[issue29137] Fix fpectl-induced ABI breakage

[Nathaniel Smith]

Nathaniel Smith added the comment:

@Dima:
> @njs: to point out that usefulness of this module is not just wishful thinking. I just used it to locate, up to the line in a Python extension module written in C, a bug in Sagemath (that has perhaps 20 FPU-using extensions, some of them as large as numpy).  (Without using it we were pulling out our hair for weeks over this)

That's pretty cool :-). But from skimming your link, it sounds like it would have been sufficient in your case to add a call to "fesetmask(FP_X_INV)" using C or Cython or ctypes (or directly in gdb), and then running the program under gdb to get a backtrace where the SIGFPE was delivered? Is that correct? Or did your debugging depend on the specific fpectl machinery for responding to that signal?

> PS. I would volunteer to fix it and maintain it, assuming I have some modest funding to support such an activity.

I'm not personally aware of any funding sources for this, if that's the question.

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue29137>
_______________________________________