Open StefanSalewski opened 6 years ago
Update: After latest RTree commit new secure crash position is:
> def rtree.nim:584:12
Segmentation fault (core dumped)
I couldn't reproduce the segfault on v0.19.9. Can you confirm that segfault is gone?
Seems the critical line numbers are moving. I am using my NEd editor as a detector, as it is designed to crash when nimsuggest crashes. With latest nim/nimsuggest one crash position seems to be line 567, see
$ nim -v
Nim Compiler Version 0.19.9 [Linux: amd64]
Compiled at 2019-03-28
Copyright (c) 2006-2019 by Andreas Rumpf
git hash: e904b3f952011ad9531b0923c2000398373af687
active boot switches: -d:release
stefan@nuc ~/RTree $ nimsuggest -v
Nim Compiler Version 0.19.9 [Linux: amd64]
Compiled at 2019-03-28
Copyright (c) 2006-2019 by Andreas Rumpf
git hash: e904b3f952011ad9531b0923c2000398373af687
active boot switches: -d:release --gc:markAndSweep
stefan@nuc ~/RTree $ nimsuggest --stdin rtree.nim
> def rtree.nim:567:12
../lib/system/fatal.nim(37) sysFatal
Error: unhandled exception: /home/stefan/Nim/compiler/sigmatch.nim(366, 13) `false` cannot resolve type: H[condenseTree.M, condenseTree.D, condenseTree.RT, condenseTree.LT] [AssertionError]
SIGABRT: Abnormal termination.
But don't waste your time, I guess there are more important issues available, RTree modules uses generics, that may be too hard for nimsuggest.
OH, I just noticed that rtree.nim does not compile with latest nim devel:
rtree.nim(65, 20) Error: undeclared identifier: 'SomeReal'
Will fix that soon. maybe saturday.
Just fixed compile error:
elif r[0].a is SomeFloat: # SomeReal:
But makes no difference, crash on line 567. Will fix the compiler warnings soon, and maybe will make a nimble package of it later.
@narimiran Note that nimsuggest has special logic that tries to help within generic code that nim check
does not do.
I can't replicate this with the latest nimsuggest.
@StefanSalewski can you still replicate it?