Bob Weinand

Another open question is: Shall we auto-instrument pcntl_fork by default, so that traces of forks always have a pcntl_fork span as parent? Instead of being on the same level with...

So, it is looping indefinitely within that topmost zend_hash_index_del call? Otherwise, I'm not seeing any particular location which loops here in the stacktrace?

The pure action of forking should not affect this at all or require any cleanup here. In fact, that should be totally unrelated to this particular PR, but be a...

It was indeed trivial to reproduce the issue (this is with a debug build and crashes as I suspected): ``` circleci(buster):~/app$ DD_TRACE_CLI_ENABLED=1 php ../r.php /usr/local/src/php/Zend/zend_hash.c(1497) : ht=0x7f9642e89040 is already destroyed...

@jlesueur I've just merged a fix for this particular issue into master and rebased this branch on master. If you want to continue testing this PR, feel free, it's appreciated!

I've got other priorities right now, but we're definitely still planning to release this quite soon, sorry that I haven't gotten to it in time for 0.77 - but I'll...

I'm looking at this again now, adding arm64 support had a higher priority. I definitely want to finish this up very soon!

@joshdifabio basically: ```php public static function getNewIdentifier() { static $i = "a"; return $i++; } ``` I do not object per se, but perhaps for some reason implementations may want...

> If watcher ID's must be unique per-driver instance then I guess that rules out object hashes? Ah right. @joshdifabio Please PR it then, I'll probably approve it.

Note that in case we shift that responsibility onto the respective classes (e.g. the DNS drivers), it does not know when a loop has been freed and will reference the...