Package dependencies

James Robertson jameswrobertson at earthlink.net
Tue Jun 17 13:59:45 PDT 2003


Tushar Teredesai wrote:
> James Robertson wrote:
> 
<snip>>
> 
> Figuring out the dependencies using the above approach would be huge 
> programming task, with no guarantees (IMO). How would a script know if 
> something in the configure script is a program or just some text. Also 
> grepping thru the source would not give you all library dependencies and 
> header files used.
> 
> It would be easier to use strace to figure out the dependencies and 
> write a script to read the strace results to find the dependencies.
> 
Thanks Tushar

That is interesting.  It does not seem to me that our dependancy lists 
are like that.  They show commands that are easily identifiable (via 
grep) in configure scripts and makefiles.  As for library dependancies, 
that is another issue.  Strace is probably a better tool for that.

Mostly I am after an easier way to ensure that the toolbox we build in 
ch5 has all the dependancies needed for all the packages we compile in 
ch6.  And, of course, if we can't or don't want to install a package in 
ch5 for a ch6 package(s), then find a way to "fake" it for the package 
in ch6.  An example of this is the inetutils bug that Greg Schager just 
recently pointed out.

I was just going to write a script with a really long case statement in 
it with output to a log file.  Maybe this approach won't work.  I also 
would need to take a closer look at strace to see what its output looks 
like.

<snip>

What do you think?

-- 
--------------------------------------------------------
James Robertson | jameswrobertson at earthlink.net
--------------------------------------------------------
Reg. Linux User: #160424  http://counter.li.org
Reg. LFS User:   #6981    http://www.linuxfromscatch.org
--------------------------------------------------------


-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list