Make pari_sig_on() macros accessible

[jdemeyer]

Up to now, the functions sig_on() and friends were redefined as _pari_sig_on() for use in sage/libs/pari.pyx. This is a hack and should be fixed by providing (and exporting) proper pari_sig_on() macros.

This ticket doesn't claim to fix all problems with PARI error handling, but at least it should improve the situation.

Also fix #10126.

Component: c_lib

Author: Jeroen Demeyer

Reviewer: Peter Bruin

Merged: sage-5.12.beta5

Issue created by migration from https://trac.sagemath.org/ticket/15124

[jdemeyer]

Description changed:

--- 
+++ 
@@ -1 +1,3 @@
 Up to now, the functions `sig_on()` and friends were redefined as `_pari_sig_on()` for use in `sage/libs/pari.pyx`. This is a hack and should be fixed by providing (and exporting) proper `pari_sig_on()` macros.
+
+Also try to fix #10126.

[jdemeyer]

Description changed:

--- 
+++ 
@@ -1,3 +1,5 @@
 Up to now, the functions `sig_on()` and friends were redefined as `_pari_sig_on()` for use in `sage/libs/pari.pyx`. This is a hack and should be fixed by providing (and exporting) proper `pari_sig_on()` macros.
 
-Also try to fix #10126.
+This ticket doesn't claim to fix all problems with PARI error handling, but at least it should improve the situation.
+
+Also fix #10126.

[pjbruin]

Reviewer: Peter Bruin

[pjbruin]

comment:4

This looks good to me. I was just thinking about this and had a very similar patch in mind (but with different names than pari_sig_on()/pari_sig_off(), which I still find slightly misleading).

I will run doctests and then give this a positive review - conditional on passing the tests of course, but not on the names for the macros. 8-)

[jdemeyer]

comment:5

Replying to @pjbruin:

This looks good to me. I was just thinking about this and had a very similar patch in mind (but with different names than pari_sig_on()/pari_sig_off(), which I still find slightly misleading).

Would pari_catch_sig_on(), pari_catch_sig_off() be a good compromise?

[pjbruin]

comment:6

Replying to @jdemeyer:

Would pari_catch_sig_on(), pari_catch_sig_off() be a good compromise?

Yes, I think that is a very good solution. A bit longer, but more descriptive.

[jdemeyer]

comment:7

Name changed to pari_catch_sig_*

[jdemeyer]

comment:8

Attachment: 15124_pari_sig_on.patch.gz

[jdemeyer]

Author: Jeroen Demeyer

[pjbruin]

comment:9

OK, all tests passed.