>>>> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org> writes:
>>>> "ms" == Michael Sperber
<sperber(a)informatik.uni-tuebingen.de> writes:
Stephen> I have. Carefully. But I don't have access to your systematic
Stephen> thoughts, as there is no design document.
The design is documented in the XEmacs manual under "Startup Paths."
I admit that not every specific decision as it pertains to each
specific environment is covered there. I wouldn't know how to do
that.
Stephen> What I expected is that XEmacs would look for a "package root" (ie,
a
Stephen> directory containing one or more *-packages subdirectories, and not
Stephen> necessarily the same as an "XEmacs root") named $datadir/xemacs.
Stephen> That's the historical location for XEmacs-related stuff not
Stephen> XEmacs-version- or architecture-specific.
Yes, indeed, if XEmacs can make a reasonably assumption that there is
a fully populated $datadir part of an XEmacs root hierarchy. (The
manual explains the concept.) In this case, there isn't.
Stephen> In a vanilla-configured XEmacs, that is
Stephen> /usr/local/lib/xemacs.
Yes, and that's precisely what you wanted me to change. I can
probably dig up the thread if you want.
Stephen> Can you point me to some of these bug reports? And if it is not
Stephen> obvious from the reports themselves, explain why the preferred
Stephen> solution in those environments is not --with-prefix=no?
These people have an installed *old* XEmacs and their compiled
in-place *new* XEmacs is picking up its packages.
--
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla