Upgrade NumPy to 1.19.1, scipy to 1.5.2, networkx to 2.4, add pybind11 package

Question

Upgrade NumPy to 1.19.1, scipy to 1.5.2, networkx to 2.4, add pybind11 package

dimpase opened this issue 5 years ago · 155 comments

dimpase commented 5 years ago

Follow-up from the update of numpy to 1.16.6 (#29429):

We update

https://pypi.org/project/numpy/ to 1.19.1 (latest as of 2020-07-21)
https://pypi.org/project/scipy/ to 1.5.2 (latest as of 2020-07-31)
networkx to the latest release, 2.4 (latest as of 2020-07-04)
All of the above versions support Python >= 3.6.

One deprecation warning is removed here, another suppressed (from #29425),
the rest it trivial.

Tarballs: see checksums.ini [upstream_url]

Depends on #29929

Upstream: Fixed upstream, but not in a stable release.

CC: @antonio-rojas @mkoeppe @kiwifb @isuruf @slel @dcoudert @tscrim @fchapoton @alexjbest

Component: packages: standard

Author: Dima Pasechnik, Matthias Koeppe

Branch: 5d4b013

Reviewer: Travis Scrimshaw, Isuru Fernando, Dima Pasechnik

Issue created by migration from https://trac.sagemath.org/ticket/29766

Answer 1 · 2020-06-01T17:52:00.000Z

Description changed:

--- 
+++ 
@@ -1 +1,4 @@
-networkx uses scipy data types which go away in scipy 2.0, so they started to do  deprecation warnings. (fixed in networkx trunk, but not in 2.4 release)
+one deprecation warning is removed here, other suppressed on ##29425, 
+the rest it trivial.
+
+tarball: https://github.com/networkx/networkx/archive/networkx-2.4.zip

Answer 2 · 2020-06-01T17:52:00.000Z

Branch: u/dimpase/packages/networkx24

Answer 3 · 2020-06-01T17:52:00.000Z

Last 10 new commits:

`fa20111`	`build/pkgs/pybind11: New, dependency of scipy`
`a5c17a8`	`build/pkgs/cython: Update to 0.29.16`
`c7cc999`	`build/pkgs/cython/checksums.ini: Use predictable URL for upstream_url pattern`
`d54b715`	`Update Cython to 0.29.17`
`5958d38`	`build/pkgs/numpy: Update numpy to 1.18.2, back out numpy-1.10.2-no-hardcode-blas.patch (does not apply)`
`0075beb`	`Update numpy to 1.18.4`
`204213d`	`Update scipy to 1.4.1`
`14bcb1f`	`build/pkgs/pybind11: New, dependency of scipy`
`1d5ea9e`	`Merge branch 'u/mkoeppe/local_conda_forge_macos_standard__build_error_with_numpy' of trac.sagemath.org:sage into networkx24`
`2310c79`	`update networkx to version 2.4`

Answer 4 · 2020-06-01T17:52:00.000Z

Dependencies: #29425

Answer 5 · 2020-06-01T17:52:00.000Z

Commit: 2310c79

Answer 6 · 2020-06-01T17:52:00.000Z

Author: Dima Pasechnik

Answer 7 · 2020-06-01T21:04:18.000Z

comment:2

You may get more stuff if you are using scipy 1.4 with that. I patched all the deprecation I could in sage-on-gentoo
https://github.com/cschwan/sage-on-gentoo/blob/master/dev-python/networkx/files/networkx-2.4-scipy-1.4.patch

Answer 8 · 2020-06-02T09:26:23.000Z

comment:3

for scipy 1.4 (cf #29425 - which is a dependency of this ticket!) I chose to
suppress these warnings in a Pythonic way, rather than doing patching already done by upstream (just not yet released)

Answer 9 · 2020-06-03T23:55:14.000Z

comment:4

Minor (perhaps trivial) point, but pulling it from that link gave a zip file named networkx-networkx-2.4.zip, which caused it to fail without manually downloading and naming the tarball correctly.

With this ticket, I get two deprecation warnings in generic_graph.py on these two tests:

File "src/sage/graphs/generic_graph.py", line 9786, in sage.graphs.generic_graph.GenericGraph.?
Failed example:
    G.pagerank(algorithm="Scipy") # abs tol 1e-9

File "src/sage/graphs/generic_graph.py", line 9793, in sage.graphs.generic_graph.GenericGraph.?
Failed example:
    G.pagerank(algorithm="Scipy", by_weight=True) # abs tol 1e-9

Answer 10 · 2020-06-04T00:23:04.000Z

comment:5

Could you post deprecation warnings you get? Cause I thought these were fixed on #29425

Answer 11 · 2020-06-04T00:25:18.000Z

comment:6

Both of them are like this:

