I *really* need help PLEASE - buildworld failing on mkdep libstdc++ can't find unwind.h but it *is* there
[Some of the maillists including -current@ are down at the moment. I hope someone on this list can help in meantime, please?] I've posted several msgs over the past two weeks stumbling on buildworld when using lang/gcc33. I've gone back to using system gcc by clearing out env-vars and ldconfig and /lib and /libexec of anything related to the lang/gcc33 (it did build some ports so the ldconfig has a last entry to find its libgcc etc.). I've rm -Rf everything under /src (/usr/src) and re-applied all CTM deltas thru early A.M. CDT today. I've downloaded the April 8 bzip2ball from current.freebsd.org and ran 'make installworld' on it in hopes to get a clean copy of world (yes there were a few tricks with symlinks to make it work). I always do 'rm -Rf /usr/obj' before building anything. The buildworld is still failing here: [...] building shared library libm.so.2 install -C -C -o root -g wheel -m 444 libm.a /usr/obj/src/i386/usr/lib install -C -s -o root -g wheel -m 444 libm.so.2 /usr/obj/src/i386/lib ln -fs /usr/obj/src/i386/lib/libm.so.2 /usr/obj/src/i386/usr/lib/libm.so install -C -C -o root -g wheel -m 444 /src/lib/msun/src/math.h /usr/obj/src/i386/usr/include === gnu/lib === gnu/lib/csu === gnu/lib/libgcc === gnu/lib/libdialog rm -f .depend mkdep -f .depend -a-I/src/gnu/lib/libdialog -DLOCALE /src/gnu/lib/libdialog/kernel.c /src/gnu/lib/libdialog/rc.c /src/gnu/lib/libdialog/checklist.c /src/gnu/lib/libdialog/inputbox.c /src/gnu/lib/libdialog/menubox.c /src/gnu/lib/libdialog/msgbox.c /src/gnu/lib/libdialog/lineedit.c /src/gnu/lib/libdialog/radiolist.c /src/gnu/lib/libdialog/textbox.c /src/gnu/lib/libdialog/yesno.c /src/gnu/lib/libdialog/prgbox.c /src/gnu/lib/libdialog/raw_popen.c /src/gnu/lib/libdialog/fselect.c /src/gnu/lib/libdialog/ui_objects.c /src/gnu/lib/libdialog/dir.c /src/gnu/lib/libdialog/notify.c /src/gnu/lib/libdialog/help.c /src/gnu/lib/libdialog/gauge.c /src/gnu/lib/libdialog/tree.c echo libdialog.so.4: /usr/obj/src/i386/usr/lib/libncurses.a .depend === gnu/lib/libregex rm -f .depend mkdep -f .depend -a-D_REGEX_RE_COMP -DSTDC_HEADERS -I/usr/obj/src/i386/usr/include/gnu /src/gnu/lib/libregex/regex.c === gnu/lib/libregex/doc === gnu/lib/libreadline === gnu/lib/libreadline/history rm -f .depend mkdep -f .depend -a-I/src/gnu/lib/libreadline/history/.. -I/src/gnu/lib/libreadline/history/../../../../contrib/libreadline -DHAVE_CONFIG_H -DRL_LIBRARY_VERSION='4.3' /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/history.c /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/histexpand.c /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/histfile.c /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/histsearch.c /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/shell.c /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/mbutil.c /src/gnu/lib/libreadline/history/../../../../contrib/libreadline/xmalloc.c === gnu/lib/libreadline/history/doc === gnu/lib/libreadline/readline rm -f .depend mkdep -f .depend -a-I/src/gnu/lib/libreadline/readline/.. -I/src/gnu/lib/libreadline/readline/../../../../contrib/libreadline -DHAVE_CONFIG_H -DRL_LIBRARY_VERSION='4.3' /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/readline.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/vi_mode.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/funmap.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/keymaps.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/parens.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/search.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/rltty.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/complete.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/bind.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/isearch.c /src/gnu/lib/libreadline/readline/../../../../contrib/libread line/display.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/signals.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/util.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/kill.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/undo.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/macro.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/input.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/callback.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/terminal.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/text.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/nls.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/misc.c /src/gnu/lib/libreadline/readline/../../../../contrib/libreadline/compat.c
Re: I *really* need help PLEASE - buildworld failing on mkdep libstdc++can't find unwind.h but it *is* there
