[GRASS5] cvs updates completed successfully.

Hamish hamish_nospam at yahoo.com
Sat Sep 18 22:23:29 EDT 2004

> > I'm referring to scripts/Makefile, not scripts/r.univar.sh/Makefile.
> I was too. That's what's strange. I did look yesterday morning. In
> fact Hamish updated this Makefile a few days ago to have r.univar.sh.
> So I am baffled as to why it is now showing r.univar (with Hamish
> still listed as updating it). 

Ok, we have:

grass51/scripts/r.univar/    (now empty)
grass51/scripts/r.univar.sh/ (now populated)

grass51/scripts/Makefile     (which script directories to look into)

the Makefile [generally] works on stuff at the same directory level.
i.e. the names listed in "grass51/scripts/Makefile" refer to the
directories that 'make' should descend into, where it will look for new
Makefiles. It does not have anything to do with the final name of the
module besides being convenient. Thus a generic name (like r.univar/) was
fine for the directory name as it gets the idea across with the minimum
unnecessary yet unambiguous information. ie it didn't really matter as long
as the directory structure and scritps/Makefile agreed.

The final name for the module is defined in the Makefile that is in the
source code directory (grass51/scripts/r.univar.sh/Makefile; "PGM = x.foo").
For scripts "make"ing just means copying to $GISBASE/scripts/, so the script
name does have to be correct there. For C code, all you have to do is rename 
PGM in the module/Makefile file to rename the module.

The problem with renaming & moving directories in CVS is that you lose
the modification history. It still exists in the Attic/ (why you shouldn't
delete old directories) but is rather confusing to follow the history that 
way. Thus it is better to try and avoid renaming files and the directory
structures if at all possible. Limitation of CVS, but somewhat logical I
guess from the perspective of keeping the change history intact.

> >> The reason for renaming it is that there is a binary module also
> >> named r.univar. According to Markus, r.univar (binary) will
> >> eventually have all the functionality (and more?) of r.univar.sh
> >> (script). However, it is still under development so we are keeping
> >> the script for awhile.
> > 
> > I was asking why you renamed the directory; I know why you renamed
> > the script.
> I renamed the directory because I thought it ought to match the
> script.

as above, that isn't required and should be avoided unless there is a 
compelling reason.

> If it is better to do it another way, please do so.

I've just modified grass51/scripts/Makefile to reflect the new directory
name. Everything should be working now, no further changes required.


More information about the grass-dev mailing list