Re: [avrdude-dev] Building Avrdude 6.3
> Did you “make clean” or otherwise clear out the directory at the same time as you deleted /opt? To completely resolve the build issue I had to completely remove the /opt tree, then remove all references to it from my $PATH setting and, finally, rebuild all the libraries I'd built with brew, as some of them had used the tools in /opt (which were quite old.) Frankly, that's one of things I dislike about solutions like macPorts and brew is that it can be hard to know just exactly what is going on under the hood when you invoked one of these automated build scripts. I'm also not that skilled with using the command line to do things, which tends to lead to panic, or despair when something doesn't work. Wayne On Sat, Mar 23, 2019 at 9:05 PM Joel Ray Holveck wrote: > Did you “make clean” or otherwise clear out the directory at the same time > as you deleted /opt? > > Did Joerg’s suggestion help? > > If the answer is “no” and “yes” respectively, then I offer the following > hypothesis. > > The Unix build process doesn’t really have a good way to see that a > previous ranlib step failed, so if it sees that the .a file exists, it > assumes that ranlib succeeded. (ranlib only changes the structure of an > existing .a file; on systems that need it, it’s generally run right after > the .a is created in the build process.) When you reran make, it thought > that ranlib succeeded previously, and that assumption broke the later build > steps. > > In Joerg’s case, it sounds like the mingw32 cross-build might not properly > detect that it needs to run ranlib. It’s not needed, for instance, on > Linux. > > joelh > > > On Mar 22, 2019, at 23:49, Joerg Wunsch wrote: > > > > > > > > Am 23. März 2019 03:14:53 MEZ schrieb Wayne Holder < > wayne.hol...@gmail.com>: > > > >> libtool: link: gcc -Wall -g -O2 -o avrdude avrdude-main.o > >> avrdude-term.o > >> ./libavrdude.a -lusb-1.0 -lusb -framework CoreFoundation -framework > >> IOKit > >> -lftdi1 -lelf -lpthread -lm -lreadline -lncurses -ltermcap > >> > >> ld: archive has no table of contents file './libavrdude.a' for > >> architecture > >> x86_64 > > > > I have a similar issue when cross-compiling the win32 version using > mingw32. I never found the actual reason, alas. > > > > Try > > > > ranlib libavrdude.a > > > > and then "make" again. > > > > > > -- > > Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet. > > > > ___ > > avrdude-dev mailing list > > avrdude-dev@nongnu.org > > https://lists.nongnu.org/mailman/listinfo/avrdude-dev > > ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
As Matthew Mondor wrote: > > Maybe the OSX linker needs it, I don't know. > > It's quite possible some ranlib magic is still missing in the current > > make/automake files. > > Very likely, Darwin being a BSD variant, all the BSDs that I know need > it as well. My main development system is FreeBSD, and the current make/automake files work there as well as on Linux. I only see that ranlib issue when cross-compiling for mingw32. Anyway, it was just an idea that occurred to me. Quite possible Wayne's problem is different. -- cheers, Joerg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/ Never trust an operating system you don't have sources for. ;-) ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
On Sun, 24 Mar 2019 08:52:28 +0100 Joerg Wunsch wrote: > Maybe the OSX linker needs it, I don't know. > It's quite possible some ranlib magic is still missing in the current > make/automake files. Very likely, Darwin being a BSD variant, all the BSDs that I know need it as well. -- Matt ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
