GenericBackend: Fix doctest of add_linear_constraint_vector

[mkoeppe]

The doctest of add_linear_constraint_vector from generic_backend.pyx (which is never run for any real backend!), when applied to COIN or GLPK leads to segfaults:

sage: coeffs = ([0, vector([1, 2])], [1, vector([2, 3])])
sage: upper = vector([5, 5])
sage: lower = vector([0, 0])
sage: from sage.numerical.backends.generic_backend import get_solver
sage: p = get_solver(solver = "Coin")      # optional - cbc
sage: p.add_linear_constraint_vector(2, coeffs, lower, upper, 'foo')
------------------------------------------------------------------------
0   signals.so                          0x0000000109df05c5 print_backtrace + 37
------------------------------------------------------------------------
Unhandled SIGSEGV: A segmentation fault occurred.
This probably occurred because a *compiled* module has a bug
in it and is not properly wrapped with sig_on(), sig_off().
Python will now terminate.
------------------------------------------------------------------------
Segmentation fault: 11
$ sage
SageMath Version 7.2.beta0, Release Date: 2016-03-24 
sage: coeffs = ([0, vector([1, 2])], [1, vector([2, 3])])
sage: upper = vector([5, 5])
sage: lower = vector([0, 0])
sage: from sage.numerical.backends.generic_backend import get_solver
sage: p = get_solver(solver = "Coin")       # optional - cbc
sage: p.add_linear_constraint_vector(2, coeffs, lower, upper)
------------------------------------------------------------------------
0   signals.so                          0x0000000109c8a5c5 print_backtrace + 37
------------------------------------------------------------------------
Unhandled SIGSEGV: A segmentation fault occurred.

CC: @dimpase @videlec

Component: numerical

Author: Matthias Koeppe

Branch/Commit: 690b1a5

Reviewer: Dima Pasechnik

Issue created by migration from https://trac.sagemath.org/ticket/20326

[dimpase]

comment:1

with the second example I only get segfault with Coin - with other backends I get gracefully handled errors.
(on a branch with lots of recent positively reviewed LP-related tickets)

[mkoeppe]

comment:2

The COIN segfaults are addressed in #20360.

[mkoeppe]

Branch: u/mkoeppe/genericbackend__fix_doctest_of_add_linear_constraint_vector

[mkoeppe]

Commit: dea0705

[mkoeppe]

New commits:

dea0705

Fix #20326: GenericBackend: Fix doctest of add_linear_constraint_vector

[dimpase]

comment:5

solver.add_variables(2) throws a NotImplementedError immediately, what's the point of the last commit then?

sage -t src/sage/numerical/backends/generic_backend.pyx
**********************************************************************
File "src/sage/numerical/backends/generic_backend.pyx", line 402, in sage.numerical.backends.generic_backend.GenericBackend.add_linear_constraint_vector
Failed example:
    solver.add_variables(2)
...
    NotImplementedError

[dimpase]

comment:6

if you want to keep these as a template you'd mark them # not tested

[sagetrac-git]

Changed commit from dea0705 to b932b89

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

b932b89

add_linear_constraint_vector: Fix fix of doctest

[mkoeppe]

comment:8

Sorry! I tested the wrong file after making this change, so I did not catch this test failure.

Yes, the idea is to make generic_backend.pyx suitable for cut and paste again.
But don't worry, I won't make many more tickets like this one; I have been making progress with #20323.

[dimpase]

comment:9

Isn't it more in line with the rest of the code to tag these statements by # optional - Nonexistent_LP_solver ? Then you don't need to insert Tracebacks...

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

b3a14bb	Fix #20326: GenericBackend: Fix doctest of add_linear_constraint_vector
690b1a5	add_linear_constraint: Use 'optional - Nonexistent_LP_solver' test as well

[sagetrac-git]

Changed commit from b932b89 to 690b1a5

[mkoeppe]

comment:11

Replying to @dimpase:

Isn't it more in line with the rest of the code to tag these statements by # optional - Nonexistent_LP_solver ? Then you don't need to insert Tracebacks...

You are right. I have now brought it in line with the doctests for the other methods. Same also for add_linear_constraint.

[dimpase]

comment:12

OK, good!

[dimpase]

Author: Matthias Koeppe

[dimpase]

Reviewer: Dima Pasechnik

[vbraun]

Changed branch from u/mkoeppe/genericbackend__fix_doctest_of_add_linear_constraint_vector to 690b1a5