The ebuild has this: if use emacs; then einfo "Configuring gnuplot-mode for GNU Emacs ..." cd "${S}/lisp" econf --with-lispdir="${EPREFIX}${SITELISP}/${PN}" EMACS=emacs fi but upstream does not have (or no longer has?) a configure in the lisp directory. Removing the block I quoted above fixes the USE=emacs build. I don’t have Xemacs installed, so I cannot test the USE=xemacs case, but something will be needed for that, too.
(In reply to comment #0) > The ebuild has this: > > if use emacs; then > einfo "Configuring gnuplot-mode for GNU Emacs ..." > cd "${S}/lisp" > econf --with-lispdir="${EPREFIX}${SITELISP}/${PN}" > EMACS=emacs > fi > > but upstream does not have (or no longer has?) a configure in the lisp > directory. > > Removing the block I quoted above fixes the USE=emacs build. > > I don’t have Xemacs installed, so I cannot test the USE=xemacs case, but > something will be needed for that, too. I am not the emacs expert, I just dropped xemacs support as building gnuplot 2x just for the el files seems a bit much. @ulm: Ok?
(In reply to comment #1) > I am not the emacs expert, I just dropped xemacs support as building gnuplot > 2x just for the el files seems a bit much. We could use the functions from elisp-common.eclass and xemacs-elisp-common.eclass to byte-compile the elisp files for both Emacs flavours. It's five cases that we must handle, namely all possible combinations of the emacs and xemacs USE flags, plus some special treatment for the live ebuild with USE="-emacs -xemacs". I can work on this, but I first I'd like to have some feedback if gnuplot.el is still working for XEmacs. > @ulm: Ok? Fine with me, but I'm the wrong person to ask. @graaff: Hans, what do you think?
I haven't looked at this particular ebuild since it is in an overlay. But in general, I don't understand what the problem is here. I don't see how things are being built twice, and even if they are: who cares. The group of people with both USE=emacs and USE=xemacs must be really small since in general people will pick one of the two. I also don't see how the removal of the code mentioned in the bug fixes things. is the emacs code actually still being built at all? Note that technically I'm not the right person to ask either since I'm no longer xemacs maintainer and I no longer have it installed... I've cc'ed Mats who hopefully will be xemacs maintainer soon.
Removing the quoted code from the ebuild fixes things because upstream removed the separate configure script from the elisp subdirectory; the top level configure script now handles the elisp subdir just like it does the other subdirectories.
I've committed gnuplot-4.6.1-r1.ebuild to the main tree, with much simplified Emacs support (because of bug 459172). Especially, no more configure is done in the lisp subdirectory. Maybe these changes could be tested for the live ebuild, too?
Changes from 4.6.1-r1 added to 4.6.9999 and 4.7.9999.