This is the mail archive of the cygwin-apps@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: nano - packaged and ready for some criticism :)


Christopher Faylor wrote:


>>*Well* beyond the newbie's capability.  But "reinstall the info package" or 
>>"run fix-info" is nice and simple.
>>
> 
> Sure.  If you've got a solution, I have no objections.  I just didn't
> think it was a big deal, either way.
> 
> And, ur, I've been one of those screwed up package maintainers.  I don't
> even currently rebuild dir in any of my packages and I really should.
> 
> Or, actually, I think either setup.exe should be intelligent enough to
> do this for you or there should be some way for a package to say "please
> run this standard bit of machinery for me".


You mean setup.exe itself should do:
   if tarball contains /usr/info/*, then
     install-info --dir-file=/usr/info/dir --info-file=$(each one)
   end

I disagree.  AFAIK, RPM doesn't do this.  I believe you have to do:

%post
install-info .....

in the spec file.  Which is analogous to each of *our* packages that have 
info files providing an /etc/postinstall/pkg.sh script to handle the 
operation for its own info files.

It's a bit painful to munge a package's Makefiles to install a "random" 
file like "autoconf.postinstall" into "/etc/postinstall/" -- but both my 
proposal and Robert's proposal use an *external* script/makefile to drive 
package building -- and in THAT script/makefile, installing the postinstall 
file is a cinch.

Point: energy barrier for maintainers to do this sort of thing will be 
lower after the new -src standard is specified.  Especially if folks just 
copy/adapt the build scripts that Robert and I have been throwing around.

--Chuck



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]