    doctest:warning
      File "/home/uqtscrim/sage-build/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1226, in run
        self.run_doctests()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 927, in run_doctests
        self.dispatcher.dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2011, in dispatch
        self.parallel_dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1906, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2178, in start
        super(DocTestWorker, self).start()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2150, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2478, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2527, in _run
        result = runner.run(test)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 878, in run
        return self._run(test, compileflags, out)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 680, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1104, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.graphs.generic_graph.GenericGraph.?[8]>", line 1, in <module>
        G.pagerank(algorithm="Scipy", by_weight=True) # abs tol 1e-9
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/graphs/generic_graph.py", line 9877, in pagerank
        dangling=dangling)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/networkx/algorithms/link_analysis/pagerank_alg.py", line 432, in pagerank_scipy
        S = scipy.array(M.sum(axis=1)).flatten()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/scipy/_lib/deprecation.py", line 19, in call
        stacklevel=stacklevel)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    DeprecationWarning: scipy.array is deprecated and will be removed in SciPy 2.0.0, use numpy.array instead
    doctest:warning
      File "/home/uqtscrim/sage-build/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1226, in run
        self.run_doctests()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 927, in run_doctests
        self.dispatcher.dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2011, in dispatch
        self.parallel_dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1906, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2178, in start
        super(DocTestWorker, self).start()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2150, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2478, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2527, in _run
        result = runner.run(test)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 878, in run
        return self._run(test, compileflags, out)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 680, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1104, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.graphs.generic_graph.GenericGraph.?[8]>", line 1, in <module>
        G.pagerank(algorithm="Scipy", by_weight=True) # abs tol 1e-9
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/graphs/generic_graph.py", line 9877, in pagerank
        dangling=dangling)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/networkx/algorithms/link_analysis/pagerank_alg.py", line 439, in pagerank_scipy
        x = scipy.repeat(1.0 / N, N)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/scipy/_lib/deprecation.py", line 19, in call
        stacklevel=stacklevel)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    DeprecationWarning: scipy.repeat is deprecated and will be removed in SciPy 2.0.0, use numpy.repeat instead
    doctest:warning
      File "/home/uqtscrim/sage-build/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1226, in run
        self.run_doctests()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 927, in run_doctests
        self.dispatcher.dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2011, in dispatch
        self.parallel_dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1906, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2178, in start
        super(DocTestWorker, self).start()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2150, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2478, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2527, in _run
        result = runner.run(test)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 878, in run
        return self._run(test, compileflags, out)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 680, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1104, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.graphs.generic_graph.GenericGraph.?[8]>", line 1, in <module>
        G.pagerank(algorithm="Scipy", by_weight=True) # abs tol 1e-9
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/graphs/generic_graph.py", line 9877, in pagerank
        dangling=dangling)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/networkx/algorithms/link_analysis/pagerank_alg.py", line 446, in pagerank_scipy
        p = scipy.repeat(1.0 / N, N)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/scipy/_lib/deprecation.py", line 19, in call
        stacklevel=stacklevel)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    DeprecationWarning: scipy.repeat is deprecated and will be removed in SciPy 2.0.0, use numpy.repeat instead
    doctest:warning
      File "/home/uqtscrim/sage-build/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1226, in run
        self.run_doctests()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 927, in run_doctests
        self.dispatcher.dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2011, in dispatch
        self.parallel_dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1906, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2178, in start
        super(DocTestWorker, self).start()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2150, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2478, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2527, in _run
        result = runner.run(test)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 878, in run
        return self._run(test, compileflags, out)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 680, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1104, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.graphs.generic_graph.GenericGraph.?[8]>", line 1, in <module>
        G.pagerank(algorithm="Scipy", by_weight=True) # abs tol 1e-9
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/graphs/generic_graph.py", line 9877, in pagerank
        dangling=dangling)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/networkx/algorithms/link_analysis/pagerank_alg.py", line 459, in pagerank_scipy
        is_dangling = scipy.where(S == 0)[0]
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/scipy/_lib/deprecation.py", line 19, in call
        stacklevel=stacklevel)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    DeprecationWarning: scipy.where is deprecated and will be removed in SciPy 2.0.0, use numpy.where instead
    doctest:warning
      File "/home/uqtscrim/sage-build/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1226, in run
        self.run_doctests()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/control.py", line 927, in run_doctests
        self.dispatcher.dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2011, in dispatch
        self.parallel_dispatch()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1906, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2178, in start
        super(DocTestWorker, self).start()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2150, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2478, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2527, in _run
        result = runner.run(test)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 878, in run
        return self._run(test, compileflags, out)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 680, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1104, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.graphs.generic_graph.GenericGraph.?[8]>", line 1, in <module>
        G.pagerank(algorithm="Scipy", by_weight=True) # abs tol 1e-9
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/sage/graphs/generic_graph.py", line 9877, in pagerank
        dangling=dangling)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/networkx/algorithms/link_analysis/pagerank_alg.py", line 467, in pagerank_scipy
        err = scipy.absolute(x - xlast).sum()
      File "/home/uqtscrim/sage-build/local/lib/python3.7/site-packages/scipy/_lib/deprecation.py", line 19, in call
        stacklevel=stacklevel)
      File "/home/uqtscrim/sage-build/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    DeprecationWarning: scipy.absolute is deprecated and will be removed in SciPy 2.0.0, use numpy.absolute instead

Answer 12 · 2020-06-04T00:26:37.000Z

comment:7

When I run the file as a standalone, I get

File "src/sage/graphs/generic_graph.py", line 4762, in sage.graphs.generic_graph.GenericGraph.minimum_cycle_basis
Failed example:
    sorted(g.minimum_cycle_basis(by_weight=True, algorithm='NetworkX'))
Expected:
    [[1, 2, 3], [1, 2, 3, 4], [5, 6, 7]]
Got:
    doctest:warning
      File "/home/uqtscrim/sage/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1226, in run
        self.run_doctests()
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/control.py", line 927, in run_doctests
        self.dispatcher.dispatch()
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2011, in dispatch
        self.parallel_dispatch()
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1906, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2178, in start
        super(DocTestWorker, self).start()
      File "/home/uqtscrim/sage/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/uqtscrim/sage/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/uqtscrim/sage/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/uqtscrim/sage/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/uqtscrim/sage/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/uqtscrim/sage/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2150, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2478, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2527, in _run
        result = runner.run(test)
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 878, in run
        return self._run(test, compileflags, out)
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 680, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1104, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.graphs.generic_graph.GenericGraph.minimum_cycle_basis[3]>", line 1, in <module>
        sorted(g.minimum_cycle_basis(by_weight=True, algorithm='NetworkX'))
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/sage/graphs/generic_graph.py", line 4805, in minimum_cycle_basis
        return networkx.minimum_cycle_basis(G, weight='weight')
      File "</home/uqtscrim/sage/local/lib/python3.7/site-packages/decorator.py:decorator-gen-354>", line 2, in minimum_cycle_basis
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/networkx/utils/decorators.py", line 82, in _not_implemented_for
        return not_implement_for_func(*args, **kwargs)
      File "</home/uqtscrim/sage/local/lib/python3.7/site-packages/decorator.py:decorator-gen-353>", line 2, in minimum_cycle_basis
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/networkx/utils/decorators.py", line 82, in _not_implemented_for
        return not_implement_for_func(*args, **kwargs)
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/networkx/algorithms/cycles.py", line 525, in minimum_cycle_basis
        nx.connected_component_subgraphs(G)), [])
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/networkx/algorithms/cycles.py", line 524, in <genexpr>
        return sum((_min_cycle_basis(c, weight) for c in
      File "/home/uqtscrim/sage/local/lib/python3.7/site-packages/networkx/algorithms/components/connected.py", line 86, in connected_component_subgraphs
        _warnings.warn(msg, DeprecationWarning)
      File "/home/uqtscrim/sage/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    DeprecationWarning: connected_component_subgraphs is deprecated and will be removedin 2.2. Use (G.subgraph(c).copy() for c in connected_components(G))
    [[1, 2, 3], [1, 2, 3, 4], [5, 6, 7]]

