octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: run BISTs for all installed packages


From: Andrew Janke
Subject: Re: run BISTs for all installed packages
Date: Tue, 5 Feb 2019 11:09:43 -0500
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.13; rv:60.0) Gecko/20100101 Thunderbird/60.5.0



On 2/5/19 11:01 AM, Carnë Draug wrote:
On Mon, 28 Jan 2019 at 03:12, Andrew Janke <address@hidden> wrote:
On 1/21/19 11:44 AM, "Markus Mützel" wrote:
I frequently cross-compile for Windows with all packages. But I very rarely 
actually test whether they work as expected.
Is there an easy way to run all BISTs of all installed packages?
Something that would be more streamlined than loading the packages one after the other 
and manually running "test" for all files.

Markus


Here's a function to do this.

__test_pkgs__     ## test all installed packages
__test_pkgs__ control   ## test a single package
__test_pkgs__ ({'control','io'})    ## test just a list of specified
packages

I'm using this in my (new, experimental) jsonstuff package:
https://github.com/apjanke/octave-jsonstuff. You can have a look there
to see how to call it on multiple OSes in a Travis CI environment.
(Windows isn't working there yet, but I think it's close.) You can also
just run it interactively; it has no dependencies except for Octave itself.

The Windows Octave installers seem to come with a bunch of packages
preinstalled, so a plain __test_pkgs__ will end up testing a lot of
stuff there.


Beware that not all package tests are in .m files.  There will be
tests in .cc files which get compiled into .oct files.  Such tests
will get discarded together with the source files.


Hmm.

In core Octave, are those the tests that end up in matlabroot()/share/octave/4.4.1/etc/tests ans *.cc-tst or in the Octave source tree as *.in.yy-tst files? Maybe we could add a mechanism to packages to similarly extract them to *-tst files?



reply via email to

[Prev in Thread] Current Thread [Next in Thread]