Should I report codepaths leading to comparing incomparable objects?

[Georgi Guninski]

I reported several bugs about Graph.vertices() and have
the impression that sage devs are aware of the problem and
don't want bug reports about this.

I have another testcase not related to vertices(), should
I report it?

[Dima Pasechnik]

On Sun, Jul 2, 2023 at 4:59 PM Georgi Guninski <ggun...@gmail.com> wrote:
>
> I reported several bugs about Graph.vertices() and have
> the impression that sage devs are aware of the problem and
> don't want bug reports about this.

What was the problem there? IMHO that was working as expected - you cannot
build graph with incomparable vertices.
It's a very minor inconvenience.

>
> I have another testcase not related to vertices(), should
> I report it?

yes, please.


>
> --
> 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 sage-devel+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CAGUWgD9Ynqi2ddLrd%3Ds9JtUGV34DG%2BGyfLmp5sa-L-dHMDKMJw%40mail.gmail.com.

[Georgi Guninski]

You definitely can construct graph with incomparable vertices
and they worked as expected in python2. python3 regressed them
and devs are doing workarounds.

The new testace is:
G=Graph([("A",1)])
G.feedback_vertex_set()
===
File /home/sc_serv/sage/src/sage/graphs/connectivity.pyx:287, in
sage.graphs.connectivity.connected_component_containing_vertex()
285
286 if sort:
--> 287 c.sort()
288 return c

[David Coudert]

Thanks for reporting this issue. I have opened an issue (https://github.com/sagemath/sage/issues/35889) and will propose a fix asap.