Am 24. März 2019 05:05:39 MEZ schrieb Joel Ray Holveck : >In Joerg’s case, it sounds like the mingw32 cross-build might not >properly detect that it needs to run ranlib. It’s not needed, for >instance, on Linux. Maybe the OSX linker needs it, I don't know. It's quite possible some ranlib magic is still missing in the current make/automake files. -- Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet. ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
Did you “make clean” or otherwise clear out the directory at the same time as you deleted /opt? Did Joerg’s suggestion help? If the answer is “no” and “yes” respectively, then I offer the following hypothesis. The Unix build process doesn’t really have a good way to see that a previous ranlib step failed, so if it sees that the .a file exists, it assumes that ranlib succeeded. (ranlib only changes the structure of an existing .a file; on systems that need it, it’s generally run right after the .a is created in the build process.) When you reran make, it thought that ranlib succeeded previously, and that assumption broke the later build steps. In Joerg’s case, it sounds like the mingw32 cross-build might not properly detect that it needs to run ranlib. It’s not needed, for instance, on Linux. joelh > On Mar 22, 2019, at 23:49, Joerg Wunsch wrote: > > > > Am 23. März 2019 03:14:53 MEZ schrieb Wayne Holder : > >> libtool: link: gcc -Wall -g -O2 -o avrdude avrdude-main.o >> avrdude-term.o >> ./libavrdude.a -lusb-1.0 -lusb -framework CoreFoundation -framework >> IOKit >> -lftdi1 -lelf -lpthread -lm -lreadline -lncurses -ltermcap >> >> ld: archive has no table of contents file './libavrdude.a' for >> architecture >> x86_64 > > I have a similar issue when cross-compiling the win32 version using mingw32. > I never found the actual reason, alas. > > Try > > ranlib libavrdude.a > > and then "make" again. > > > -- > Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet. > > ___ > avrdude-dev mailing list > avrdude-dev@nongnu.org > https://lists.nongnu.org/mailman/listinfo/avrdude-dev ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
Am 23. März 2019 03:14:53 MEZ schrieb Wayne Holder : >libtool: link: gcc -Wall -g -O2 -o avrdude avrdude-main.o >avrdude-term.o >./libavrdude.a -lusb-1.0 -lusb -framework CoreFoundation -framework >IOKit >-lftdi1 -lelf -lpthread -lm -lreadline -lncurses -ltermcap > >ld: archive has no table of contents file './libavrdude.a' for >architecture >x86_64 I have a similar issue when cross-compiling the win32 version using mingw32. I never found the actual reason, alas. Try ranlib libavrdude.a and then "make" again. -- Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet. ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
On Mar 22, 2019, at 13:30, Wayne Holder wrote: > > My system is up to date with the latest version of > Xcode and is running macOS 0.14.2 (18C54) "Mojave". Just to make sure: have you launched Xcode’s GUI since you last updated it? That will set up many of the command-line tools that are needed for builds. joelh ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
I guess some of my problems came from having once used macPorts. So, I deleted the /opt folder, fixed my $PATH and then reran configure and make. This gets me further, but still no cigar. Tha current error is now: libtool: link: ( cd ".libs" && rm -f "libavrdude.la" && ln -s "../ libavrdude.la" "libavrdude.la" ) gcc -DHAVE_CONFIG_H -I. -DCONFIG_DIR=\"/usr/local/etc\" -Wall -g -O2 -MT avrdude-main.o -MD -MP -MF .deps/avrdude-main.Tpo -c -o avrdude-main.o `test -f 'main.c' || echo './'`main.c mv -f .deps/avrdude-main.Tpo .deps/avrdude-main.Po gcc -DHAVE_CONFIG_H -I. -DCONFIG_DIR=\"/usr/local/etc\" -Wall -g -O2 -MT avrdude-term.o -MD -MP -MF .deps/avrdude-term.Tpo -c -o avrdude-term.o `test -f 'term.c' || echo './'`term.c