Answer 13 · 2020-06-04T00:30:28.000Z

comment:8

You shouldn't have this warning with networkx 2.4. It implies you are still using 2.2.

Answer 14 · 2020-06-04T00:33:08.000Z

comment:9

I didn't fully merge #29425 into this branch. I can do it (must I?), or leave it to the release manager.

as to the url - I guess it's something to do with GitHub naming of archives.

Answer 15 · 2020-06-04T00:41:23.000Z

comment:10

@kiwifb I am using 2.4 as far as Sage knows:

./sage --installed
...
networkx................................2.4 (2.4)

@dimpase I didn't realize the entire branch of #29425 had not been merged in. It looked like it had from a quick glance. Once I did that then indeed I do not get the deprecation warnings.

Answer 16 · 2020-06-04T00:42:34.000Z

comment:11

Since the doctests do pass for me with #29425 (and you don't need to merge it in any further I think), positive review.

Answer 17 · 2020-06-04T00:42:34.000Z

Reviewer: Travis Scrimshaw

Answer 18 · 2020-06-04T14:41:04.000Z

comment:12

Replying to @tscrim:

Minor (perhaps trivial) point, but pulling it from that link gave a zip file named networkx-networkx-2.4.zip, which caused it to fail without manually downloading and naming the tarball correctly.

I don't quite understand what "pulling it from that link" means. The idea of
upstream_url in checksum.ini is to allow automatic download of the package archive from that URL. (which can either be enabled with -o, as follows, ./sage -i -o networkx of if one runs ./configure with --enable-download-from-upstream-url option).

And this automation works for me on Linux. The ULR is copied verbatim from https://github.com/networkx/networkx/releases/tag/networkx-2.4

Answer 19 · 2020-06-05T00:38:42.000Z

comment:13

Replying to @dimpase:

Replying to @tscrim:

Minor (perhaps trivial) point, but pulling it from that link gave a zip file named networkx-networkx-2.4.zip, which caused it to fail without manually downloading and naming the tarball correctly.

I don't quite understand what "pulling it from that link" means. The idea of
upstream_url in checksum.ini is to allow automatic download of the package archive from that URL. (which can either be enabled with -o, as follows, ./sage -i -o networkx of if one runs ./configure with --enable-download-from-upstream-url option).

And this automation works for me on Linux. The ULR is copied verbatim from https://github.com/networkx/networkx/releases/tag/networkx-2.4

Okay, I tried with just make instead of directly with sage -i (which worked by itself when I did this on another spkg previously), so maybe that is why it didn't work. I was just thinking it came from the zip file having an unexpected name.

Answer 20 · 2020-06-05T02:08:14.000Z

Description changed:

--- 
+++ 
@@ -1,4 +1,4 @@
 one deprecation warning is removed here, other suppressed on ##29425, 
 the rest it trivial.
 
-tarball: https://github.com/networkx/networkx/archive/networkx-2.4.zip
+tarball: see checksums.ini [upstream_url]

Answer 21 · 2020-06-06T04:39:40.000Z

comment:15

From #29425: running the testsuite of networkx requires pytest.

Answer 22 · 2020-06-06T04:55:49.000Z

comment:16

Replying to @mkoeppe:

From #29425: running the testsuite of networkx requires pytest.

I'll add pytest in #29813.

Answer 23 · 2020-06-06T05:11:32.000Z

Changed dependencies from #29425 to #29425, #29813

Answer 24 · 2020-06-06T05:11:32.000Z

Description changed:

--- 
+++ 
@@ -1,4 +1,4 @@
-one deprecation warning is removed here, other suppressed on ##29425, 
+one deprecation warning is removed here, other suppressed on #29425, 
 the rest it trivial.
 
 tarball: see checksums.ini [upstream_url]

Answer 25 · 2020-06-06T05:24:39.000Z

Changed branch from u/dimpase/packages/networkx24 to u/mkoeppe/packages/networkx24

Answer 26 · 2020-06-06T05:31:10.000Z

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

`9a3617d`	`build/pkgs/cython: Update to 0.29.16`
`4a606b6`	`build/pkgs/cython/checksums.ini: Use predictable URL for upstream_url pattern`
`e9a7cb0`	`Update Cython to 0.29.17`
`15911cb`	`build/pkgs/numpy: Update numpy to 1.18.2, back out numpy-1.10.2-no-hardcode-blas.patch (does not apply)`
`96e7a41`	`Update numpy to 1.18.4`
`78b4907`	`Update scipy to 1.4.1`
`54595b0`	`build/pkgs/pybind11: New, dependency of scipy`
`c005562`	`suppress scipy deprecation warnings`
`56a6fe1`	`build/pkgs/numpy: Update to 1.18.5`
`e592bdc`	`Merge branch 't/29425/trac/local_conda_forge_macos_standard__build_error_with_numpy' into t/29766/packages/networkx24`

Answer 27 · 2020-06-06T05:31:10.000Z

Changed commit from 2310c79 to e592bdc

Answer 28 · 2020-06-06T05:31:32.000Z

comment:20

Merged in updated #29425 and #29813.

