Re: Suggestion: Allow cleanup code in AC_LANG_POP

OK, forget it - I just found out that it doesn't work he way I expected. Sorry to bother, Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Suggestion: Allow cleanup code in AC_LANG_POP

ANG_ABBREV[]_whatever = ... I think this is a clean solution that does not affect running code but might add more opportunities for language switching. Regards Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-3

Bug in AC_PROG_F77_C_O ??

heers, Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

exit status of conftest program

Is it ok for autoconf tests to create a C test program such as this if ( == ) exit (0); else exit (1); and check the exit status with AC_TRY_RUN, or may this fail on some systems? -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D

Re: Strange warnings from AC_REQUIRE

se we might # AC_REQUIRE `AC_LANG_COMPILER(C)' for instance, the latter must be # AC_DEFUN'd, not just define'd. define([AC_LANG_COMPILER], [_AC_LANG_DISPATCH([$0], _AC_LANG, $@)]) in aclang.m4 is sort of misleading, I think. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tr

Strange warnings from AC_REQUIRE

Hi, after two months, I'm trying to get my Fortran/cpp and Fortran 90 macros to work with cvs autoconf. So far I have utterly failed. I was able to track down (at least part of) my problems to this trivial configure.in, totally unrelated to Fortran stuff: AC_INIT(fpptest, 0.0.1, [EMAIL PROTECT

Re: sh compatibility question

cking tool. However, I still have to explore other possibilities. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361 ___ Autoconf mailing

sh compatibility question

I wonder why autoconf doesn't use shell functions. Is that a compatibility issue? According to the bash info pages, shell functions belong to the POSIX standard for the Bourne shell. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04

Re: aclang.m4 problem (Cygwin + VC++)

docs I included in my message simply say that it is impossible to cancel builtin suffix rules without resorting to pattern rules or explicit rules. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: aclang.m4 problem (Cygwin + VC++)

ith no commands. [ End GNU make info ] GNU make at least provides the command-line option "-r" to to cancel builtin rules, but I have no idea if this is portable. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: aclang.m4 problem (Cygwin + VC++)

on some systems initially) and must be explicitly activated by maintainers by calling, say, AC_CHECK_SOURCE_SUFFIX before doing compiler checks. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: aclang.m4 problem (Cygwin + VC++)

r my .F/.f stuff, because the first rtule will be ignored by make if CXXEXT != .cc and if there is a built-in rule .cc.o. Your first suggestion seems threfore more robust to me, because it rewrites the .cc.o rule. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Perm

Re: aclang.m4 problem (Cygwin + VC++)

lly like to know why at all compilers are so picky about file suffixes - if a file has correct syntax, what the heck does the suffix matter? -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: aclang.m4 problem (Cygwin + VC++)

es - IMO that's all that counts. (unlike your old xlf :-( ) >From what you wrote so far, this compiler doesn't beat xlf. I think it wouldn't even deserve a top 3 rank in the "weirdest fortran compilers" list. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: 25-ac-lang-compiler-inext

0 extension in some way :-( Could you perhaps consult your manual and find out whether that old version had an equivalent to the -qsuffix option? Regards Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: aclang.m4 problem (Cygwin + VC++)

upports it in the make environment, > not in autoconf). I'd be grateful if you detail a bit about this compiler's peculiarities. To date, xlf is still my favourite, perhaps Microsoft can beat him... :-) I don't work with M$ systems at all (lucky me!), so I have no idea. -

Re: -c -o

.o: $(CC) $(CFLAGS) ... @CC_C_O@ $< where CC_C_O would either be "-c -o $@" or simply "-c". -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Fortran dependency tracking

at all in automake/autoconf, - if yes, which solution we should go for ? Regards, Martin P.S. This problem seems to be trivial compared to providing dependency tracking support for Fortran 90 modules. But that's a different issue... -- Martin Wilck <[EMAIL PROTECTED]> I

-c -o

ds for C/C++, unless I'm completely mistaken. If such a test can make sense, then automake must provide a configure variable that autoconf can insert in the Makefile in order to assert correct compilation behaviour, right ? Regards Martin -- Martin Wilck <[EMAIL PROTECTED]> Inst

Re: Help wanted: compile error detection

Dear Thomas E. Dickey, you wrote on Today: > I've seen some compilers generate a zero-length (or nearly-zero ;-) > object file on error. (no, I don't recall which ones) That's why we should use test -s rather than test -f, like AC_LINK_IFELSE. -- Martin Wilck <[EMA