mv -f .deps/avrdude-term.Tpo .deps/avrdude-term.Po /bin/sh ./libtool --tag=CC --mode=link gcc -Wall -g -O2 -o avrdude avrdude-main.o avrdude-term.o ./libavrdude.a -lusb-1.0 -lusb -framework CoreFoundation -framework IOKit -lftdi1 -lelf -lpthread -lm -lreadline -lncurses -ltermcap libtool: link: gcc -Wall -g -O2 -o avrdude avrdude-main.o avrdude-term.o ./libavrdude.a -lusb-1.0 -lusb -framework CoreFoundation -framework IOKit -lftdi1 -lelf -lpthread -lm -lreadline -lncurses -ltermcap ld: archive has no table of contents file './libavrdude.a' for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make[2]: *** [avrdude] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 Wayne On Fri, Mar 22, 2019 at 6:55 PM Wayne Holder wrote: > It appears I had an old copy of gm4 in opt/local/bin and, when I removed > it, I got much further in the process. I now have only the following > errors printed: > > ar cru libavrdude.a libavrdude_a-config_gram.o libavrdude_a-lexer.o > libavrdude_a-arduino.o libavrdude_a-avr.o libavrdude_a-avr910.o > libavrdude_a-avrftdi.o libavrdude_a-avrftdi_tpi.o libavrdude_a-avrpart.o > libavrdude_a-bitbang.o libavrdude_a-buspirate.o libavrdude_a-butterfly.o > libavrdude_a-config.o libavrdude_a-confwin.o libavrdude_a-crc16.o > libavrdude_a-dfu.o libavrdude_a-fileio.o libavrdude_a-flip1.o > libavrdude_a-flip2.o libavrdude_a-ft245r.o libavrdude_a-jtagmkI.o > libavrdude_a-jtagmkII.o libavrdude_a-jtag3.o libavrdude_a-linuxgpio.o > libavrdude_a-lists.o libavrdude_a-par.o libavrdude_a-pgm.o > libavrdude_a-pgm_type.o libavrdude_a-pickit2.o libavrdude_a-pindefs.o > libavrdude_a-ppi.o libavrdude_a-ppiwin.o libavrdude_a-safemode.o > libavrdude_a-serbb_posix.o libavrdude_a-serbb_win32.o > libavrdude_a-ser_avrdoper.o libavrdude_a-ser_posix.o > libavrdude_a-ser_win32.o libavrdude_a-stk500.o libavrdude_a-stk500v2.o > libavrdude_a-stk500generic.o libavrdude_a-usbasp.o > libavrdude_a-usb_libusb.o libavrdude_a-usbtiny.o libavrdude_a-update.o > libavrdude_a-wiring.o > /opt/local/bin/ranlib: object: libavrdude.a(libavrdude_a-config_gram.o) > malformed object (unknown load command 1) > ar: internal ranlib command failed > make[2]: *** [libavrdude.a] Error 1 > make[1]: *** [all-recursive] Error 1 > make: *** [all] Error 2 > > Wayne > > > On Fri, Mar 22, 2019 at 4:55 PM Joel Ray Holveck wrote: > >> On Mar 22, 2019, at 13:30, Wayne Holder wrote: >> > >> > My system is up to date with the latest version of >> > Xcode and is running macOS 0.14.2 (18C54) "Mojave". >> >> Just to make sure: have you launched Xcode’s GUI since you last updated >> it? That will set up many of the command-line tools that are needed for >> builds. >> >> joelh >> > ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
It appears I had an old copy of gm4 in opt/local/bin and, when I removed it, I got much further in the process. I now have only the following errors printed: ar cru libavrdude.a libavrdude_a-config_gram.o libavrdude_a-lexer.o libavrdude_a-arduino.o libavrdude_a-avr.o libavrdude_a-avr910.o libavrdude_a-avrftdi.o libavrdude_a-avrftdi_tpi.o libavrdude_a-avrpart.o libavrdude_a-bitbang.o libavrdude_a-buspirate.o libavrdude_a-butterfly.o libavrdude_a-config.o libavrdude_a-confwin.o libavrdude_a-crc16.o libavrdude_a-dfu.o libavrdude_a-fileio.o libavrdude_a-flip1.o libavrdude_a-flip2.o libavrdude_a-ft245r.o libavrdude_a-jtagmkI.o libavrdude_a-jtagmkII.o libavrdude_a-jtag3.o libavrdude_a-linuxgpio.o libavrdude_a-lists.o libavrdude_a-par.o libavrdude_a-pgm.o libavrdude_a-pgm_type.o libavrdude_a-pickit2.o libavrdude_a-pindefs.o libavrdude_a-ppi.o libavrdude_a-ppiwin.o libavrdude_a-safemode.o libavrdude_a-serbb_posix.o libavrdude_a-serbb_win32.o libavrdude_a-ser_avrdoper.o libavrdude_a-ser_posix.o libavrdude_a-ser_win32.o