David,
Many thanks for the quick turn round on these issues. I have got quite a bit further. The ProofPower build is now failing as a result of the following issue:
5) In string comparisons, the leading character is being treated as negative if it has character code greater than 127:
E.g.,
"\128" < "\000";
val it = true: bool
"\128a" < "\000a";
val it = true: bool
Character comparisons are OK:
#"\128" < #"\000";
val it = false: bool
(This is on Mac OS X after building Poly/ML with Apple's Xcode tool chain.)
Regards,
Rob.
On 18 Sep 2016, at 16:36, David Matthews <David.Matthews at prolingua.co.uk> wrote:
Rob, Thanks for that and for cutting it down to something manageable. As a result I've been able to track down and fix the problem. Let me know as and when you find anything else.
Regards, David
On 17/09/2016 16:20, Rob Arthan wrote:
David,
On 17 Sep 2016, at 14:49, Rob Arthan <rda at lemma-one.com> wrote:
?
? I am getting a segfault somewhere in the ProofPower parser generator. I will report again when I have isolated that.
To the list of three issues in my previous e-mail, I can now add:
- In some circumstances a function call can lead to a segfault.
I?ve attached a short extract from the parser generator code that demonstrates the problem. If you execute the body of the function empty_non_terminals interactively with the parameter bound to the test data, then nothing goes wrong. If you call the function with the test data as parameter (as on the last line of the attached file) you get a segfault.
Regards,
Rob.
polyml mailing list polyml at inf.ed.ac.uk http://lists.inf.ed.ac.uk/mailman/listinfo/polyml