Answer 29 · 2020-06-06T05:34:02.000Z

comment:21

Tests at https://github.com/mkoeppe/sage/pull/38/checks

Answer 30 · 2020-06-06T05:43:10.000Z

Changed author from Dima Pasechnik to Dima Pasechnik, Matthias Koeppe

Answer 31 · 2020-06-06T21:16:53.000Z

comment:24

#29813 needs to be fixed.

Answer 32 · 2020-06-07T03:02:57.000Z

Changed dependencies from #29425, #29813 to #29480, #29813

Answer 33 · 2020-06-07T03:02:57.000Z

Description changed:

--- 
+++ 
@@ -1,4 +1,12 @@
-one deprecation warning is removed here, other suppressed on #29425, 
+Follow-up from the update of numpy to 1.16.6 (#29429):
+
+We update
+* NumPy to the latest release, 1.18.5
+* scipy to the latest release, 1.4.1.
+* networkx to the latest release, 2.4.
+(latest as of 2020-06-05; all versions (and also the upcoming numpy 1.19) support python >= 3.6)
+
+one deprecation warning is removed here, another suppressed (from #29425), 
 the rest it trivial.
 
-tarball: see checksums.ini [upstream_url]
+tarballs: see checksums.ini [upstream_url]

Answer 34 · 2020-06-07T03:02:57.000Z

Changed reviewer from Travis Scrimshaw to Travis Scrimshaw, Isuru Fernando, Dima Pasechnik

Answer 35 · 2020-06-07T03:04:55.000Z

comment:26

Merged ticket #29425 into this ticket because #29245 could not be separately tested because of networkx 2.2 testsuite errors triggered by updates.

Answer 36 · 2020-06-07T06:35:59.000Z

Changed commit from e592bdc to 11bf67f

Answer 37 · 2020-06-07T06:35:59.000Z

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

`bedc7ae`	`build/make/Makefile.in: Allow pip packages as dependencies`
`11bf67f`	`Merge branch 't/29813/add_pytest_as_a_type_optional__source_pip_package' into t/29766/packages/networkx24`

Answer 38 · 2020-06-08T06:05:25.000Z

comment:29

Tests at https://github.com/mkoeppe/sage/actions/runs/128112212:

On ubuntu-eoan-standard (https://github.com/mkoeppe/sage/runs/748390625), LOTS of failures in the networkx testsuite:

Ran 3618 tests in 386.630s

FAILED (errors=808)

Is this expected?

Answer 39 · 2020-06-08T14:47:36.000Z

comment:30

Failure on cygwin-standard (https://github.com/mkoeppe/sage/runs/748608397):

  g++: scipy/optimize/rectangular_lsap/rectangular_lsap.cpp
  scipy/optimize/rectangular_lsap/rectangular_lsap.cpp:117:43: error: ‘int64_t’ has not been declared
    117 | solve(int nr, int nc, double* input_cost, int64_t* output_col4row)
        |                                           ^~~~~~~
  scipy/optimize/rectangular_lsap/rectangular_lsap.cpp:186:41: error: ‘int64_t’ has not been declared
    186 |                                         int64_t* col4row)
        |                                         ^~~~~~~

Answer 40 · 2020-06-08T23:52:47.000Z

comment:31

The int64_t issue is fixed in scipy/scipy#11320
for the upcoming 1.5.0, marked "backport-candidate", but so far no 1.4.x with the fix has been released...

Answer 41 · 2020-06-09T00:07:58.000Z

Changed commit from 11bf67f to db0fa41

Answer 42 · 2020-06-09T00:07:58.000Z

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

`db0fa41`	`build/pkgs/scipy: Add patch for scipy pr 11320`

Answer 43 · 2020-06-10T23:34:41.000Z

comment:35

Everything builds and works for me on Ubuntu.

Answer 44 · 2020-06-12T06:25:34.000Z

comment:36

lgtm

Answer 45 · 2020-06-12T18:59:23.000Z

comment:37

Comment 29 is unresolved

Answer 46 · 2020-06-12T20:32:23.000Z

comment:38

a package fails its own testsuite. it's not the 1st one (happens on Gentoo as well)

Answer 47 · 2020-06-12T20:39:09.000Z

comment:39

OK, good to know. No objection then

Answer 48 · 2020-06-22T15:02:52.000Z

Description changed:

--- 
+++ 
@@ -4,9 +4,9 @@
 * NumPy to the latest release, 1.18.5
 * scipy to the latest release, 1.4.1.
 * networkx to the latest release, 2.4.
-(latest as of 2020-06-05; all versions (and also the upcoming numpy 1.19) support python >= 3.6)
+(latest as of 2020-06-05; all versions (and also the upcoming numpy 1.19) support Python >= 3.6)
 
-one deprecation warning is removed here, another suppressed (from #29425), 
+One deprecation warning is removed here, another suppressed (from #29425), 
 the rest it trivial.
 
