You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is pretty much working as intended, given that \w covers all possible unicode character lettering ranges, which then get converted to byte ranges. Doing a + after in the regex requires that we append another one of those at the end of each leaf, which can be simulated with a r"\w\w" pattern. This is bound to be slow, but it should still compile in finite time.
Given that this pattern compiled in 0.9 (as mentioned in #68), this is a regression in 0.10.
The text was updated successfully, but these errors were encountered:
The loop seems to happen inside code generator. It might not even be infinite, but my machine runs out of RAM trying to handle it (with 32GB, that really shouldn't happen).
My current hunch is that it's due to hashing and attempting to de-duplicate branches or patterns...
Related to #68.
regex = r"\w+"
makes the derive macro fall into an infinite loop. A single\w
produces a tree:This is pretty much working as intended, given that
\w
covers all possible unicode character lettering ranges, which then get converted to byte ranges. Doing a+
after in the regex requires that we append another one of those at the end of each leaf, which can be simulated with ar"\w\w"
pattern. This is bound to be slow, but it should still compile in finite time.Given that this pattern compiled in 0.9 (as mentioned in #68), this is a regression in 0.10.
The text was updated successfully, but these errors were encountered: