Re: using make to run binaries

[Andreas Maunz]

Hi everybody,

> This technique has never failed me: take the command
>
>     /path/to/binary -i input.data [more args] > output.data
>
> and change it to
>
>     echo /path/to/binary -i input.data [more args]; $(SHELL) -i
>
> Then run make.

Now this is a *very* clever idea. I am going to pursue that. Meanwhile I 
have tried to set $SHELL=/bin/bash in the Makefile, since for manual 
launch I also use bash. Indeed, make then really uses bash to run the 
above command line. However, this does not affect the problem in any way.

> The situation smells like undefined behaviour to me; if you're using
> an uninitialised variable or in some other way depending on random
> memory contents, that could trigger the different behaviour in your
> program. Your description of the problem is a bit vague, but by
> "more chemical information", I'm guessing the library is either
> returning more records, or more data fields in each record.

The output consists mainly of chemical compounds. Every compound has a 
score attached (a floating point number). When run with make, lower 
scores are attached and thus more chunks fall below a predefined 
threshold (0.3) which results in less output. The score is a measure of 
similarity between the compounds and one specific compound, called the 
query compound. For similarity computation the already mentioned 
libraries are accessed. It really seems that is has something to do with 
program internal computation and/or the way libraries are accessed.

The manual results seem to make much more sense, so in the meantime I 
have decided to refer to them as "correct".

Regards, Andreas
--
http://www.maunz.de

           If you are passed on the right, you're in the wrong lane.