Mail Archives: djgpp-workers/2000/08/13/06:39:58
Hello.
Laurynas Biveinis wrote:
>
> Eli Zaretskii wrote:
> > > You can specify in DSM `install-info' as a post-installation command
> > > to execute.
> >
> > Yes, but what I meant to say that the command to invoke install-info
> > is rather generic, so zippo could have done that automatically.
>
> I recall I've objected it earlier; however now I don't see why
> adding something like
>
> install-info-docs: gcc.inf
>
> to DSM would make things worse. Rich?
I'm not sure that adding yet another directive to the DSM specification is
a good idea. I think we have all we need right now.
Anyway, from reading the earlier discussion, the viewpoints were as
follows:
* Eli thought (& still thinks) that zippo should do this, since it's so
standard.
* Laurynas thought that zippo should be kept as generic as possible and
that this operation should be perform using a post-install-script.
* I (Rich) was undecided. I hadn't used install-info at the time of the
discussion.
Now that I've used install-info, I agree that zippo could do this
automatically for files with filenames of the form info/*.info and
info/*.inf. My concerns are:
* How do you know info/*.inf is an info file and not, for example, a
Windows .inf file? Unlikely, but AFAICT there's no easily identifiable
magic string at the start of an info file.
[ Incidentally, is there a port of the 'file' program ]
[ (available on Linux) to DJGPP? ]
* In the future zippo may not be the only DJGPP installer (*). If zippo
installs info docs automatically rather than making people using
post-install-script, then we set a precedent. A better way to fix this
problem is to write the code for zippo to support post-install-script.
(*) A fair proportion of zippo is libzippo, which could be used in other
installers, e.g. Laurynas's.
Personally I think that supporting 'post-install-script' is the best plan.
Bye, Rich =]
--
Richard Dawe
[ mailto:richdawe AT bigfoot DOT com | http://www.bigfoot.com/~richdawe/ ]
- Raw text -