-tarballs: see checksums.ini [upstream_url]
+Tarballs: see checksums.ini [upstream_url]

Answer 49 · 2020-06-28T00:15:20.000Z

comment:41

Lots of failures of the form

**********************************************************************
File "src/sage/modular/modform_hecketriangle/abstract_space.py", line 2235, in sage.modular.modform_hecketriangle.abstract_space.FormsSpace_abstract.ZZ
Failed example:
    qexp_int = WF.rationalize_series(qexp)
Expected nothing
Got:
    doctest:warning
      File "/home/release/Sage/src/bin/sage-runtests", line 178, in <module>
        err = DC.run()
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/control.py", line 1231, in run
        self.run_doctests()
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/control.py", line 931, in run_doctests
        self.dispatcher.dispatch()
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2045, in dispatch
        self.parallel_dispatch()
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1940, in parallel_dispatch
        w.start()  # This might take some time
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2212, in start
        super(DocTestWorker, self).start()
      File "/home/release/Sage/local/lib/python3.7/multiprocessing/process.py", line 112, in start
        self._popen = self._Popen(self)
      File "/home/release/Sage/local/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
        return _default_context.get_context().Process._Popen(process_obj)
      File "/home/release/Sage/local/lib/python3.7/multiprocessing/context.py", line 277, in _Popen
        return Popen(process_obj)
      File "/home/release/Sage/local/lib/python3.7/multiprocessing/popen_fork.py", line 20, in __init__
        self._launch(process_obj)
      File "/home/release/Sage/local/lib/python3.7/multiprocessing/popen_fork.py", line 74, in _launch
        code = process_obj._bootstrap()
      File "/home/release/Sage/local/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap
        self.run()
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2184, in run
        task(self.options, self.outtmpfile, msgpipe, self.result_queue)
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2513, in __call__
        doctests, extras = self._run(runner, options, results)
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 2562, in _run
        result = runner.run(test)
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 912, in run
        return self._run(test, compileflags, out)
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 707, in _run
        self.compile_and_execute(example, compiler, test.globs)
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/doctest/forker.py", line 1138, in compile_and_execute
        exec(compiled, globs)
      File "<doctest sage.modular.modform_hecketriangle.abstract_space.FormsSpace_abstract.ZZ[4]>", line 1, in <module>
        qexp_int = WF.rationalize_series(qexp)
      File "/home/release/Sage/local/lib/python3.7/site-packages/sage/modular/modform_hecketriangle/abstract_space.py", line 2309, in rationalize_series
        warn("Using an experimental rationalization of coefficients, please check the result for correctness!")
      File "/home/release/Sage/local/lib/python3.7/warnings.py", line 110, in _showwarnmsg
        msg.file, msg.line)
    :
    UserWarning: Using an experimental rationalization of coefficients, please check the result for correctness!
**********************************************************************
2 items had failures:
   1 of  29 in sage.modular.modform_hecketriangle.abstract_space.FormsSpace_abstract.ZZ
   1 of  41 in sage.modular.modform_hecketriangle.abstract_space.FormsSpace_abstract.construct_quasi_form
    [566 tests, 2 failures, 7.37 s]
----------------------------------------------------------------------
sage -t --long --warn-long 37.2 src/sage/modular/modform_hecketriangle/abstract_space.py  # 2 doctests failed
----------------------------------------------------------------------

Answer 50 · 2020-06-28T02:38:41.000Z

comment:42

This looks strange. I have done a test run of sage 9.2.beta2 with numpy-1.19/scipy-1.5/networkx-2.4 in sage-on-gentoo and I have seen no such things [in fact the number of doctest failures in sage-on-gentoo is at its lowest in years - 3 doctests across 2 files].

I would think another ticket would be the cause of this, although I am at a loss as to which one when inspecting your last commits. Running my own tests on your branch ASAP.

Answer 51 · 2020-06-28T12:00:51.000Z

comment:43

This is caused by the

warnings.filterwarnings('ignore', '.*Deprec.*scipy.*instead',)

line. This is run on every examples block, and resets the registry of already displayed warnings - so warnings which are not affected by a filter, which by default are only displayed the first time they are hit, are shown again.

Answer 52 · 2020-06-28T14:04:34.000Z

comment:44

oh dear. ok, we can suppress these, too

Answer 53 · 2020-07-05T00:38:32.000Z

comment:45

From the SciPy 1.5.1 Release Notes:

SciPy 1.5.1 is a bug-fix release with no new features
compared to 1.5.0. In particular, an issue where DLL loading
can fail for SciPy wheels on Windows with Python 3.6 has been
fixed.

Answer 54 · 2020-07-05T00:54:13.000Z

Description changed:

--- 
+++ 
@@ -1,10 +1,10 @@
 Follow-up from the update of numpy to 1.16.6 (#29429):
 
 We update
-* NumPy to the latest release, 1.18.5
-* scipy to the latest release, 1.4.1.
-* networkx to the latest release, 2.4.
-(latest as of 2020-06-05; all versions (and also the upcoming numpy 1.19) support Python >= 3.6)
+* https://pypi.org/project/numpy/ to 1.18.5 (latest: 1.19 as of 2020-07-04)
+* https://pypi.org/project/scipy/ to 1.4.1 (latest: 1.5.1 as of 2020-07-04)
+* networkx to the latest release, 2.4 (latest as of 2020-07-04)
+All of the above versions support Python >= 3.6.
 
 One deprecation warning is removed here, another suppressed (from #29425), 
 the rest it trivial.

Answer 55 · 2020-07-05T01:18:37.000Z

Changed commit from db0fa41 to a2e1978

Answer 56 · 2020-07-05T01:18:37.000Z

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

`2efe841`	`Merge tag '9.2.beta3' into t/29766/packages/networkx24`
`634ce73`	`build/pkgs: Update numpy to 1.19, scipy to 1.5.1`
`a2e1978`	`build/pkgs/scipy/patches/pr-11320.patch: Remove`

Answer 57 · 2020-07-05T01:33:45.000Z

comment:48

Tests run at https://github.com/mkoeppe/sage/actions/runs/157933479

Answer 58 · 2020-07-05T01:34:15.000Z

Description changed:

--- 
+++ 
@@ -1,7 +1,7 @@
 Follow-up from the update of numpy to 1.16.6 (#29429):
 
 We update
-* https://pypi.org/project/numpy/ to 1.18.5 (latest: 1.19 as of 2020-07-04)
+* https://pypi.org/project/numpy/ to 1.18.5 (latest: 1.19.0 as of 2020-07-04)
 * https://pypi.org/project/scipy/ to 1.4.1 (latest: 1.5.1 as of 2020-07-04)
 * networkx to the latest release, 2.4 (latest as of 2020-07-04)
 All of the above versions support Python >= 3.6.

