Re: iota core dump

2025-04-23 Thread Hans-Peter Sorge
Hi Christian, thank you for your information. Correct - this is Fedora 42. One of those silent improvements one can run into after the release but not read about before the release. In this special case it looks like the immediate output fails: x←'0123456789abcde'⍳ '345'  x 4 5 6  '

Re: iota core dump

2025-04-23 Thread Christian Robert
Hi Hans-Peter,  you are probably on the brand new Fedora 42 ? if so, Jürgen is actually working on this problem. may take some time because it affect std::vector<> class and there are a lot of them in gnu apl. On fedora 42, the new library check for *all* out of bound access and there are s

iota core dump

2025-04-23 Thread Hans-Peter Sorge
Hi, this is with most recent SVN: ⎕AV ⍳ 'a' /usr/include/c++/15/bits/stl_vector.h:1263: std::vector<_Tp, _Alloc>::reference std::vector<_Tp, _Alloc>::operator[](size_type) [with _Tp = Token; _Alloc = std::allocator; reference = Token&; size_type = long unsigned int]: Assertion '__n < thi