Hi Chuck, - Original Message - From: Chuck Swiger [EMAIL PROTECTED] Date: Mon, 12 Apr 2004 16:02:53 -0400 To: Paul Seniura [EMAIL PROTECTED] Subject: Re: I *really* need help PLEASE - buildworld failing on mkdep libstdc++can't find unwind.h but it *is* there Paul Seniura wrote: [Some of the maillists including -current@ are down at the moment. I hope someone on this list can help in meantime, please?] I can confirm that your email is making it to the list, anyway... No one is getting -current@ since 'midnight' today (whatever timezone the mailserver uses ;) which is where I originally posted this same msg. About half the lists are down including CTM-via-email (FTP is current okay there). Simon Nielson from -www@ says the postmaster has been notified. -questions@ is working, fornunately. I've posted several msgs over the past two weeks stumbling on buildworld when using lang/gcc33. ...and I remember seeing these messages go by, too. :-) I'd seen another user ask about using the gcc ports, so I thought I'd take a plunge and report back. Trying to build -CURRENT with a compiler from a port rather than with the default compiler is a wizard-mode operation (unfortuantely). I suspect that you would have better luck tracking RELENG_5_2 or RELENG_4. I think I mentioned some time back that the port net/tn3270 I'm suppose to be working on needs to track 5-Current because it relies on copying the /src/contrib/telnet tree. And mentioned I'm trying to beat the deprecated functions, too, by making sure we can use the next planned release gcc3.3.4, even. Well, we cannot use 'em as I've said, just too many things need work. If you want to experiment with a gcc from ports, try rebuilding a specific application and benchmarking that before trying to do so with the entire world. If you wanted to perform a broader test, consider recompiling KDE and X11. Yeah, I know. But 'til someone finds time to help, I can't do it alone, _way_ too many problems. [ ... ] PLEASE Help - I do not want to blow everything away and start completely over, as Da Bosses are looking rather stern that this project is taking way too long. It is not clear to me what problem you are trying to solve by the activities you are pursuing: perhaps you ought to install 5.2.1 or 4.9 from a .iso image and get on with other tasks, and revisit the issue of recompiling world later? Nutshell: I've gone back to using the system gcc. Now we're not able compile libstdc++ and other related pieces; the headers _are_ there as mentioned in the earlier msg. Just about every other thing under world _does_ compile link properly -- it is just the libstdc-type stuff. We should be running an installworld made from April 8's bzip2all from current.freebsd.org, and fresh srcs from the CTM deltas up to this morning. I don't know what else to check on, I'm needing another pair of eyes. ;) -- -Chuck -- thx, Paul Seniura. -- ___ Sign-up for Ads Free at Mail.com http://promo.mail.com/adsfreejump.htm ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: I *really* need help PLEASE - buildworld failing on mkdep libstdc++can't find unwind.h but it *is* there
At 2:41 PM -0600 4/12/04, P.D. Seniura wrote: Chuck Swiger wrote: It is not clear to me what problem you are trying to solve by the activities you are pursuing: perhaps you ought to install 5.2.1 or 4.9 from a .iso image and get on with other tasks, and revisit the issue of recompiling world later? Nutshell: I have gone back to using the system gcc. But now we are not able compile libstdc++ and other related pieces; the headers _are_ there as mentioned in the earlier msg. Just about every other thing under world _does_ compile link properly -- it is just the libstdc-type stuff. ... I don't know what else to check on, I'm needing another pair of eyes. ;) I am not a gcc or gcc++ expert. I can offer the following observation, but don't ask me what it means. gcc is a major project in its own right, and I do not know the ins-and-outs of it. In your logfile, you have the sequence: === gnu/lib/libstdc++ sed -e ...etc... strstream-fixed.cc rm -f .depend mkdep -f .depend -a-DIN_GLIBCPP_V3 -DHAVE_CONFIG_H -I/src/gnu/lib/libstdc++ -I/src/gnu/lib/libstdc++/../../../contrib/libstdc++/libsupc++ -I/src/gnu/lib/libstdc++/../../../contrib/gcc /src/gnu/lib/libstdc++/../../../contrib/libstdc++/libmath/nan.c ...etc... mkdep -f .depend -a /src/gnu/lib/libstdc++/../../../contrib/libstdc++/src/bitset.cc ...etc... The second one does not have the -DIN_GLIBCPP_V3 -DHAVE_CONFIG_H or the three settings of -I. In a logfile of one of my own buildworlds, both of those mkdep's seem to start out with the same set of options. I expect the missing options are significant, but I do not know why they would be missing, or what to do about them. -- Garance Alistair Drosehn= [EMAIL PROTECTED] Senior Systems Programmer or [EMAIL PROTECTED] Rensselaer Polytechnic Instituteor [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
