"Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> wrote:
I don't understand why you don't have anything in your
module-load-path, but what's in mine (and Yoshiki's) I believe is
coming because there is no startup code in current CVS to set
module-load-path; it's only done at dump time AFAICT. I wonder if
maybe Yoshiki's got a modules -> $srcdir/modules symlink in the build
directory?
Maybe you fixed this already and forgot to diff this file?
A quick experiment on my end seems to indicate that the value dumped in
module-load-path is different in these two situations:
- a (previous) build of XEmacs is installed in the place where this
build will eventually be installed (it's enough for the directories
to exist);
- the directories will be created when XEmacs
I was puzzled about this, too, so I rebuilt it ... and module-load-path
was not nil. But then it occurred to me...
- The first time I built XEmacs 21.2 and got nil in my
module-load-path, I didn't have an installed XEmacs 21.2 tree.
- The second time I built XEmacs 21.2 and got a non-nil
module-load-path,
I had an installed XEmacs 21.2 tree (from the first build).
A quick experiment seems to confirm this: the value dumped into
module-load-path is affected by the existence of directories in the
location where XEmacs will be installed once it is dumped. Ugh!
I'll note that this is *with* my last patch, though. I'll try backing
the patch out and see if the same thing happens. I suspect it will,
though, since it was getting nil the first time that took me down the
path to creating that patch.
--
Jerry James