Answer 59 · 2020-07-05T01:44:51.000Z

comment:50

The warnings.filterwarnings issue needs to be addressed

Answer 60 · 2020-07-05T08:56:24.000Z

Description changed:

--- 
+++ 
@@ -2,7 +2,7 @@
 
 We update
 * https://pypi.org/project/numpy/ to 1.18.5 (latest: 1.19.0 as of 2020-07-04)
-* https://pypi.org/project/scipy/ to 1.4.1 (latest: 1.5.1 as of 2020-07-04)
+* https://pypi.org/project/scipy/ to 1.5.1 (latest: 1.5.1 as of 2020-07-04)
 * networkx to the latest release, 2.4 (latest as of 2020-07-04)
 All of the above versions support Python >= 3.6.

Answer 61 · 2020-07-05T08:57:21.000Z

Description changed:

--- 
+++ 
@@ -1,7 +1,7 @@
 Follow-up from the update of numpy to 1.16.6 (#29429):
 
 We update
-* https://pypi.org/project/numpy/ to 1.18.5 (latest: 1.19.0 as of 2020-07-04)
+* https://pypi.org/project/numpy/ to 1.19.0 (latest: 1.19.0 as of 2020-07-04)
 * https://pypi.org/project/scipy/ to 1.5.1 (latest: 1.5.1 as of 2020-07-04)
 * networkx to the latest release, 2.4 (latest as of 2020-07-04)
 All of the above versions support Python >= 3.6.

Answer 62 · 2020-07-05T09:31:57.000Z

comment:53

Replying to @slel:

From the SciPy 1.5.1 Release Notes:

SciPy 1.5.1 is a bug-fix release with no new features
compared to 1.5.0. In particular, an issue where DLL loading
can fail for SciPy wheels on Windows with Python 3.6 has been
fixed.

no worries, the ticket description was not updated, that's all.

Answer 63 · 2020-07-05T11:02:44.000Z

comment:54

wrong tarball hash for numpy. See numpy/numpy#16754

the usual GH crappy tarballs stuff I guess.

Answer 64 · 2020-07-05T11:06:04.000Z

comment:55

Currently it is

--- a/build/pkgs/numpy/checksums.ini
+++ b/build/pkgs/numpy/checksums.ini
@@ -1,5 +1,5 @@
 tarball=numpy-VERSION.tar.gz
-sha1=0b1792c8a828d6a9fb148fb33dfa11549391d858
-md5=3f5ce88a859302f0a1aceb5f75b563fc
-cksum=1603544858
+sha1=a6c099c0929e0b437a72fd60aad76d23061087c6
+md5=e3baa59da67348c2d6596803514129db
+cksum=1607167529
 upstream_url=https://github.com/numpy/numpy/releases/download/vVERSION/numpy-VERSION.tar.gz

but of course it's not really stable...

Answer 65 · 2020-07-05T11:12:42.000Z

comment:56

And now md5 as advertised on Numpy pages:

--- a/build/pkgs/numpy/checksums.ini
+++ b/build/pkgs/numpy/checksums.ini
@@ -1,5 +1,5 @@
 tarball=numpy-VERSION.tar.gz
-sha1=0b1792c8a828d6a9fb148fb33dfa11549391d858
-md5=3f5ce88a859302f0a1aceb5f75b563fc
-cksum=1603544858
+sha1=ba35592527bd01f53244ae1bb55c896101bbf0f3
+md5=d59aadf47354bd10c7b9996032ba4da0
+cksum=3500020755

Sucks.

Answer 66 · 2020-07-05T20:01:06.000Z

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

`0879c3b`	`build/pkgs/{numpy,scipy}/checksums.ini: Set upstream_url to pypi.io`

Answer 67 · 2020-07-05T20:01:06.000Z

Changed commit from a2e1978 to 0879c3b

Answer 68 · 2020-07-05T20:01:31.000Z

comment:58

Hopefully the sources packages on PyPI are more stable.

Answer 69 · 2020-07-06T00:28:59.000Z

comment:60

Replying to @mkoeppe:

The warnings.filterwarnings issue needs to be addressed

It appears I figured out this issue; doing

--- a/src/sage/all.py
+++ b/src/sage/all.py
@@ -87,6 +87,8 @@ warnings.filterwarnings('ignore', category=RuntimeWarning,
 # Ignore all deprecations from IPython etc.
 warnings.filterwarnings('ignore', category=DeprecationWarning,
     module='.*(IPython|ipykernel|jupyter_client|jupyter_core|nbformat|notebook|ipywidgets|storemagic|jedi)')
+warnings.filterwarnings('ignore', category=DeprecationWarning,
+    module='.*(scipy|networkx)')
 # Ignore collections.abc warnings, there are a lot of them but they are
 # harmless.
 warnings.filterwarnings('ignore', category=DeprecationWarning,
diff --git a/src/sage/doctest/forker.py b/src/sage/doctest/forker.py
index 9646ddee18..bd7e1a5a97 100644
--- a/src/sage/doctest/forker.py
+++ b/src/sage/doctest/forker.py
@@ -881,7 +881,7 @@ class SageDocTestRunner(doctest.DocTestRunner, object):
         #             and will be removed in SciPy 2.0.0, use numpy.array instead
         # This affects networkx 2.2 up and including 2.4
         # We filter them out here:
-        warnings.filterwarnings('ignore', '.*Deprec.*scipy.*instead',) # cf trac #29425
+ #       warnings.filterwarnings('ignore', '.*Deprec.*scipy.*instead',) # cf trac #29425
 
         warnings.showwarning = showwarning_with_traceback
         self.running_doctest_digest = hashlib.md5()

appears to fix it - still testing, but expect a success.
Basically, we already did suppression of deprecation warnings
from other packages right, now the same has to be done for scipy and networkx

