-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
Conversation
Codecov Report
@@ Coverage Diff @@
## debian #4285 +/- ##
==========================================
+ Coverage 66.53% 66.56% +0.03%
==========================================
Files 302 302
Lines 29919 29895 -24
Branches 4900 4893 -7
==========================================
- Hits 19906 19900 -6
+ Misses 8563 8550 -13
+ Partials 1450 1445 -5
Continue to review full report at Codecov.
|
Context for interested readers: this PR is based on a branch which copies the existing packaging from https://github.com/matrix-org/package-synapse-debian/tree/debian/debian. It supercedes #4212 and matrix-org/package-synapse-debian#39. The idea for now is that, because the changes here are a substantial rewrite, we will keep the existing python 2 packaging (in https://github.com/matrix-org/package-synapse-debian), and offer the packages built from this source as alternatives for those wanting to try out python 3. (at some point in the future the old python 2 packages will be deprecated and we'll do the debian package shenanigans to make these packages replace them.) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ok so lots of comments here which range from "nitpicky things about trailing NLs" to "I think you should do X" to "This may be entirely valid but I don't quite understand why we are doing it". I'll let you figure out which is which :)
debian/install
Outdated
@@ -1,2 +1,3 @@ | |||
debian/homeserver.yaml etc/matrix-synapse | |||
debian/log.yaml etc/matrix-synapse | |||
debian/matrix-synapse.service lib/systemd/system |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
fwiw we can do it the "right way" when we rename the package, I didn't find a way for dh_systemd to install it under a name that wasn't the package name. :/
I'm totally failing to follow your logic here. We aren't using dh_systemd* at all currently (because we are on debhelper compat 9, which doesn't enable it by default, and instead dh_installinit
does the installation of the service file).
Either way, the solution is just to add a --name
argument in the relevant override_dh_*
sections.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I just took a quick look and haven't tested it out, but it looks plausible.
So that the container build doesn't modify the original source
... so that dpkg-source generates a source package with the right name
so that we don't have to invoke sudo
... so that dput works right
the stdlib has distutils on stretch
... to avoid problems running the tests with -j
We're supposed to match the debhelper compat level with our debhelper dependency. Xenial only has debhelper 9, so we should use that. I think the reason for this is basically that compat level 10 was considered unstable until debhelper 10 was released - basically we're getting some experimental pre-alpha version of compat level 10 by asking for it in xenial. It may be fine, or it may not. Let's play safe.
we don't really need these.
57cea1c
to
f4dbe67
Compare
These just emit warnings, as of #4290
f4dbe67
to
f0176b6
Compare
No description provided.