Hi!
I've noticed some quite nasty thing about
gcl 2.6.7 ansi build. After make install main
executable saved_ansi_gcl is partially broken.
"Partially" because the problems is not so easy
to see. This gcl builds Maxima, passes all tests -
everything seems to be OK. The problem is that
lisp error-debugger loop is broken. It is enough
to type unbound variable to see some weird
message instead of usual lisp debugger prompt.
The problem is caused by erroneous (reset-sys-paths)
executed during make install. Patch which fixes
the trouble is attached. I'm a bit surprised.
Why the bug didn't show up with gcl 2.6.6 and earlier?
--
Vadim V. Zhytnikov
<[EMAIL PROTECTED]>
<[EMAIL PROTECTED]>
diff -uNr gcl-2.6.7-orig/makefile gcl-2.6.7/makefile
--- gcl-2.6.7-orig/makefile 2004-08-06 02:49:39 +0400
+++ gcl-2.6.7/makefile 2005-08-31 22:27:43 +0400
@@ -173,8 +173,7 @@
if gcc --version | grep -i mingw >/dev/null 2>&1 ; then if grep -i
oncrpc makedefs >/dev/null 2>&1 ; then cp /mingw/bin/oncrpc.dll
$(DESTDIR)$(INSTALL_LIB_DIR)/$(PORTDIR); fi ; fi
cd $(DESTDIR)$(INSTALL_LIB_DIR)/$(PORTDIR) && \
mv $(FLISP)$(EXE) temp$(EXE) && \
- echo '(reset-sys-paths "$(INSTALL_LIB_DIR)/")\
- (si::save-system "$(FLISP)$(EXE)")' | ./temp$(EXE) && \
+ echo '(reset-sys-paths "$(INSTALL_LIB_DIR)/")(si::save-system
"$(FLISP)$(EXE)")' | ./temp$(EXE) && \
rm -f temp$(EXE)
if [ -e "unixport/rsym$(EXE)" ] ; then cp unixport/rsym$(EXE)
$(DESTDIR)$(INSTALL_LIB_DIR)/unixport/ ; fi
# ln $(SYMB) $(INSTALL_LIB_DIR)/$(PORTDIR)/$(FLISP)$(EXE) \
_______________________________________________
Gcl-devel mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/gcl-devel
