RE: (ITS#4269) ldapsearch no LDIF wrap option

[mike.patnode@centrify.com]

> -----Original Message-----
> From: Kurt D. Zeilenga [mailto:Kurt@OpenLDAP.org] 
> Sent: Tuesday, December 20, 2005 5:23 PM
> To: Mike Patnode
> Cc: openldap-its@OpenLDAP.org
> Subject: RE: (ITS#4269) ldapsearch no LDIF wrap option
> 
> At 04:07 PM 12/20/2005, mike.patnode@centrify.com wrote:
> >> -----Original Message-----
> >> From: Kurt D. Zeilenga [mailto:Kurt@OpenLDAP.org]
> >> 
> >> At 06:45 AM 12/20/2005, ando@sys-net.it wrote:
> >> >> The Netscape version could optionally turn off LDIF wrap.  
> >> Handy for
> >> >> scripting and pipes etc.  I'm surprised no one has ever 
> done this 
> >> >> before. Is there a religious movement against it?
> >> >
> >> >It's so trivial to workout otherwise that it does not deserve the 
> >> >coding effort.
> >> 
> >> Personally, I rather spend my time making other programs properly 
> >> accepting conformant LDIF, including line wraps, than to modifying 
> >> OpenLDAP Software to produce non-compliant LDIF files 
> (note that some 
> >> line wrapping is required by the open standard).
> >> 
> >
> >OK, then why add the integer option?  
> 
> Note that my -r 0 suggestion was in response to Ando's 
> suggestion that ldapsearch handle -r with an optional integer.  Due to
> getopt(3) restrictions, if there is an integer to be expected 
> with the flag, it cannot be optional.  That is, it has to be 
> getopt(argc,argv,"r" ...) or getopt(argc,argv,"r:" ...).
> 
> >> >> Diffs attached for ldapsearch -r (turn off line wrap)
> >> >
> >> >I suggest making the -r switch accept an optional arg 
> that specifies 
> >> >the wrap column, defaulting to LDIF_LINE_WIDTH; so
> >> >
> >> >(nothing) means wrap at column LDIF_LINE_WIDTH
> >> >
> >> >-r alone means don't wrap at all
> >> 
> >> -r 0 instead.
> >> 
> >> 
> >> >-r <positive integer> means wrap at given column
> >
> >Is -r 3 really an interesting option, or is it just 
> code/feature bloat?
> 
> Some would argue that adding a no-line-wrap option to 
> ldapsearch(1) is just code/feature bloat as, in general, any 
> program purporting to accept LDIF input (or output of 
> ldapsearch(1)) is generally expected to handle LDIF, 
> including line wraps.
> 
> >Either you live in an LDIF conformant world and you want the 
> standard 
> >wrap, or you're trying to write a one line script, and you'd 
> rather not 
> >deal with it.
> 
> Well, if I was writing a one line script, I'd rather that 
> script (still only one line) properly handle LDIF line wraps 
> than to waste my time and others with getting users and 
> programs, not just ldapsearch(1), to produce output without 
> standard line-wraps.
> 
> >Is there really an interesting in-between usage case?
> 
> If a user say wants to view ldapsearch(1) output on a 
> terminal of width 50, I would think having the capability to 
> wrap at 50 would be useful to that user.

Or they could just use -r, and let the terminal wrap the lines. :) 
Anyone ever actually ask for it?

My vote would be kill it if you want to be pure, or just do the -r with
no columns.

mp