Alessandro Lai

Results 270 comments of Alessandro Lai

Using an extension is definitely the way to go: monkey-patching code would make many users recoil and flee from such a tool, since it would mess with a lot of...

As said, changing the default is not feasible. Doc issue reported as https://github.com/getsentry/sentry-docs/issues/4579

> I will make sure send_attempts gets documented but I don't think the SDK is handling the 429 correctly either :) No we don't, you're right! Opened #1279 for that.

> **How does the user get to it?** > > * getting the sidecar and starting it up > > * in PHP it could download the code and spawn...

If the end users will still have the option to deploy Relay manually, that would be a great strategy IMHO. In that way, I can make the conscious trade-off between...

IIRC this was intentional, to avoid creating a loop where, if Sentry is hooked to the logger, a transport failure triggers another event, that triggers a new failure transport... @ste93cry...

Currently, there are 3 methods that register an handler of some kind: exceptions, errors, fatals * https://github.com/getsentry/sentry-php/blob/51af981636bbe97fe9271dab76819d02593c3191/src/ErrorHandler.php#L194 * https://github.com/getsentry/sentry-php/blob/51af981636bbe97fe9271dab76819d02593c3191/src/ErrorHandler.php#L131 * https://github.com/getsentry/sentry-php/blob/51af981636bbe97fe9271dab76819d02593c3191/src/ErrorHandler.php#L167 All 3 of them are used ONLY in the...

I have two opposing views on this: * on one hand, it would mitigate https://github.com/getsentry/sentry-symfony/issues/421 by avoiding interfering with the error handler * OTOH, the fact that you avoid registering...

IMHO it should be solved on the project side: autoloaders are nestable, so you should have a single one that's able to reach every piece of code\class.

This should wait for #1275