[rkward-devel] rkwardtests & docs

[Thomas Friedrichsmeier]

Hi,

On Sunday 31 October 2010, meik michalke wrote:
> i spent some time on adding setwd() calls to a lot of tests, using
> rktest.getTempDir(). if i haven't overlooked something all tests should now
> create their special files and/or graphs only in the temp dir as well, so
> by calling a test the standards dir is entirely left untouched. now i
> wonder if this could be better solved if it was part of the tests
> functions somehow, though.

Sounds like a good idea. But in fact, this should probably be part of 
rktests.runRKTest[Suite]() to keep the tests themselves slim. Of course the 
follow up problem is that using relative paths to read in data files will not 
work for the import plugin tests, so weed need an additional function to 
locate the test-source directory.

> i worked around most errors i could find. but i have no clue how to deal
> with the errors i get by "import_spss" and "import_stata":

It seems I was a bit too optimistic in including non-latin1 characters in 
these data files. Interestingly, both tests work fine for me. Probably it 
depends on which locales are installed on the system, but I did not find the 
time to investigate this, yet.

Either way, the core of the problem is in read.spss() / read.dta(), not in our 
plugin code, so to some degree it is a false alarm. On the other hand, not 
being able to import these localized files is an issue, and we should try to 
improve on this, somehow.

BTW, I see that read.spss() has gained a parameter "reencode" meanwhile, 
perhaps using this would help?

> somebody please have a look at the "dotchart" test in "plots.R". i get an
> error caused by the second plot call:
> 
> <error>
> Error in plot.xy(xy.coords(x, y), type = type, ...) : invalid plot type
> </error>

Looks like a genuine bug, but I don't understand where it is coming from.
   x <- table (datasets::warpbreaks[["tension"]])
   dotchart (unclass (x))	# works
   dotchart (x)  # fails
Frankly, I don't understand, why. And I know it worked in an earlier version 
of R. Ideas?

> last but not least, an error from "plot_history_basics" in
> "rkward_application_tests.R":
> 
> <error>
> Error : rk.record.plot$sP.length == 6 is not TRUE
> </error>

I don't see this one with R 2.12.0, lattice 0.19-15.

> btw. this works perfectly here -- and it seems the previous hardcoded list
> of needed packages was outdated anyway, wasn't it?

It probably was. Hardcoded stuff tends to get outdated real fast...

Regards
Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20101101/8809b5d5/attachment.sig>