libavrdude_a-stk500.o libavrdude_a-stk500v2.o libavrdude_a-stk500generic.o libavrdude_a-usbasp.o libavrdude_a-usb_libusb.o libavrdude_a-usbtiny.o libavrdude_a-update.o libavrdude_a-wiring.o /opt/local/bin/ranlib: object: libavrdude.a(libavrdude_a-config_gram.o) malformed object (unknown load command 1) ar: internal ranlib command failed make[2]: *** [libavrdude.a] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 Wayne On Fri, Mar 22, 2019 at 4:55 PM Joel Ray Holveck wrote: > On Mar 22, 2019, at 13:30, Wayne Holder wrote: > > > > My system is up to date with the latest version of > > Xcode and is running macOS 0.14.2 (18C54) "Mojave". > > Just to make sure: have you launched Xcode’s GUI since you last updated > it? That will set up many of the command-line tools that are needed for > builds. > > joelh > ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
Thanks for trying to help me. When I run the command you suggested: > /bin/sh ./ylwrap lexer.l lex.yy.c lexer.c -- flex I get an immediate "gm4 had quit unexpectedly" message and no other response on the command line. And, when I then run configure and make, I get the same result as before. I also tried running brew to install flex, byacc and bison and then retried the configure and make, but that also produced no change. Trying to use brew to install lex, but got Error: No available formula with the name "lex" > You'll probably need the patches mentioned by David Mosberger. Which build should I download to get these patches? Also, any thought on how to a fix the "quit unexpectedly" messages? Wayne On Fri, Mar 22, 2019 at 2:47 PM Joerg Wunsch wrote: > As Wayne Holder wrote: > > > As part of trying to diagnose the issue I'm having with serbb and the TPI > > protocol (see my other thread on this topic), I decided to try and build > > the 6.3 release on my Mac Pro system and then see if I could resolve the > > issue by making a few tweaks to the timing bit timing. > > You'll probably need the patches mentioned by David Mosberger. > > Anyway, ... > > > /bin/sh ./ylwrap lexer.l lex.yy.c lexer.c -- flex > > > > make: *** [lexer.c] Error 1 > > Do you have flex and bison or [b]yacc installed? > > Alas, autoconf (which generates the configure script) has no > precautions to really test for the presence of flex or lex - it can > only distinguish which one of both it is. That's why there is no > better error message for it. > > -- > cheers, Joerg .-.-. --... ...-- -.. . DL8DTL > > http://www.sax.de/~joerg/ > Never trust an operating system you don't have sources for. ;-) > > ___ > avrdude-dev mailing list > avrdude-dev@nongnu.org > https://lists.nongnu.org/mailman/listinfo/avrdude-dev > ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev
Re: [avrdude-dev] Building Avrdude 6.3
As Wayne Holder wrote: > As part of trying to diagnose the issue I'm having with serbb and the TPI > protocol (see my other thread on this topic), I decided to try and build > the 6.3 release on my Mac Pro system and then see if I could resolve the > issue by making a few tweaks to the timing bit timing. You'll probably need the patches mentioned by David Mosberger. Anyway, ... > /bin/sh ./ylwrap lexer.l lex.yy.c lexer.c -- flex > > make: *** [lexer.c] Error 1 Do you have flex and bison or [b]yacc installed? Alas, autoconf (which generates the configure script) has no precautions to really test for the presence of flex or lex - it can only distinguish which one of both it is. That's why there is no better error message for it. -- cheers, Joerg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/ Never trust an operating system you don't have sources for. ;-) ___ avrdude-dev mailing list avrdude-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avrdude-dev