Don Simons wrote:
>
> I've been experimenting with my latest PMX distribution from
> http://www.dslnorthwest.net/~don/pmx/pmx.html . It was created and
> ZIPped under Win/DOS. Using "unzip" I decompressed the file on a linux
> box. After making the two usual syntax changes and renaming source.for
> to pmxab.f it compiled OK under g77. But when processing the unzipped
> sample .pmx files, Linux-PMX had all sorts of problems related to text
> line terminations. If on the other hand I retrieve the same .pmx files
> by ascii-mode FTP from a Win/DOS box, there are no such problems.
>
> I suppose the CR-LF's are somehow surviving the DOS-ZIP/Linux-unzip
> process, then confusing the Linux-compiled PMX.
>
> I wonder why the Linux compilers (g77 and also lf95) did not suffer the
> same confusion that PMX did regarding the ends of lines.
>
> I can imagine that this behavior would irritate people who only work in
> Linux, but I can't really recall anyone complaining in the past, and I
> really don't relish changing PMX to try to compensate for it. Any
> thoughts, discussion, suggestions?
>
Suggestions - unfortunately no, observations - unfortunately (?) yes!
For a test I attempted to let pmx23 (built with f2c & gcc) process the
pmx output from the mtx source of my latest published madrigal by Mogens
Pedersøn (Amor, per tua mercè). The outcome of that was an immediate
segmentation fault :-( This problem is not with pmx 2298!
Would that be a clue?
--
Christian Mondrup, Computer Programmer
Scandiatransplant, Skejby Hospital, University Hospital of Aarhus
Brendstrupgaardsvej, DK 8200 Aarhus N, Denmark
Phone: +45 89 49 53 01
