Since this overwrites (I noticed the "1" at the end), the server
operator cannot force a particular protocol with their server
configuration, no?
Would a weaker form to use 0 (set if there isn't any, but keep the
value if somebody else already has set it) work OK? Would that have
a downside?
Since this overwrites (I noticed the "1" at the end), the server
operator cannot force a particular protocol with their server
configuration, no?
Would a weaker form to use 0 (set if there isn't any, but keep the
value if somebody else already has set it) work OK? Would that have
a downside?
Yeah, I wondered about that while writing it, but struggled to think of
a reason why the server operator would want to set it at all.
But maybe as a workaround for some misbehaving client (e.g., recognizing
it by a user-agent header). Or we could even perhaps use this in our
tests to test the v2-client-to-v0-server fallback behavior.
I'll re-roll with that change, plus some documentation changes adapted
to this new approach.
-Peff
From: Jeff King <hidden> Date: 2021-09-10 14:02:36
On Fri, Sep 10, 2021 at 07:39:54AM -0400, Jeff King wrote:
I'll re-roll with that change, plus some documentation changes adapted
to this new approach.
Here's what I came up with. I think this should replace both
jk/http-backend-handle-proto-header and kr/doc-webserver-config-for-v2.
The latter does give some specific nginx tips which I didn't carry over,
but they shouldn't be necessary after the change in http-backend. If we
do want to include them, they can be mentioned as optional if we later
add an nginx example config to the http-backend manpage.
[1/5]: t5551: test v2-to-v0 http protocol fallback
[2/5]: http-backend: handle HTTP_GIT_PROTOCOL CGI variable
[3/5]: docs/http-backend: mention v2 protocol
[4/5]: docs/git: discuss server-side config for GIT_PROTOCOL
[5/5]: docs/protocol-v2: point readers transport config discussion
Documentation/git-http-backend.txt | 26 ++++++++++++++++++++++++-
Documentation/git-upload-pack.txt | 8 ++++++++
Documentation/git.txt | 15 ++++++++++++++
Documentation/technical/protocol-v2.txt | 8 +++++++-
http-backend.c | 4 ++++
t/lib-httpd/apache.conf | 7 +++++--
t/t5551-http-fetch-smart.sh | 9 +++++++++
7 files changed, 73 insertions(+), 4 deletions(-)
-Peff
From: Jeff King <hidden> Date: 2021-09-10 14:04:46
Since we use the v2 protocol by default, the connection of a v2 client
to a v2 server is well covered by the test suite. And with the
GIT_TEST_PROTOCOL_VERSION knob, we can easily test a v0 client
connecting to a v2-aware server (which will then just speak v0). But we
have no regular tests that a v2 client, when encountering a non-v2-aware
server, will correctly fall back to using v0.
In theory this is a job for the cross-version tests in t/interop, but:
- they cover only git:// and file:// clones
- they are not part of the usual test suite, so nobody ever runs them
anyway
Since using v2 over http requires configuring the web server to pass
along the Git-Protocol header, we can easily create a situation where
the server does not respect the v2 probe, and the conversation falls
back to v0.
This works just fine. This new test is not about fixing any particular
bug, but just making sure that the system works (and continues to work)
as expected.
Signed-off-by: Jeff King <redacted>
---
Not strictly necessary for this series, but it seemed like a good chance
to beef up test coverage, and make sure the setenv() overwrite flag in
the next patch was set sensibly. :)
t/lib-httpd/apache.conf | 5 +++++
t/t5551-http-fetch-smart.sh | 9 +++++++++
2 files changed, 14 insertions(+)
@@ -558,4 +558,13 @@ test_expect_success 'http auth forgets bogus credentials' 'expect_askpassbothuser@host'+test_expect_success'client falls back from v2 to v0 to match server''+GIT_TRACE_PACKET=$PWD/trace\+GIT_TEST_PROTOCOL_VERSION=2\+gitclone$HTTPD_URL/smart_v0/repo.gitrepo-v0&&+# check for v0; there the HEAD symref is communicated in the capability+# line; v2 uses a different syntax on each ref advertisement line+grepsymref=HEAD:refs/heads/trace+'+ test_done
From: Jeff King <hidden> Date: 2021-09-10 14:05:50
When a client requests the v2 protocol over HTTP, they set the
Git-Protocol header. Webservers will generally make that available to
our CGI as HTTP_GIT_PROTOCOL in the environment. However, that's not
sufficient for upload-pack, etc, to respect it; they look in
GIT_PROTOCOL (without the HTTP_ prefix).
