FWIW, on Debian testing, once solved the above problem of upgrading from
`9.3.rc0` to 9.3.rc2`, thanks to Matthias, ptestlong gave exactly the same
results as for the last two previous releases : one timeout (parigp.py),
one cosmetic bug (gap_packages.py), one real bug (division by zero in
It's an annoying shortcoming of our build system. This particular problem
is explained in https://trac.sagemath.org/ticket/29711, on which you are
already cc'd. Overall I agree that possible remedies for failures in
incremental builds are not sufficiently explained anywhere in our
Thanks. This seems to have done it.
One more question ; how should I have divined this ? I can't remember
reading anything about this in the build instructions nor in the
developer's guide...
Le dimanche 11 avril 2021 à 00:16:59 UTC+2, matthia...@gmail.com a écrit :
> On Saturday, April 10,
On Saturday, April 10, 2021 at 1:23:53 PM UTC-7 emanuel.c...@gmail.com
wrote:
> It seems that something (what ?) is not updated in the build process
> scripts and still points to the Singular version used in 9.3.rc0. I
> suppose that rebuildig would fix the problem [...]
>
It suffices to
On Debian testing, upgrading from 9.3.rc0 to 9.3.rc2 fails at documentation
building. From install.log :
make[3]: rien à faire pour « all-sage ».
[dochtml] Traceback (most recent call last):
[dochtml] File "/usr/lib/python3.9/runpy.py", line 188, in _run_module_as_main
[dochtml]