frames without the HTTP message exchange semantics defined in
RFC 7540 Section 8. Other RPC protocols like gRPC tunnel are built
on top of HTTP/2 and thus preserve its semantics of stream per
RPC invocation. Even QUIC does this. We could attempt to invent a
higher-level stream that spans HTTP/2 streams. But this would be
violating HTTP/2 because there is no guarantee that HTTP/2 streams
are routed to the same server. The best we can do - which is what
this protocol does - is shoehorn all request and response data into
a single HTTP message and create streams within. At that point, we've
defined a Content-Type in HTTP parlance. It just so happens our
media type can also work as a standalone, stream-based protocol,
without leaning on HTTP or similar protocol.
Differential Revision: https://phab.mercurial-scm.org/D2907
#require symlink
== tests added in 0.7 ==
$ hg init test-symlinks-0.7; cd test-symlinks-0.7;
$ touch foo; ln -s foo bar; ln -s nonexistent baz
import with add and addremove -- symlink walking should _not_ screwup.
$ hg add
adding bar
adding baz
adding foo
$ hg forget bar baz foo
$ hg addremove
adding bar
adding baz
adding foo
commit -- the symlink should _not_ appear added to dir state
$ hg commit -m 'initial'
$ touch bomb
again, symlink should _not_ show up on dir state
$ hg addremove
adding bomb
Assert screamed here before, should go by without consequence
$ hg commit -m 'is there a bug?'
$ cd ..
== fifo & ignore ==
$ hg init test; cd test;
$ mkdir dir
$ touch a.c dir/a.o dir/b.o
test what happens if we want to trick hg
$ hg commit -A -m 0
adding a.c
adding dir/a.o
adding dir/b.o
$ echo "relglob:*.o" > .hgignore
$ rm a.c
$ rm dir/a.o
$ rm dir/b.o
$ mkdir dir/a.o
$ ln -s nonexistent dir/b.o
$ mkfifo a.c
it should show a.c, dir/a.o and dir/b.o deleted
$ hg status
M dir/b.o
! a.c
! dir/a.o
? .hgignore
$ hg status a.c
a.c: unsupported file type (type is fifo)
! a.c
$ cd ..
== symlinks from outside the tree ==
test absolute path through symlink outside repo
$ p=`pwd`
$ hg init x
$ ln -s x y
$ cd x
$ touch f
$ hg add f
$ hg status "$p"/y/f
A f
try symlink outside repo to file inside
$ ln -s x/f ../z
this should fail
$ hg status ../z && { echo hg mistakenly exited with status 0; exit 1; } || :
abort: ../z not under root '$TESTTMP/x'
$ cd ..
== cloning symlinks ==
$ hg init clone; cd clone;
try cloning symlink in a subdir
1. commit a symlink
$ mkdir -p a/b/c
$ cd a/b/c
$ ln -s /path/to/symlink/source demo
$ cd ../../..
$ hg stat
? a/b/c/demo
$ hg commit -A -m 'add symlink in a/b/c subdir'
adding a/b/c/demo
2. clone it
$ cd ..
$ hg clone clone clonedest
updating to branch default
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
== symlink and git diffs ==
git symlink diff
$ cd clonedest
$ hg diff --git -r null:tip
diff --git a/a/b/c/demo b/a/b/c/demo
new file mode 120000
--- /dev/null
+++ b/a/b/c/demo
@@ -0,0 +1,1 @@
+/path/to/symlink/source
\ No newline at end of file
$ hg export --git tip > ../sl.diff
import git symlink diff
$ hg rm a/b/c/demo
$ hg commit -m'remove link'
$ hg import ../sl.diff
applying ../sl.diff
$ hg diff --git -r 1:tip
diff --git a/a/b/c/demo b/a/b/c/demo
new file mode 120000
--- /dev/null
+++ b/a/b/c/demo
@@ -0,0 +1,1 @@
+/path/to/symlink/source
\ No newline at end of file
== symlinks and addremove ==
directory moved and symlinked
$ mkdir foo
$ touch foo/a
$ hg ci -Ama
adding foo/a
$ mv foo bar
$ ln -s bar foo
$ hg status
! foo/a
? bar/a
? foo
now addremove should remove old files
$ hg addremove
adding bar/a
adding foo
removing foo/a
commit and update back
$ hg ci -mb
$ hg up '.^'
1 files updated, 0 files merged, 2 files removed, 0 files unresolved
$ hg up tip
2 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ cd ..
== root of repository is symlinked ==
$ hg init root
$ ln -s root link
$ cd root
$ echo foo > foo
$ hg status
? foo
$ hg status ../link
? foo
$ hg add foo
$ hg cp foo "$TESTTMP/link/bar"
foo has not been committed yet, so no copy data will be stored for bar.
$ cd ..
$ hg init b
$ cd b
$ ln -s nothing dangling
$ hg commit -m 'commit symlink without adding' dangling
abort: dangling: file not tracked!
[255]
$ hg add dangling
$ hg commit -m 'add symlink'
$ hg tip -v
changeset: 0:cabd88b706fc
tag: tip
user: test
date: Thu Jan 01 00:00:00 1970 +0000
files: dangling
description:
add symlink
$ hg manifest --debug
2564acbe54bbbedfbf608479340b359f04597f80 644 @ dangling
$ readlink.py dangling
dangling -> nothing
$ rm dangling
$ ln -s void dangling
$ hg commit -m 'change symlink'
$ readlink.py dangling
dangling -> void
modifying link
$ rm dangling
$ ln -s empty dangling
$ readlink.py dangling
dangling -> empty
reverting to rev 0:
$ hg revert -r 0 -a
reverting dangling
$ readlink.py dangling
dangling -> nothing
backups:
$ readlink.py *.orig
dangling.orig -> empty
$ rm *.orig
$ hg up -C
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
copies
$ hg cp -v dangling dangling2
copying dangling to dangling2
$ hg st -Cmard
A dangling2
dangling
$ readlink.py dangling dangling2
dangling -> void
dangling2 -> void
Issue995: hg copy -A incorrectly handles symbolic links
$ hg up -C
0 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ mkdir dir
$ ln -s dir dirlink
$ hg ci -qAm 'add dirlink'
$ mkdir newdir
$ mv dir newdir/dir
$ mv dirlink newdir/dirlink
$ hg mv -A dirlink newdir/dirlink
$ cd ..