Time to abandon recursive pulling of dependencies?

[Stephen Montgomery-Smith]

Alexander Leidinger wrote:
> Quoting Stephen Montgomery-Smith <stephen at math.missouri.edu> (Wed, 16 May 2007 07:59:11 -0500):
> 
>> Alexander Leidinger wrote:
>>> Quoting Stephen Montgomery-Smith <stephen at math.missouri.edu> (from Tue, 
>>> 15 May 2007 16:53:35 -0500):
>>>
>>>> Ulrich Spoerlein wrote:
>>>>> Stephen Montgomery-Smith wrote:
>>>>>> 2.  Sorting the dependencies in pkg_create.  My fix now cuts this  
>>>>>> out completely.  Thus there is no need to change the structure of  
>>>>>> /var/db/pkg to use databases or anything like this.
>>>>> I haven't had time to look at your patches, but are your removing second
>>>>> order dependancies? This would cut down the tree size considerably.
>>>> No.  No second order dependencies are removed.  The new algorithm does
>>>> exactly the same as the old one did, only it works a lot faster.
>>> Would you please tell me what kind of output you need (which order of 
>>> pkgname & origin and which seperator between them), I want to provide an 
>>> awk script to you which produces this based upon +CONTENTS. It should 
>>> cut down some program invocations and a make call (IIR your patch 
>>> correctly).
>> pkgname:origin
> 
> ---snip---
> awk -F '( |:)' '/@pkgdep / { printf "%s:", $2 } /@comment DEPORIGIN:/ { printf "%s\n", $3 }' /var/db/pkg/yelp-2.18.1/+CONTENTS
> 
> [...]
> libbonoboui-2.18.0:x11-toolkits/libbonoboui
> firefox-2.0.0.3,1:www/firefox
> libgnomeui-2.18.1:x11-toolkits/libgnomeui
> ---snip---

Thanks.  But this program is very sensitive to errors in the +CONTENTS 
file.  Suppose that there is a @pkgdep without an accompanying 
DEPORIGIN.  I would want an output

pkgname