Fix standard_packages(), optional_packages(), and experimental_packages()

[ncohen]

Those functions are broken (and not doctested) in the latest beta. This branches fixes it.

[git]

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

​c400ee6

trac #18407: Fix standard_packages(), optional_packages(), and experimental_packages()

[vdelecroix]

Is it disjoint from the bash scripts that we have

$ sage -standard
Using Sage Server http://www.sagemath.org/packages

Standard packages:

atlas-3.10.1.p7 ........................ installed version: 3.10.2
boehm_gc-7.2d.p0 ....................... already installed
boost_cropped-1.52.0 ................... already installed
...
tachyon-0.98.9.p5 ...................... already installed
zlib-1.2.8.p0 .......................... already installed
zn_poly-0.9.p11 ........................ already installed

$ sage -optional
Optional packages:

4ti2-1.6.2 ............................. not installed
PyQt_x11-4.9.1.p0 ...................... not installed
TOPCOM-0.17.4 .......................... not installed
autotools-20121217 ..................... not installed
...
trac-0.11.5.p0 ......................... not installed
valgrind-3.8.1.p0 ...................... not installed
zeromq-3.2.0.p0 ........................ installed version: 4.0.5

[leif]

Hmmm,

• that's only part of the problem,
• I was going to rather fix it by adding a --machine-readable (say) option to sage-list-packages (at #16759, but I haven't written any code yet, also because of #15642).

[ncohen]

Is it disjoint from the bash scripts that we have

No, it parses the output.

[ncohen]

• that's only part of the problem,
• I was going to rather fix it by adding a --machine-readable (say) option to sage-list-packages (at #16759, but I haven't written any code yet, also because of #15642).

Well, that branch fixes the old function, and has the nice property of being ready and implemented. We will still need those functions in the long run, so it is not incompatible with what you plan on doing anyway. I really have nothing against seeing this parsing code rewritten in the future.

I am writing a doc patch at the moment, which uses these functions to make sure that all packages appear in our doc.

Nathann

[leif]

Hmmm ok, although it now exposes the brokenness of the shell script in Python as well... ;-)

One tiny thing:

        print("Standard package list (shown above) appears to be currently")

=>

        print("List of %s packages (shown above) appears to be currently" % package_type)

[ncohen]

Can you add a commit? I am working on another patch at the moment.

[git]

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

​10933bc

Fix exception message of _package_lists_from_sage_output() (#18407)

[ncohen]

Thanks !! I am almost done with my other patch which lists all packages in the doc. I also noticed some problem in the output of this one, which I will fix real soon ;-)

[leif]

Replying to ncohen:

Can you add a commit? I am working on another patch at the moment.

Done, but

$ ./sage -c 'print optional_packages()'
(['4ti2', 'git', 'pyzmq', 'zeromq'], ['PyQt_x11', 'TOPCOM', 'autotools', 'beautifulsoup', 'benzene', 'biopython', 'brian', 'buckygen', 'cbc', 'ccache', 'chomp', 'cluster_seed', 'coxeter3', 'cryptominisat', 'cunningham_tables', 'database_cremona_ellcurve', 'database_gap', 'database_jones_numfield', 'database_kohel', 'database_odlyzko_zeta .................. not installe', 'database_pari', 'database_stein_watkins_mini.p0 ......... not installe', 'database_symbolic_data', 'dot2tex', 'extra_docs', 'gap_packages', 'gcc', 'gdbm', 'ginv-1.9', 'gnuplotpy', 'guppy', 'java3d', 'kash3-2008-07', 'knoboo', 'libogg', 'libtheora', 'lie', 'lrs', 'mpi4py', 'nauty', 'normaliz', 'nose', 'nzmath', 'openmpi', 'openssl', 'ore_algebra', 'p_group_cohomology', 'patchbot', 'phc', 'pybtex', 'pycryptoplus-20100809', 'pyx', 'qhull', 'sage_mode', 'sip', 'trac', 'valgrind'])
$ ./sage -c 'print experimental_packages()'
(['pexpect'], ['PyQt4', 'PyVTK', 'QScintilla2', 'asymptote', 'bison', 'boost_1_34_1 ........................... not installe', 'cadabra', 'clapack', 'clisp', 'cmake', 'csdp', 'dvipng', 'ets', 'fes', 'flex', 'gcc', 'gdb', 'gnofract4d', 'gnuplot', 'graphviz', 'latte_int', 'libcprops', 'libjpeg', 'libsigsegv', 'macaulay2-1.1', 'mayavi_2.2.1 ........................... not installe', 'meataxe', 'modglue', 'mpich2', 'numarray', 'numeric', 'openopt', 'pcre', 'phcpack', 'polymake', 'processing', 'pygame', 'pygsl', 'pygtk', 'pynifti', 'pyqt', 'pyrexembed-0.1.1', 'qasm', 'qepcad', 'quantlib', 'quantlib_swig', 'reallib3-linux', 'sandpile', 'scitools++ ............................. not installe', 'semigroupe-2.0', 'simpqs', 'sip', 'soya', 'soya_cvs', 'superlu', 'surf', 'vtk_meta', 'wxPython', 'yafray', 'yassl'])

[leif]

Seems you already noticed the pas installé(?)

[git]

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

​cc02304

trac #18407: better parsing of package name

[ncohen]

Seems you already noticed the pas installé(?)

Indeed. But come on, that's written in french on my console too ;-)

Nathann

[leif]

Replying to ncohen:

Seems you already noticed the pas installé(?)

Indeed. But come on, that's written in french on my console too ;-)

I thought you'd just s/not/pas/...

And I won't complain about removing lambda I love so much, nor efficiency w.r.t. using in and .rfind(). ;-)

[ncohen]

And I won't complain about removing lambda I love so much, nor efficiency w.r.t. using in and .rfind(). ;-)

I would have loved to do it with lambda functions only, but I really needed to define a variable here :-P

Thank you very much for the review. The list of packages is built by #18408 :-)

Nathann

[leif]

Hmmm, may I revert the status or do we have to open a follow-up? XD

We have to s/# internet/# optional -- internet/ I think...

Sorry, tested but missed that somehow.

[ncohen]

Hmmm, may I revert the status or do we have to open a follow-up? XD

Oh True, True :-/

Just a second.

Nathann

[git]

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

​dd02493

trac #18407: 'internet' -> 'optional internet'