Re: Help wanted: compile error detection

esn't - that should work. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: Help wanted: compile error detection

;t it be smart to include "test -x" in AC_LINK_IFELSE as well? Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: Help wanted: compile error detection

ce that tests whether an output file was generated. There may be better ones, though. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Help wanted: compile error detection

I have come across one compiler that returns an exit status of 0 if compile errors occur :-{ Thus, AC_COMPILE_IFELSE can't be used to test this compiler's features. Suggestions? -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318

Re: Autoconf Fortran suggestions

regular basis. This may result in autoconf finding a GNU compiler, trying to use it, and fail. The Linux systems are the only ones where the GNU software is always cutting edge. I don't know whether that is typical, but it might be an (albeit poor) argument in favour of f77. Regards, Martin --

Re: AC_F77_LIBRARY_LDFLAGS patch (again:)

t consider that stuff very important, it just solves a minor logical quirk that probably didn't pose any practical problem, and gets rid of some non-"ac" shell variables. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig

Autoconf Fortran suggestions

"F77=f77 ./configure" all the time. Therefore, I suggest to put "g77" after "f77" and perhaps after other names of good native Fortran 77 compilers in the AC_CHECK_TOOLS list. Of course, g77 should come before "fc" and any Fortran 90 compilers. -- Martin W

Re: F77_LIBRARY_FLAGS patch revised

way because I thought you had ask me to resubmit. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

top_srcdir: Bug in cvs version -?

ALthough my test package is flat, since today (07-07-2000) ./configure inserts top_srcdir=.. in the Makefile, causing errors: make: Fatal error: Don't know how to make target `../configure.in' (top_srcdir should be ".") Martin

Re: [patch] acgeneral.m4 (dirname usage)

> Hm, I've used this many times, and have never been trapped. Which sed > are you thinking about? Sorry, I was wrong. Problem was: I had a space between the "!" and the sed command. :( *That* is only supported by GNU sed.

Re: [patch] acgeneral.m4 (dirname usage)

on is not portable - actually, it seems to work only with GNU sed. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: autoconf fortran 90 support

ary to use the same compiler for f77 and f90 source. Actually, autoconf/automake must detect this case, and take special measures to handle it. This would probably imply a lot of further complications with make and automake rules. -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: AC_F77_NAME_MANGLING question

ver, I'm waiting what the maintainers say. Best regards, Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: AC_F77_LIBRARY_LDFLAGS patch (part 1)

Dear Akim, > | + * aclang.m4 (AC_F77_LIBRARY_LDFLAGS): > | + don't swallow -lm (if a Fortran compiler needs it, it is likely > | + to be required when linking C/C++ with Fortran) > | + > | + -YP,... option: replace ^-YP, with -L > > This is not what you did: you refer to -YP without the

F77_NAME_MANGLING/F77_WRAPPERS rewrite attempt

me,NAME)], [NAME ## _]) AC_DEFINE([F77_FUNC_(name,NAME)], [NAME ## _]) ;; "upper case, underscore, extra underscore") AC_DEFINE([F77_FUNC(name,NAME)], [NAME ## _]) AC_DEFINE([F77_FUNC_(name,NAME)], [NAME ## __]) ;; *) AC_MSG_WARN(unknown Fortran 77 name-mangling scheme) ;; esac ])# AC_F77_WRAPPERS -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

Re: AC_F77_NAME_MANGLING question

dn't produce errors on any system I've tried it on, so it may be ok for practical purposes. Is the orignal author present on this list? Regards, Martin -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

AC_F77_LIBRARY_LDFLAGS

ld rather be "-L/Some/directory" ? No error occurs if the compiler doesn't set LD_RUN_PATH. I don't quite understand the reason why LD_RUN_PATH is considered in the macro, thus I couldn't actually "fix" it - I just commented it out and had no proble

AC_F77_NAME_MANGLING question

underscore to symbol names containing one already, f77_underscore will have the wrong value "double" and AC_F77_WRAPPERS will produce incorrect code. Please correct me if I'm wrong! -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04

Re: autoconf fortran 90 support

ying/linking - other things I haven't thought of yet ? -- Martin Wilck <[EMAIL PROTECTED]> Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany Tel. +49-341-2352151 / Fax +49-341-2352361

autoconf fortran 90 support

ran code. Although these certainly need to be cleaned up and adapted to the new code structure, they might be useful to include in autoconf sooner or later. Again, I'd like to know whether there is interest for this in the autoconf community. Best regards Martin -- Martin Wilck <[EMAIL PROTE