The git-remote-curl backend detects if the remote server supports
the git-receive-pack service, and if so, runs git-send-pack in a
pipe to dump the command and pack data as a single POST request.
The advertisements from the server that were obtained during the
discovery are passed into git-send-pack before the POST request
starts. This permits git-send-pack to operate largely unmodified.
For smaller packs (those under 1 MiB) a HTTP/1.0 POST with a
Content-Length is used, permitting interaction with any server.
The 1 MiB limit is arbitrary, but is sufficent to fit most deltas
created by human authors against text sources with the occasional
small binary file (e.g. few KiB icon image). The configuration
option http.postBuffer can be used to increase (or shink) this
buffer if the default is not sufficient.
For larger packs which cannot be spooled entirely into the helper's
memory space (due to http.postBuffer being too small), the POST
request requires HTTP/1.1 and sets "Transfer-Encoding: chunked".
This permits the client to upload an unknown amount of data in one
HTTP transaction without needing to pregenerate the entire pack
file locally.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
CC: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This removes the last parameter of recv_sideband, by which the callers
told which channel bands #2 and #3 should be written to.
Sayeth Shawn Pearce:
The definition of the streams in the current sideband protocol
are rather well defined for the one protocol that uses it,
fetch-pack/receive-pack:
stream #1: pack data
stream #2: stderr messages, progress, meant for tty
stream #3: abort message, remote is dead, goodbye!
Since both callers of the function passed 2 for the parameter, we hereby
remove it and send bands #2 and #3 to stderr explicitly using fprintf.
This has the nice side-effect that these two streams pass through our
ANSI emulation layer on Windows.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Acked-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently the code looks for line break characters in order to prepend
"remote: " to every line received as many lines can be sent in a single
chunk. However the opposite might happen too, i.e. a single message
line split amongst multiple chunks. This patch adds support for the
later case to avoid displays like:
remote: Compressing objeremote: cts: 100% (313/313), done.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "clear to end of line" sequence is used to nicely output the progress
indicator without leaving garbage on the terminal. However, this works
only on ANSI capable terminals. We use the same check as in color.c to
find out whether the terminal supports this feature and use a workaround
(a few spaces in a row) if it does not.
[jc: as an old fashoned git myself, and given the fact that the
possible prefix and suffix are small number of short constant strings,
I actually prefer a simpler-and-more-stupid approach. This is with
Nico's clean-up.]
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It is possible for the remote summary line to be displayed over the
local progress display line, and therefore that local progress gets
bumped to the next line. However, if the progress line is long enough,
it might not be entirely overwritten by the remote summary line. This
creates a messed up display such as:
remote: Total 310 (delta 160), reused 178 (delta 112)iB/s
Receiving objects: 100% (310/310), 379.98 KiB | 136 KiB/s, done.
So we have to clear the screen line before displaying the remote message
to make sure the local progress is not visible anymore on the first
line.
Yet some Git versions on the remote side might be sending updates to the
same line and terminate it with \r, and a separate packet with a single
\n might be sent later when the progress display is done. This means
the screen line must *not* be cleared in that case.
Since the sideband code already has to figure out line breaks in the
received packet to properly prepend the "remote:" prefix, we can easily
determine if the remote line about to be displayed is empty. Only when
it is not then a proper suffix is inserted before the \r or \n to clear
the end of the screen line.
Also some magic constants related to the prefix length have been
replaced with a variable, making it similar to the suffix length
handling. Since gcc is smart enough to detect that the variable is
constant there is no impact on the generated code.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A single sideband packet may sometimes contain multiple lines of progress
messages, but we prepend "remote: " only to the whole buffer which creates
a messed up display in that case. Make sure that the "remote: " prefix
is applied to every remote lines.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
It has been a few times that I ended up with such a confusing display:
|remote: Generating pack...
|remote: Done counting 17 objects.
|remote: Result has 9 objects.
|remote: Deltifying 9 objects.
|remote: 100% (9/9) done
|remote: Unpacking 9 objects
|Total 9, written 9 (delta 8), reused 0 (delta 0)
| 100% (9/9) done
The confusion can be avoided in most cases by writing the remote message
in one go to prevent interleacing with local messages. The buffer
declaration has been moved inside recv_sideband() to avoid extra string
copies.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The server side support; this is just the very low level, and the
caller needs to know which band it wants to send things out.
Signed-off-by: Junio C Hamano <junkio@cox.net>
(cherry picked from b786552b67878c7780c50def4c069d46dc54efbe commit)
This moves the receiver side of the sideband support from
fetch-clone.c to sideband.c and its header file, so that
archiver protocol can use it.
Signed-off-by: Junio C Hamano <junkio@cox.net>
