Drop the unnecessary indirection of mc::BufferStreamFactory
#3368
Conversation
We only create a `mc::Stream` in like 1 place, the implementation of `BufferStreamFactory` is trivial, and I'm about to remove *all* the parameters from the `mc::Stream` constructor anyway. There's no need to keep this indirection (and the tests that check that we correctly call through that indirection)
RAOF
force-pushed
the
drop-buffer-stream-factory
branch
from
12d0d62 to
7789335
Compare
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #3368 +/- ##
==========================================
+ Coverage 77.52% 77.66% +0.13%
==========================================
Files 1065 1071 +6
Lines 67869 70102 +2233
==========================================
+ Hits 52618 54446 +1828
- Misses 15251 15656 +405 ☔ View full report in Codecov by Sentry. |
| @@ -33,16 +32,6 @@ namespace mc = mir::compositor; | |||
| namespace ms = mir::scene; | |||
| namespace mg = mir::graphics; | |||
|
|
|||
| std::shared_ptr<ms::BufferStreamFactory> | |||
| mir::DefaultServerConfiguration::the_buffer_stream_factory() | |||
There was a problem hiding this comment.
Do we need to do anything in the symbols.map file to remove this?
There was a problem hiding this comment.
There was a problem hiding this comment.
🎉
Hurray for extra CI checks!
|
Eh, it still seems to want |
|
Huh. Thanks, symbol checks, for picking that up. |
There was a problem hiding this comment.
I'm not against this, but with the_buffer_stream_factory() we had a customisation point.
I have no recollection of why that existed (apparently it was part of a refactoring I did 11 years ago), but are we sure hypothetical downstreams (or ourselves) we don't need it?
I think it was a configuration point for testing purposes? Our |
|
Poking, because we've waited too long for statuses to be reported |
|
Hm. What needs to be done here? #3418 includes this. The symbols file changes have been made (and those tests pass) |
We only create a
mc::Streamin like 1 place, the implementation ofBufferStreamFactoryis trivial, and I'm about to remove all the parameters from themc::Streamconstructor anyway.There's no need to keep this indirection (and the tests that check that we correctly call through that indirection)