Re: 20020125: ncdump printing of multi-line global attributes

>To: address@hidden
>cc: address@hidden
>From: Willie Kantrowitz <address@hidden>
>Subject: ncdump printing of multi-line global attributes
>Organization: MIT
>Keywords: 200201251855.g0PItXx02262 ncdump global attributes

Hi William,

> ncdump prints global attributes indented two tabs worth.  If the value
> of the attribute is a multi-line string the printing of continuation
> lines uses a slight indentation off  the left margin rather than an
> indentation off the printing of the name of the global attribute.  This
> makes it hard to read, especially if there are many global attributes
> with many multi-line continuations.
> 
> Example (simulating effect of tabs);
> 
> // global attributes:
>                            :name = "\n",
>      "first line of information\n",
>      "second line of information\n",
>      "third line of information" ;
>                            :another = "\n",
>      "a line of information\n",
>      "another line of information" ;
> 
> This format is hard to read and find the attribute names.  The following
> outline-style format would be easier to read.
> 
> // global attributes:
>                            :name = "\n",
>                                        "first line of information\n",
>                                        "second line of information\n",
>                                        "third line of information" ;
>                            :another = "\n",
>                                        "a line of information\n" ,
>                                        "another line of information" ;
> 
> Since white-space is no problem for ncgen it could read such an
> outline-style format without problem.
> 
> This is an annoyance.  I would hope/expect that it would not be too
> difficult to correct,  especially since ncdump already knows to do a
> small indentation for continuation lines.
> 
> (While I have not tried it, I assume the same indentation scheme is used
> for printing attributes of variables.  A similar correction should hold
> there too.)

Thanks, this is a good suggestion.  I'll try to get it into the 3.5.1
release.

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu