[fpc-devel]Some uncatchable ansistring bug?

[Aleksey V. Vaneev]

Hello,

  I've encountered some strange problem (go32v2).

  I have an expression like "s := s + c" buried somewhere deep in my
  code.

  In v1.0.5 of FPC (dated 4-aug-2001) execution of program
  compiled with opt -glh gives no memory losses. But if I'm trying to
  use v1.1 of FPC (dated 19-sep-2001 up to 19-dec-2001) debug reports
  some memory losses pointing to that concat expression.
  For you info, v1.1 dated 19-aug-2001 gives no such error (much like
  v1.0.5).

  I was unable to reproduce this error in some simple manner, so I
  think you could check updates starting from 19-aug to 19-sep - maybe
  some bug (maybe in areas like ansistrings and try blocks) was seeded along
  with some improvements.

  In fact I was really mucked up when the latest Dec 1.1 verison gave me
  EXE that was not working correctly at all (in fact it just hangs
  somewhere), while same Dec 1.0.5 version gives a working executable
  (that runs without any memory loss reports).

Best regards,
 Aleksey                          mailto:picoder at sbis.komi.ru