This looks like a buffering issue: The test was apparently written
under the expectation that the Python print() would get flushed to
stdout before the GMP error gets printed in the library call, but
this is untrue if Python caches the output until after calling GMP.
Hence we're seeing the output lines in the wrong order.

I suspect passing flush=True to the print() call should fix it.


On Thu, 21 Oct 2021 09:09:11 +0200, Frédéric Chapoton <[email protected]> 
wrote:
> Hello,
> 
> it seems that all patchbots currently break on a doctest in integral.pyx, see
> below. The cause is not identified so far. Could be random-seed usage,
> changes in gmp, or anything else. Any idea ?
> 
> Frédéric
> 
> File "src/sage/rings/integer.pyx", line 6434, in
> sage.rings.integer.Integer._shift_helper Failed example:
>     try:
>         print('Possible error output from gmp')
>         1 << (2^60)
>     except (MemoryError, OverflowError, RuntimeError):
>         pass
>     else:
>         print("Failed to raise exception")
> Expected:
>     Possible error output from gmp...
> Got:
>     gmp: overflow in mpz type
>     Possible error output from gmp

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/20211021154315.1b959739%40l.

Reply via email to