Either the webserver or the CGI is responsible for relaying that HTTP
header into the GIT_PROTOCOL variable. Traditionally, our tests have
configured the webserver to do so, but that's a burden on the server
admin. We can make this work out of the box by having the http-backend
CGI copy the contents of HTTP_GIT_PROTOCOL to GIT_PROTOCOL.
There are no new tests here. By removing the SetEnvIf line from our
test Apache config, we're now relying on this behavior of http-backend
to trigger the v2 protocol there (and there are numerous tests that fail
if this doesn't work).
There is one subtlety here: we copy HTTP_GIT_PROTOCOL only if there is
no existing GIT_PROTOCOL variable. That leaves the webserver admin free
to override the client's decision if they choose. This is unlikely to be
useful in practice, but is more flexible. And indeed, it allows the
v2-to-v0 fallback test added in the previous commit to continue working.
Signed-off-by: Jeff King <redacted>
---
http-backend.c | 4 ++++
t/lib-httpd/apache.conf | 2 --
2 files changed, 4 insertions(+), 2 deletions(-)
From: Jeff King <hidden> Date: 2021-09-10 14:09:16
Historically there was a little bit of configuration needed at the
webserver level in order to get the client's v2 protocol probes to Git.
But when we introduced the v2 protocol, we never documented these.
As of the previous commit, this should mostly work out of the box
without any explicit configuration. But it's worth documenting this to
make it clear how we expect it to work, especially in the face of
webservers which don't provide all headers over the CGI interface. Or
anybody who runs across this documentation but has an older version of
Git (or _used_ to have an older version, and wonders why they still have
a SetEnvIf line in their Apache config and whether it's still
necessary).
Signed-off-by: Jeff King <redacted>
---
I used the vague "older versions of Git" here, which might not be as
helpful as we could be. But it seemed presumptuous to say "before
2.34", since we don't know the fate of the topic, nor even that we for
sure will call the next version 2.34.
I guess we could swap out "older versions" for "before 2.34" later,
after the topic graduates and the release gets closer, but it does seem
like something that we'd easily forget to do. I dunno.
Documentation/git-http-backend.txt | 26 +++++++++++++++++++++++++-
1 file changed, 25 insertions(+), 1 deletion(-)
@@ -16,7 +16,9 @@ A simple CGI program to serve the contents of a Git repository to Git clients accessing the repository over http:// and https:// protocols. The program supports clients fetching using both the smart HTTP protocol and the backwards-compatible dumb HTTP protocol, as well as clients-pushing using the smart HTTP protocol.+pushing using the smart HTTP protocol. It also supports Git's+more-efficient "v2" protocol if properly configured; see the+discussion of `GIT_PROTOCOL` in the ENVIRONMENT section below. It verifies that the directory has the magic file "git-daemon-export-ok", and it will refuse to export any Git directory
@@ -77,6 +79,18 @@ Apache 2.x:: SetEnv GIT_PROJECT_ROOT /var/www/git SetEnv GIT_HTTP_EXPORT_ALL ScriptAlias /git/ /usr/libexec/git-core/git-http-backend/++# This is not strictly necessary using Apache and a modern version of+# git-http-backend, as the webserver will pass along the header in the+# environment as HTTP_GIT_PROTOCOL, and http-backend will copy that into+# GIT_PROTOCOL. But you may need this line (or something similar if you+# are using a different webserver), or if you want to support older Git+# versions that did not do that copying.+#+# Having the webserver set up GIT_PROTOCOL is perfectly fine even with+# modern versions (and will take precedence over HTTP_GIT_PROTOCOL,+# which means it can be used to override the client's request).+SetEnvIf Git-Protocol ".*" GIT_PROTOCOL=$0 ---------------------------------------------------------------- + To enable anonymous read access but authenticated write access,
@@ -264,6 +278,16 @@ a repository with an extremely large number of refs. The value can be specified with a unit (e.g., `100M` for 100 megabytes). The default is 10 megabytes.+Clients may probe for optional protocol capabilities (like the v2+protocol) using the `Git-Protocol` HTTP header. In order to support+these, the contents of that header must appear in the `GIT_PROTOCOL`+environment variable. Most webservers will pass this header to the CGI+via the `HTTP_GIT_PROTOCOL` variable, and `git-http-backend` will+automatically copy that to `GIT_PROTOCOL`. However, some webservers may+be more selective about which headers they'll pass, in which case they+need to be configured explicitly (see the mention of `Git-Protocol` in+the Apache config from the earlier EXAMPLES section).+ The backend process sets GIT_COMMITTER_NAME to '$REMOTE_USER' and GIT_COMMITTER_EMAIL to '$\{REMOTE_USER}@http.$\{REMOTE_ADDR\}', ensuring that any reflogs created by 'git-receive-pack' contain some
From: Jeff King <hidden> Date: 2021-09-10 14:09:58
The v2 protocol requires that the GIT_PROTOCOL environment variable gets
passed around, but we don't have any documentation describing how this
is supposed to work. In particular, we need to note what server admins
might need to configure to make things work.
