[Bioc-devel] bug-report for combine

"Laurent Gautier" <lgautier at gmail.com> writes:

2007/7/18, Seth Falcon <sfalcon at fhcrc.org>:

Hi Laurent,

"Laurent Gautier" <lgautier at gmail.com> writes:

It seems to be a problem with 'identical'.

The following patched "combine" seems to fix the problem
(workaround for zero-rows data.frames:
length(sharedRows) == 0))
)

Thanks for the report and the fix.  Would you be willing to resend as
a real patch?  By that I mean:
   - make the change in an up to date svn working copy of Biobase
   - send output of 'svn diff'

I would be willing
(...'just have to find the time to do it ;-) )

(and I also fixed a probable other bug since my last post)

You can run svn diff on just a single file if that helps.  You can,
with care, edit hunks out of a patch file to isolate a change.

I know how these things go, but perhaps you can appreciate that it is
a bit funny to have time to send an email along with a fix, but not
have time to send a patch and expect others to generate it by hand...
or maybe you aren't asking and are just saying that you need a spot of
time which is quite understandable...

Moving off topic, I've been using git [1] as my local version control
system and it really helps with this sort of problem because it allows
you to commit changes locally and then revise those commits prior to
sending them on to svn using git-svn, of course.  This goes a long way
towards preventing patches with random unrelated fixes sprinkled
throughout.  And it makes generating patches for picky picky people on
mailing lists much easier ;-)

We should also add your test case to the unit tests in inst/UnitTests

I was about to suggest it.

Should be an easy patch :-)

+ seth

[1] http://git.or.cz/

Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center
http://bioconductor.org