Answer 70 · 2020-07-06T06:39:33.000Z

comment:61

Basically, we already did suppression of deprecation warnings
from other packages right, now the same has to be done for scipy and networkx

Doesn't that also hide deprecation warnings coming from sage code, and thus would prevent detecting such deprecated code use with future scipy/networkx updates? That sort of defeats the point of the previous

    deprecationWarning = ('ignore', None, DeprecationWarning, None, 0)
    if deprecationWarning in warnings.filters: warnings.filters.remove(deprecationWarning)

Answer 71 · 2020-07-06T07:01:36.000Z

comment:62

Replying to @antonio-rojas:

Basically, we already did suppression of deprecation warnings
from other packages right, now the same has to be done for scipy and networkx

Doesn't that also hide deprecation warnings coming from sage code, and thus would prevent detecting such deprecated code use with future scipy/networkx updates? That sort of defeats the point of the previous
    deprecationWarning = ('ignore', None, DeprecationWarning, None, 0)
    if deprecationWarning in warnings.filters: warnings.filters.remove(deprecationWarning)

module='.*(scipy|networkx)' suppresses deprecation warnings emitted by code in scipy or in networkx. It does not do anything to these warnings from Sage code itself.
Indeed, the tests in src/sage/tests/deprecation_test.py still pass.

Answer 72 · 2020-07-06T07:04:32.000Z

comment:63

needless to say, each upgrade to scipy|networkx will need to re-evaluate these.

We should add something to the Developer manual on this.

Answer 73 · 2020-07-06T07:16:43.000Z

Changed branch from u/mkoeppe/packages/networkx24 to u/dimpase/packages/networkx24

Answer 74 · 2020-07-06T07:16:43.000Z

comment:64

correct suppression of warnings

New commits:

`8a70ebe`	`right place for warnings.filterwarnings() for networkx & scipy`

Answer 75 · 2020-07-06T07:16:43.000Z

Changed commit from 0879c3b to 8a70ebe

Answer 76 · 2020-07-12T05:04:04.000Z

Changed dependencies from #29480, #29813 to #29480, #29813, #30118

Answer 77 · 2020-07-12T05:18:04.000Z

Changed branch from u/dimpase/packages/networkx24 to u/mkoeppe/packages/networkx24

Answer 78 · 2020-07-12T05:18:14.000Z

Changed dependencies from #29480, #29813, #30118 to #29480, #29813, #29929, #30118

Answer 79 · 2020-07-12T05:18:14.000Z

Changed commit from 8a70ebe to 0356be9

Answer 80 · 2020-07-12T05:18:14.000Z

Last 10 new commits:

`b7bf43b`	`build/bin/write-dockerfile.sh: ADD src/bin for bootstrapping, needed by src/doc/bootstrap after #29884`
`365ce61`	`Merge branch 'u/mkoeppe/fix_tox_docker_builds_broken_by__29884' of git://trac.sagemath.org/sage into HEAD`
`fb61a31`	`Merge branch 'u/mkoeppe/tox_ini__debian_bullseye___sid_have_python3_8_instead_of_3_7' of git://trac.sagemath.org/sage into 9.2.beta3+ci-fixes`
`bfb2fb4`	`tox.ini: Add environment local-homebrew-usrlocal`
`aded967`	`Merge commit '8195cb821e01e0cb627240d7135d5634c265cab0' of git://trac.sagemath.org/sage into t/29929/tox_ini__add_a_macos_environment_without_homebrew__conda`
`07d657d`	`Merge tag '9.2.beta4' into t/29929/tox_ini__add_a_macos_environment_without_homebrew__conda`
`bd15fb9`	`build/bin/write-dockerfile.sh, tox.ini: Do not run testsuites that need pip (ssl)`
`25393b0`	`Handle SAGE_CHECK_PACKAGES in build/make/Makefile.in, not sage-spkg`
`5aca238`	`Merge branch 't/30118/handle_sage_check_packages_in_build_make_makefile_in__not_sage_spkg' into t/29929/tox_ini__add_a_macos_environment_without_homebrew__conda`
`0356be9`	`Merge branch 't/29929/tox_ini__add_a_macos_environment_without_homebrew__conda' into t/29766/packages/networkx24`

Answer 81 · 2020-07-12T18:41:48.000Z

comment:68

Elevating this upgrade ticket to "blocker" because homebrew builds are currently broken because they now ship gfortran 10.

Answer 82 · 2020-07-12T18:43:26.000Z

comment:69

(see for example https://github.com/sagemath/sage/runs/862451287)

Answer 83 · 2020-07-13T21:39:14.000Z

Changed commit from 0356be9 to 6954deb

Answer 84 · 2020-07-13T21:39:14.000Z

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

`6954deb`	`build/pkgs/networkx/dependencies: Conditionalize on SAGE_CHECK_networkx`

Answer 85 · 2020-07-14T02:10:15.000Z

comment:71

Tests as part of Meta-ticket #29456 (Support gcc/gfortran 10) at https://github.com/mkoeppe/sage/actions/runs/167913617

