update qepcad to B 1.74

[dimpase]

As noted in #36835, we missed an update to qepcad done in 2021. It has made it into Debian and Fedora,
and we use Debian's tarball and patches (most of the latter are updated versions of our B 1.72 patches)

We removed 'B' from the version number to accommodate the tarfile name from Debian.

We also update saclib to the latest version

Fixes #36831
Fixes #36835

[dimpase]

qepcad tests pass on x86_64 Linux and on macOS M1 (arm64)

[mkoeppe]

Testing at https://github.com/mkoeppe/sage/actions/runs/7135923618

[dimpase]

I took the liberty to promote qepcad to optional. It's in a reasonably good shape on all platforms, sort of maintained upstream, etc.

[mkoeppe]

Trivial doctest failure can be seen in the CI Linux Incremental runs, e.g. ubuntu-focal-standard https://github.com/sagemath/sage/actions/runs/7141754593/job/19449654300?pr=36837#step:11:7027

ile "src/sage/interfaces/qepcad.py", line 1092, in sage.interfaces.qepcad.Qepcad._parse_answer_stats
Failed example:
    stats
Expected:
    '-----------------------------------------------------------------------------\r\n0 Garbage collections, 0 Cells and 0 Arrays reclaimed, in 0 milliseconds.\r\n492514 Cells in AVAIL, 500000 Cells in SPACE.\r\n\r\nSystem time: 16 milliseconds.\r\nSystem time after the initialization: 4 milliseconds.\r\n-----------------------------------------------------------------------------\r\n'
Got:
    '-----------------------------------------------------------------------------\r\n0 Garbage collections, 0 Cells and 0 Arrays reclaimed, in 0 milliseconds.\r\n492273 Cells in AVAIL, 500000 Cells in SPACE.\r\n\r\nSystem time: 5 milliseconds.\r\nSystem time after the initialization: 2 milliseconds.\r\n-----------------------------------------------------------------------------\r\n'
**********************************************************************
1 item had failures:
   1 of   6 in sage.interfaces.qepcad.Qepcad._parse_answer_stats
    [343 tests, 1 failure, 30.35 s]

[mkoeppe]

I took the liberty to promote qepcad to optional. It's in a reasonably good shape on all platforms, sort of maintained upstream, etc.

Yes, that seems like a good idea.

[mkoeppe]

We removed 'B' from the version number to accommodate the tarfile name from Debian.

Good idea, that also matches how the versions are commonly represented according to repology.

[github-actions]

Documentation preview for this PR (built with commit 95b80ca; changes) is ready! 🎉

[mkoeppe]

LGTM.

[dimpase]

with a later develop branch, I see

sage -t --warn-long 14.2 --random-seed=42485267800152194636581559978378394047 src/sage/interfaces/qepcad.py
**********************************************************************
File "src/sage/interfaces/qepcad.py", line 833, in sage.interfaces.qepcad.Qepcad.?
Failed example:
    qepcad(conds, memcells=2000000) # optional - qepcad
Expected:
    2 x - 1 > 0 /\ z > 0 /\ z - y < 0 /\ 3 z + 3 y + x - 1 < 0
Got:
    finish  &
    <BLANKLINE>
**********************************************************************
1 item had failures:
   1 of   6 in sage.interfaces.qepcad.Qepcad.?

which can be fixed by

--- a/src/sage/interfaces/qepcad.py
+++ b/src/sage/interfaces/qepcad.py
@@ -833,7 +833,7 @@ class Qepcad:
                            x^2 + x*y + 3*x*z + 2*y*z + 2*z^2 - x - z < 0, \
                            -2*x + 1 < 0, -x*y - x*z - 2*y*z - 2*z^2 + z < 0, \
                            x + 3*y + 3*z - 1 < 0]
-            sage: qepcad(conds, memcells=2000000) # optional - qepcad
+            sage: qepcad(conds, memcells=3000000) # optional - qepcad
             2 x - 1 > 0 /\ z > 0 /\ z - y < 0 /\ 3 z + 3 y + x - 1 < 0
         """
         self._cell_cache = {}

[dimpase]

doctest to be fixed in #36895