Closed nikophil closed 9 months ago
Should we have TestTransport::ignoreDelayStamp()|supportDelayStamp()
methods?
Should we have TestTransport::ignoreDelayStamp()|supportDelayStamp() methods?
because I wanted support_delay_stamp
to be the de facto behavior, I didn't wanted the feature to be disabled. Since the IMO tests with delay stamp are wrongly handled for now, the support_delay_stamp
was only a BC layer for me, waiting all user to migrate. But it seems harder that I was thinking
Should we have TestTransport::ignoreDelayStamp()|supportDelayStamp()
Let's not add until asked and re-evaluate when/if that happens.
@kbond the PR is ready I think
the last problem is that I required symfony/clock
as a dev dep in order to test the behavior, but this component has never been compatible with php 8.0, so tests are failing in some cases
any idea how to mitigate this?
any idea how to mitigate this?
Not without some CI craziness (requiring symfony/clock on 8.1+). I'd be fine requiring 8.1 for this package. 8.0 is almost EOL. WDYT?
let's go for this!
@kbond most of the runs are green, any idea how to fix the deprecation?
on my computer, even when using php 8.1 + --prefer-lowest, this deprecation is not a direct one... :shrug:
on my computer, even when using php 8.1 + --prefer-lowest, this deprecation is not a direct one...
I was able to recreate locally, needed to bump phpunit min to 9.6+ (I pushed this to your branch)
Hi!
here is a suggestion for supporting the
DelayStamp
in this lib.Please, consider this PR as fully WIP, I'll add tests, and cleaner code once we're OK on some approach.
Why is it needed?
we have some kind of home-made scheduler which permits to program some actions in the future, based on delay stamp, and we would like to be able to test this functionally by making "time jumps"
example:
Some problem it may raise
&disable_retries=false
inside the dsn of their transports), if they activate this feature along with retries, they'd need to make their app sleep in order to validate the retry. Or we may deactivate the feature for the retried messages, but this sounds cumbersome: delays will be handled, but not in all cases...WDYT?