[GRASS5] [bug #2402] (grass) r.colors - new rules=name option not yet working

[Helena]

Glynn Clements wrote:
> Markus Neteler wrote:
> 
> 
>>>>Subject: r.colors - new rules=name option not yet working
>>>>
>>>>I tried the new rules=[filename] option. It generates ERROR: Unable to
>>>>open rules file [filename]. Then it stops all processing until a
>>>>ctrl-C is pressed. I tried a simple rules file:
>>>>
>>>>1 blue
>>>>2 green
>>>>end
>>>>
>>>>(I tried it with and without 'end' and got the same result)
>>>
>>>The rules file needs to be in the $GISBASE/etc/colors/ directory. 
>>
>>I have added the path to the message. Now the user knows where
>>the file is searched:
>>
>>r.colors N45E008.meters rules=test.rul
>>ERROR: Unable to open rules file test.rul in
>>       /nfsmnt/ssi0/ssi/neteler/grass57/dist.i686-pc-linux-gnu/etc/colors/test.rul
> 
> 
> This is the wrong solution; a user doesn't need this information. A
> better solution would be to list the set of valid options. If the list
> was constructed and set as opt4->options, G_parser() would do this
> automatically; however, the list could easily become quite large.
> 
> The original problem seems to stem from the user thinking that rules=
> was meant for using user-supplied rule files. It isn't; color=rules
> and shell redirection is the correct solution in that case.

I very much agree with Glynn, as shown in his previous message
the user supplied rule file should be done through redirection of the rules file

r.colors mapname color=rules < rulesfile

adding the rules= would be confusing (sorry for saying it this late,
I was away and could not respond when it was suggested).

Helena



  I'll fix
> that problem by removing the word "file" from the option's
> description, e.g.
> 
>     opt4->description = "name of predefined rule set";
> 
> rules= is meant to be analogous to color=, and may well be eliminated
> in future (i.e. rules= will be renamed to color=). I just didn't want
> to mess up the existing color= behaviour until the new code had some
> testing and the "exceptions" (i.e. color=grey.eq/grey.log/random/rules)
> had been dealt with.
>