Awesome. Since this is a release candidate, I am going to try and reproduce the same FCL memory leak I detected in csvdataset here where the variable
FParser isn't being freed correctly. I'd rather the release not contain this same bug.
You can read about my research into this leak, and how I managed to eventually resolve it by reading my last post on this thread:
https://forum.lazarus.freepascal.org/index.php/topic,49096.0.htmlAnd if this isn't a bug, then please let me know how I should be freeing FParser in my own code. I will update this post once I test the RC against the leak mentioned in my thread.
Looks like this memory leak is addressed in FPC 3.2.0 RC1! Thank you guys! Glad I wasn't the first one to notice this issue, however, it would have been nice if someone in the forums let me know it was an issue, and is going to be resolved in the next release. After I did a test compile, heaptrc no longer showed a memory leak. I checked the source code in FPC trunk, and I see that a
FreeAndNil(FParser); was definitely added to the Destroy destructor, resolving the memory leak.