The definition of the GIT_PROTOCOL variable is probably the best place
for this, since:
- we deal with multiple transports (ssh, http, etc).
Transport-specific documentation (like the git-http-backend bits
added in the previous commit) are helpful for those transports, but
this gives a broader overview. Plus we do not have a specific
transport endpoint program for ssh, so this is a reasonable place to
mention it.
- the server side of the protocol involves multiple programs. For now,
upload-pack is the only endpoint which uses GIT_PROTOCOL, but that
will likely expand in the future. We're better off with a central
discussion of what the server admin might need to do. However, for
discoverability, this patch adds a pointer from upload-pack's
documentation.
Signed-off-by: Jeff King <redacted>
---
Documentation/git-upload-pack.txt | 8 ++++++++
Documentation/git.txt | 15 +++++++++++++++
2 files changed, 23 insertions(+)
@@ -44,6 +44,14 @@ OPTIONS <directory>:: The repository to sync from.+ENVIRONMENT+-----------++`GIT_PROTOCOL`::+ Internal variable used for handshaking the wire protocol. Server+ admins may need to configure some transports to allow this+ variable to be passed. See the discussion in linkgit:git[1].+ SEE ALSO -------- linkgit:gitnamespaces[7]
@@ -894,6 +894,21 @@ for full details. Contains a colon ':' separated list of keys with optional values 'key[=value]'. Presence of unknown keys and values must be ignored.+++Note that servers may need to be configured to allow this variable to+pass over some transports. It will be propagated automatically when+accessing local repositories (i.e., `file://` or a filesystem path), as+well as over the `git://` protocol. For git-over-http, it should work+automatically in most configurations, but see the discussion in+linkgit:git-http-backend[1]. For git-over-ssh, the ssh server may need+to be configured to allow clients to pass this variable (e.g., by using+`AcceptEnv GIT_PROTOCOL` with OpenSSH).+++This configuration is optional. If the variable is not propagated, then+clients will fall back to the original "v0" protocol (but may miss out+on some performance improvements or features). This variable currently+only affects clones and fetches; it is not yet used for pushes (but may+be in the future). `GIT_OPTIONAL_LOCKS`:: If set to `0`, Git will complete any requested operation without
From: Jeff King <hidden> Date: 2021-09-10 14:10:22
We recently added tips for server admins to configure various transports
to support v2's GIT_PROTOCOL variable. While the protocol-v2 document is
pretty technical and not of interest to most admins, it may be a
starting point for them to figure out how to turn on v2. Let's put some
pointers from there to the other documentation.
Signed-off-by: Jeff King <redacted>
---
Documentation/technical/protocol-v2.txt | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
@@ -42,7 +42,8 @@ Initial Client Request In general a client can request to speak protocol v2 by sending `version=2` through the respective side-channel for the transport being used which inevitably sets `GIT_PROTOCOL`. More information can be-found in `pack-protocol.txt` and `http-protocol.txt`. In all cases the+found in `pack-protocol.txt` and `http-protocol.txt`, as well as the+`GIT_PROTOCOL` definition in `git.txt`. In all cases the response from the server is the capability advertisement. Git Transport
@@ -58,6 +59,8 @@ SSH and File Transport When using either the ssh:// or file:// transport, the GIT_PROTOCOL environment variable must be set explicitly to include "version=2".+The server may need to be configured to allow this environment variable+to pass. HTTP Transport ~~~~~~~~~~~~~~
@@ -81,6 +84,9 @@ A v2 server would reply: Subsequent requests are then made directly to the service `$GIT_URL/git-upload-pack`. (This works the same for git-receive-pack).+The server may need to be configured to pass this header's contents via+the `GIT_PROTOCOL` variable. See the discussion in `git-http-backend.txt`.+ Capability Advertisement ------------------------