About six weeks ago I brought up the issue of making a system-ghostscript package (message reproduced below). Jeff objected (also reproduced below) but I would like to revisit the issue, because there is a forthcoming OzTeX CD which will have Gerben Wierda's teTeX/ghostscript installer on it, as well as a Fink installer, and I'm trying to set things up so that users of this CD can easily install Fink's TeX-related packages without having to go online to do so.
Jeff proposed a policy about system-foo packages: we should only allow them if the package they are replacing is huge (like system-xfree86 or system-tetex). I would like to propose that we could have them if the package is huge, OR if the package has a Restrictive license (like ghostscript). In the case of a Restrictive license, we won't ever be able to distribute the deb files via Fink; but users might have legitimately obtained that package another way with an appropriate license, so we should accomodate them if possible. Anyway, Jeff's other objection about this particular case was that the system-ghostscript package would be fragile and easily broken. I'm willing to assume responsibility for that, maintaining the package and jumping in to help users where appropriate. Other opinions about this would be welcome. -- Dave > On Thu, 28 Feb 2002, David R. Morrison wrote: > > > I have written a system-ghostscript package, but I am undecided about > > whether it should be included in fink or not. > > > > As some of you know, there is a widely-used teTeX distribution (which > > accompanies the program TeXShop), and we support this as an alternative > > with fink's system-tetex package. The installer which is used for this > > teTeX distribution also installs a copy of ghostscript 6.01 in /usr/local, > > and there have been periodic requests from users for a system-ghostscript > > package that recognizes this installation. > > > > The package I wrote sets up lots of symbolic links: each of the executables > > is individually linked from its location in /usr/local/bin to /sw/bin; > > each of the man pages is individually linked from its location in > > /usr/local/man/man1 to /sw/share/man/man1; and the rest of ghostscript > > is linked directory by directory from /usr/local/share/ghostscript to > > appropriate places in the fink hierarchy. > > > > This would certainly be convenient for users who have installed teTeX > > and ghostscript together. The question is, though: is this setting a > > bad precedent? We certainly don't want to have lots of fink packages > > which set up symbolic links to things that were installed within > > /usr/local by some other installation method. > > > > I won't submit the package until I've gotten some feedback on this. > > > > -- Dave > > > > Jeff Whitaker <[EMAIL PROTECTED]> wrote: > > David: I would vote no. The reason I never did this (even though I was > asked to several times) is exactly the one you mentioned - these > system-foo packages that install symlinks are very inelegant and easily > broken. I think we should only do it in those cases (like tetex and > XFree86) where the packages are huge, and very painful to install twice. > I don't think ghostscript falls into that category (my ghostscript6 deb > file is under 2mb). > > -Jeff > _______________________________________________ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
