Jérôme Marant <[EMAIL PROTECTED]> writes: > Le jeudi 26 octobre 2006 21:06, David Kastrup a écrit : > >> It _is_ already true. It is not possible to find anybody on either >> the Emacs or XEmacs developer lists who could understand or use the >> the Debian packaging implementation. I tried, as AUCTeX >> maintainer. That has nothing to do with the doc/code split. I am >> just saying that one need not consider maintainers/developers of >> Emacs or XEmacs based applications since they can't work with the >> Debian packaging, anyway. Asking on either developer list will >> have people tell you "I tried a few times, then gave up". So >> active Emacs/XEmacs package developers are out of the loop, anyway. >> The Debian packages cater to users only. > > David, you keep complaining about the way Emacs is packaged in > Debian. We got that. But you never came up with solutions that > would fullfil both you and the practical benefits we've tried to > offer our users through the current implementation so far (even if > you think we failed to do so).
It is not like I have not mentioned them several times. For one thing, don't play games with the directory structure of the Emacsen. XEmacs has a package system with directories. Adding another tree for Debian packages that is late in the load order is perfectly well supported. Similarly, adding a directory or tree to search in Emacs is perfectly well supported, too. Debian installs source Elisp files and the compiled files into different directories. This is a mistake. Call the command M-x list-load-path-shadows RET and see how Emacs and XEmacs complain. Whether you copy the files or symlink them: the Elisp files that Emacs/XEmacs find have to be in the same directories as the corresponding compiled Lisp files, and certainly in the same position in load-path. That is a fundamental assumption of both Emacs and XEmacs that gets violated by the Debian packaging system. As a result, installing user packages into the (XEmacs) user package tree overriding Debian packages is not really feasible. > Now, how about you describe straight how you'd see it to be > implemented and we begin constructive discussions instead, please? > Thanks. It is not like I have not explained this several times already. -- David Kastrup, Kriemhildstr. 15, 44793 Bochum -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
