Re: diffutils 2.8.1 ISO C90 compliance patch

[Paul Edwards]

From: "Dan Hipschman" <address@hidden>
> > The C90 standard exists precisely because there is no way to
> > do it.  It is the common subset of all environments.  There's
> > nowhere in the C90 standard that says how you can invoke an
> > external program and capture the output.  You can't do it.  Not
> > portably, and not on MVS.  You can however pass in parameters
> > of filenames.
> 
> FWIW, a while ago I did some work on abstracting the diff functionality
> into a diff library, but on a lower level than what diff3 would use.
> However, the ultimate goal of the project was to provide diff as a
> library so programs like diff3 and wdiff didn't have to exec diff, and
> more importantly, programs like cvs and git didn't have to statically
> link with older versions of the diff sources all in the name of
> performance.  So I think diff3 could be written without using pipes this
> way, but if we do it like that, it should be done right so everybody can
> use the diff library.

Yes, I agree.  This is the best way to do it.  It just takes a lot
more work than my simple change.

BFN.  Paul.