Since most of what Collate does is to cat files together in a
particular order, and you are doing automatic conversion, why not just
sacrifice the individual files and cat them together during the
conversion that you are doing, instead of package building?
You aren't planning on editing them by hand, are you, or
replacing/updating piecemeal?
Alternatively, do this as part of a configure script or makefile?
(i.e. on build, concatenate the files into one and remove the little
pieces)?
best,
-tony
On 4/24/05, Ali - <saveez@hotmail.com> wrote:> Last week I managed to automatically wrap over 600 C++ classes into R in
> less than 24 hours. It was not because I am an expert in R programming, it
> was simply because I wrapped the classes as S3. Not only the wrapping
> process was easy, but also installation process and loading the final
> library was fast too.
>
> Now I am also trying to do the wrapping into the 'modern' S4
classes. This
> is the 5th day I am working on this and I have to say S4 was not design for
> big packages!
>
> (1) Because of its meta-data design, the setClass commands must be sorted
in
> the order of inheritance and this is a real pain when it comes to a few
> hundered classes. I wrote some R code to read the annotated source files
and
> return the ordered file names.
>
> (2) After all, when I put the ordered file list in front of Collade in the
> DESCRIPTION file, I got this error induring installation:
>
> ---------- Making package vtkr ------------
> adding build stamp to DESCRIPTION
> Error in .read_description(dfile) : file 'DESCRIPTION' is
> not in valid DCF format
> Execution halted
> make[2]: *** [frontmatter] Error 1
> make[1]: *** [all] Error 2
> make: *** [pkg-vtkr] Error 2
> *** Installation of vtkr failed ***
>
> After a bit playing with th enumber of th efiles, it seems to me Collade
> cannot handle (probably) more than 256 files.
>
> For many good reasosns, I am not going to break the package up into
> different pieces. If I am not missing anything here, I hope some R
developer
> can let me know about some work around on this bug. (hey even java and
> python don't have this limitation, not sure about Ada83.)
>
> ______________________________________________
> R-devel@stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
--
best,
-tony
"Commit early,commit often, and commit in a repository from which we can
easily
roll-back your mistakes" (AJR, 4Jan05).
A.J. Rossini
blindglobe@gmail.com