Technically this "works" too (if it works properly). Let's see.

Sage-the-distro is still tested using ci-incremental.

(if you mean ci-linux-incremental) technically this is false since the test is only run on a pull request in the (relatively rare) case where some package has changed.

the put(result, False) is strange, and it was added all the way back in 9b1c246 .

Maybe try changing it to .put(result, block=True, timeout=10)? That way at least it tolerates the reader is slow, but also appropriately complain after a while if the reader malfunction.

the put(result, False) is strange, and it was added all the way back in 9b1c246 .

Maybe try changing it to .put(result, block=True, timeout=10)? That way at least it tolerates the reader is slow, but also appropriately complain after a while if the reader malfunction.

Do you understand what's going on here? The queue has a max limit of 1, because it is supposed to hold only one doctest result. I don't see why it would try to put more than one result in there. The code also doesn't seem to reuse the queue across workers/doctests.

Do you understand what's going on here? The queue has a max limit of 1, because it is supposed to hold only one doctest result. I don't see why it would try to put more than one result in there. The code also doesn't seem to reuse the queue across workers/doctests.

no idea either, but some StackOverflow post claims this may be caused when the object being passed around is too large (I can't reproduce this on my side though). In that case it sounds like blocking=True might help.

Worst case we can add some debugging there (print out what's the current length), modulo race conditions.

sage -t --long --warn-long 30.0 --random-seed=308424213503200308932561197805792472951 src/sage/dynamics/arithmetic_dynamics/projective_ds.py
Error: sage: P1.<x,y> = ProjectiveSpace(QQ,1) ## line 169 ##
sage: DynamicalSystem_projective([y, 2*x]) ## line 170 ##
Dynamical System of Projective Space of dimension 1 over Rational Field
  Defn: Defined on coordinates by sending (x : y) to
        (y : 2*x)
sage: R.<t> = QQ[] ## line 178 ##
sage: DynamicalSystem_projective(t^2 - 3) ## line 179 ##
Dynamical System of Projective Space of dimension 1 over Rational Field
  Defn: Defined on coordinates by sending (X : Y) to
        (X^2 - 3*Y^2 : Y^2)
sage: DynamicalSystem_projective(1/t^2) ## line 183 ##
Dynamical System of Projective Space of dimension 1 over Rational Field
  Defn: Defined on coordinates by sending (X : Y) to
        (Y^2 : X^2)
sage: R.<x> = PolynomialRing(QQ,1) ## line 190 ##
sage: DynamicalSystem_projective(x^2, names=['a','b']) ## line 191 ##
Dynamical System of Projective Space of dimension 1 over Rational Field
  Defn: Defined on coordinates by sending (a : b) to
        (a^2 : b^2)
sage: x,y = var('x,y')                             
Process DocTestWorker-1129:
Traceback (most recent call last):
  File "/usr/share/miniconda/envs/sage-dev/lib/python3.12/multiprocessing/process.py", line 314, in _bootstrap
    self.run()
  File "/usr/share/miniconda/envs/sage-dev/lib/python3.12/site-packages/sage/doctest/forker.py", line 2301, in run
    task(self.options, self.outtmpfile, msgpipe, self.result_queue,
  File "/usr/share/miniconda/envs/sage-dev/lib/python3.12/site-packages/sage/doctest/forker.py", line 2637, in __call__
    result_queue.put(result, block=True, timeout=10)
  File "/usr/share/miniconda/envs/sage-dev/lib/python3.12/multiprocessing/queues.py", line 90, in put
    raise Full
queue.Full

it's not helping either. I'll see if I can reproduce the issue locally…

edit: I cannot. debugging with #40161

edit: feels like a weird bug in multiprocessing.Queue to me. By debugging I determine that qsize() returns 1, yet get() fails.

Now continue debugging.

sanity check, is test-new currently running long tests? (it's supposed to be.)

Documentation preview for this PR (built with commit a5aa2b4; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

Needs #39872 to be fixed.