vendor: update to the latest version of ferny #19555
Conversation
allisonkarlitskaya
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
allisonkarlitskaya
force-pushed
the
new-ferny
branch
from
a56fa2d to
5f42c56
Compare
allisonkarlitskaya
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
allisonkarlitskaya
force-pushed
the
new-ferny
branch
from
5f42c56 to
40e1e2f
Compare
allisonkarlitskaya
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
allisonkarlitskaya
force-pushed
the
new-ferny
branch
from
40e1e2f to
4b4c167
Compare
allisonkarlitskaya
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
src/cockpit/superuser.py
Outdated
| @@ -66,9 +73,6 @@ async def do_connect_transport(self) -> None: | |||
| transport = await self.spawn(self.args, env, stderr=agent, start_new_session=True) | |||
|
|
|||
| if '# cockpit-bridge' in self.args: | |||
There was a problem hiding this comment.
I suppose splitting this code into two parts was done due to some changed ordering requirements? Although it's not clear to me why writing the askpass helper above would depend on an already initialized InteractionAgent -- in fact, the old ordering made more sense to me, as the agent surely depends on a working askpass script?
This ordering change could use at least some justification in the commit message.
Testing for this magic # cockpit-bridge string twice is a DRY. Not a strong veto, but it may be nicer to set a boolean flag for this, if the code split really is necessary?
There was a problem hiding this comment.
It's because the .add_handler() call on the interaction agent disappeared. All handlers have to be set at the start now, which necessitates the earlier creation of the handler.
I have an idea for how this can be done more elegantly.
This has some breaking API changes, so make adjustments as necessary. One slightly non-obvious change comes from the removal of InteractionAgent.add_handlers(), which had the potential to be racy in case a message arrived before its handler was registered. We now have to specify all handlers up-front, which means that the block for doing that needs to move up to before the transport is connected. Of course, we can only write the stage1 bootloader after the transport is online, so we need to keep that in its current location.
allisonkarlitskaya
force-pushed
the
new-ferny
branch
from
4b4c167 to
d220099
Compare
allisonkarlitskaya
temporarily deployed
to
gh-cockpituous
GitHub Actions
Inactive
|
testSudo failure is an old acquaintance, but this unexpected message is entirely novel and exciting. Retrying, and on my todo list. |
This has some breaking API changes, so make adjustments as necessary.