On debian-jessie-standard (https://github.com/mkoeppe/sage/runs/866989168) and debian-jessie-minimal:

gcc: build/src.linux-x86_64-3.7/numpy/core/src/umath/loops.c
In file included from numpy/core/src/umath/loops.c.src:50:0:
numpy/core/src/umath/simd.inc.src: In function ‘avx512_conjugate_ps’:
numpy/core/src/umath/simd.inc.src:1843:5: warning: implicit declaration of function ‘_mm512_castps_si512’ [-Wimplicit-function-declaration]
     __m512i cast_x = _mm512_cast@vsub@_si512(x);
     ^
numpy/core/src/umath/simd.inc.src:1843:22: error: incompatible types when initializing type ‘__vector(8) long long int’ using type ‘int’
     __m512i cast_x = _mm512_cast@vsub@_si512(x);
                      ^
numpy/core/src/umath/simd.inc.src:1846:5: warning: implicit declaration of function ‘_mm512_castsi512_ps’ [-Wimplicit-function-declaration]
     return _mm512_castsi512_@vsub@(res);
     ^
numpy/core/src/umath/simd.inc.src:1846:5: error: incompatible types when returning type ‘int’ but ‘__vector(16) float’ was expected

Answer 86 · 2020-07-14T02:19:58.000Z

comment:72

debian-jessie uses gcc version 4.9.2 (Debian 4.9.2-10+deb8u2), and a fix for something that looks related is in gcc 4.9.3 - https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61878

Answer 87 · 2020-07-14T02:23:22.000Z

comment:73

This was added to numpy in numpy/numpy@5562a8c
(new in 1.19.0)
(numpy/numpy#15408)

Answer 88 · 2020-07-14T02:40:27.000Z

Upstream: Reported upstream. No feedback yet.

Answer 89 · 2020-07-14T02:43:52.000Z

comment:75

Let's go back to 1.18.5 (released Jun 3, 2020).

Answer 90 · 2020-07-14T02:48:41.000Z

Changed commit from 6954deb to d39817f

Answer 91 · 2020-07-14T02:48:41.000Z

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

`d39817f`	`build/pkgs/numpy: Downgrade to 1.18.5 to fix build error on debian-jessie`

Answer 92 · 2020-07-14T02:49:11.000Z

Description changed:

--- 
+++ 
@@ -1,7 +1,7 @@
 Follow-up from the update of numpy to 1.16.6 (#29429):
 
 We update
-* https://pypi.org/project/numpy/ to 1.19.0 (latest: 1.19.0 as of 2020-07-04)
+* https://pypi.org/project/numpy/ to 1.18.5 (latest: 1.19.0 as of 2020-07-04)
 * https://pypi.org/project/scipy/ to 1.5.1 (latest: 1.5.1 as of 2020-07-04)
 * networkx to the latest release, 2.4 (latest as of 2020-07-04)
 All of the above versions support Python >= 3.6.

Answer 93 · 2020-07-14T02:51:35.000Z

comment:78

New tests at https://github.com/mkoeppe/sage/actions/runs/168163111

Answer 94 · 2020-07-14T17:04:41.000Z

comment:79

Actually https://github.com/mkoeppe/sage/actions/runs/168165380

Answer 95 · 2020-07-14T17:30:50.000Z

comment:80

In these tests, fedora-32-standard uses gcc/gfortran 10 from the system, fedora-32-minimal uses gcc 10 from the system and gfortran from spkg, and the various homebrew-macos tests use gcc = apple clang, gfortran 10 from homebrew.

They all succeed. Still waiting for other tests to finish.

Answer 96 · 2020-07-14T17:51:53.000Z

comment:81

On local-conda-forge-macos-maximal (https://github.com/mkoeppe/sage/runs/867731009), numpy 1.18.5 builds but building sagelib fails as follows:

  [sagelib-9.2.beta5] error installing, exit status 1. End of log file:
  [sagelib-9.2.beta5]       from . import core
  [sagelib-9.2.beta5]     File "/Users/runner/work/sage/sage/.tox/local-conda-forge-macos-maximal/local/lib/python3.7/site-packages/numpy/core/__init__.py", line 50, in <module>
  [sagelib-9.2.beta5]       raise ImportError(msg)
  [sagelib-9.2.beta5]   ImportError: 
  [sagelib-9.2.beta5]   
  [sagelib-9.2.beta5]   IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE!
  [sagelib-9.2.beta5]   
  [sagelib-9.2.beta5]   Importing the numpy C-extensions failed. This error can happen for
  [sagelib-9.2.beta5]   many reasons, often due to issues with your setup or how NumPy was
  [sagelib-9.2.beta5]   installed.
...
  [sagelib-9.2.beta5]   Original error was: dlopen(/Users/runner/work/sage/sage/.tox/local-conda-forge-macos-maximal/local/lib/python3.7/site-packages/numpy/core/_multiarray_umath.cpython-37m-darwin.so, 2): Symbol not found: _aheapsort_bool
  [sagelib-9.2.beta5]     Referenced from: /Users/runner/work/sage/sage/.tox/local-conda-forge-macos-maximal/local/lib/python3.7/site-packages/numpy/core/_multiarray_umath.cpython-37m-darwin.so
  [sagelib-9.2.beta5]     Expected in: flat namespace
  [sagelib-9.2.beta5]    in /Users/runner/work/sage/sage/.tox/local-conda-forge-macos-maximal/local/lib/python3.7/site-packages/numpy/core/_multiarray_umath.cpython-37m-darwin.so

Also building matplotlib fails with a similar error.

Answer 97 · 2020-07-14T22:19:20.000Z

Changed upstream from Reported upstream. No feedback yet. to Reported upstream. Developers acknowledge bug.

Answer 98 · 2020-07-14T22:21:47.000Z

comment:83

numpy 1.19 had worked OK on local-conda-forge-macos when that was tested (https://github.com/mkoeppe/sage/runs/858960241).

Answer 99 · 2020-07-14T22:23:49.000Z

comment:84

Trying next with numpy/numpy#16871

Answer 100 · 2020-07-14T22:52:13.000Z

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

`fcddf69`	`Revert "build/pkgs/numpy: Downgrade to 1.18.5 to fix build error on debian-jessie"`
`3558737`	`build/pkgs/numpy/patches/16871.patch: New`