I am attempting to build a project with C and F90 sources. I have found a
way around autoconf and automake attemting to use f77 to compile the Fortran
sources, by placing "F77=f90" in Makefile.am. However, I haven't been able
to work around automake and its choice of link program. The program's
There is still one test which fails, but it's because I don't know
what happens: subobj2.test.
=== Running test ./subobj2.test
Makefile.am:0: @AMDEP@CXXDEPMODE multiply defined in condition TRUE
@AMDEP@CXXDEPMODE (Automake, where = 0) =
{
TRUE => @CXXDEPMODE@
}
FAIL: subobj2.test
I y
I had to write this one under the pressure of variable_define which
doesn't want one variable to be defined several times. Given that it
was due to a hack, fixing it using the regular mechanism seems the
right choice. In addition the code is much smaller.
Index: ChangeLog
from Akim Demaille <
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* data.am, header.am, java.am, libs.am, lisp.am, ltlib.am,
* progs.am, python.am: Use if/endif instead of ?INSTALL?.
* java.am (_am_installdirs): Be sure to set it.
Index: data.am
--- data.am Mon, 05 Mar 2001 22:33
There was some confusion between `#' and `FALSE' representing negated
conditions.
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&make_condition): Return '#' when FALSE.
(&conditional_string): Return FALSE, not '#' in the corresponding
cases.
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* distdir.am: Use and abuse of if/endif.
* scripts.am: Using if/endif with variables is fine.
Index: distdir.am
--- distdir.am Sat, 10 Mar 2001 16:24:49 +0100 akim (am/h/17_distdir.am 1.15 644)
+++ distdir.am Sun, 11 Mar 2
This patch adjusts the last details so that we can use if/endif in
file_contents just like we do in read_am_file.
In particular the previous splitter considered as a single paragraph
something like
if TRUE
do something
which prevents from handing the if. It now works properly, and it's
I must say I don't understand very well why this patch is needed, but
it is. I found it back in the old code, which was doing it for
unconditionally defined sources, but not for conditionally defined
sources (which was a bug).
Now it works for both.
Index: ChangeLog
from Akim Demaille <[EMAIL
Index: 0.175/am.prj
--- 0.175/am.prj
+++ 0.175(w)/am.prj
@@ -8,7 +8,10 @@
* automake.in (&handle_source_transform): Better locality of the
variables.
Use `next' to skip nonexistent variables.")
-(New-Version-Log "")
+(New-Version-Log "2001-03-11 Akim Demaille <[EMAIL PROT
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&handle_source_transform): Better locality of the
variables.
Use `next' to skip nonexistent variables.
Index: automake.in
--- automake.in Sun, 11 Mar 2001 12:53:20 +0100 akim (am/f/39
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&variable_dump): If a variable is undefined, say
it.
(&check_ambiguous_conditional): Give finer error messages.
(&variable_define): Check that a `+=' variable is not set with `='.
Check
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&variable_defined): Some callers, e.g.,
&am_primary_prefixes, really want to know if the variable is
defined for any condition, not `TRUE' by default.
(&am_primary_prefixes): Give a more precis
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* tests/defs, tests/depend3.test: Remove useless code.
Index: tests/defs
--- tests/defs Wed, 07 Mar 2001 21:02:27 +0100 akim (am/f/2_defs 1.6 644)
+++ tests/defs Sat, 10 Mar 2001 20:45:43 +0100 akim (am/f/2_defs 1.6 644)
@@ -35,8
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&variable_conditions_reduce): FALSE is absorbent.
(&variable_conditions): Don't include `FALSE' in the result.
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
(&variable_conditions): Don't
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&handle_source_transform): Use
&variable_conditions, don't read $conditional{$var}.
(&variable_conditions_sub): When the call is the top level call,
generate all the permutations of the condi
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* depend2.am (@AMDEP@%FPFX%DEPMODE): Define it when GENERIC so
that it does not get noticed twice by automake, which now
complains for multiple definitions.
* tests/vartar.test: s/INSTALL/install/g, automake
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&read_am_file, &file_contents): Avoid name clashed
on $cond.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 18:48:14 +0100 akim (am/f/39_automake.i 1.153 755)
+++ automake.in Sat, 10 Mar 2001 18:58:10 +
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&generate_makefile) Use variable_define to define
SOURCES and OBJECTS.
(&variable_defined): Don't check for $conditional{VAR}{COND} as
this would make perl create $condition{VAR}, which we don'
Index: 0.162/ChangeLog
--- 0.162/ChangeLog Sat, 10 Mar 2001 17:45:55 +0100 akim (am/f/47_ChangeLog 1.146 644)
+++ 0.163(w)/ChangeLog Sat, 10 Mar 2001 18:20:38 +0100 akim (am/f/47_ChangeLog 1.147
+644)
@@ -1,5 +1,15 @@
2001-03-10 Akim Demaille <[EMAIL PROTECTED]>
+ * automake.in (&variab
This patch and a couple of following patches are breaking Automake.
That was the easiest path to finish cleaning up the variable handling.
This series of patches ends with patches that fix some of these issues.
It took me hours to do that, I'd like not to be asked to rewrite
everything so that Au
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&handle_source_transform, &read_main_am_file):
Simplify loops which were split in two parts, conditional vs
unconditional variables.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 17:12:12 +0100
This patch is long, and has been quite difficult to write. I really
could not make it smaller, I tried several times to split it, but
systematically failed.
Actually, more is needed, and cleanups are to come. But precisely I
meant to keep it not even bigger.
Index: ChangeLog
from Akim Demail
I had to simplify, clean up and harmonize notations :(
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&variable_conditionally_defined): Rename as...
(&variable_unconditionally_defined): this.
(&variable_dump): Fix the output.
(&variable_d
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&handle_footer, &handle_installdirs)
(&read_main_am_file): Don't read %contents directly, use
&variable_value.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 13:36:31 +0100 akim
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&read_am_file): Define the variables in a single
shot.
Factor the $saw_bk code.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 13:19:51 +0100 akim (am/f/39_automake.i 1.144 755)
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&conditional_dump): Rename as...
(&variables_dump): this. Use...
(&variable_dump): this new sub.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 11:44:56 +0100 akim (am/f/39_auto
I used variables_dump to check that there are no difference (at the
end of a run, the variables are in the same state before and after
this patch), and there are none.
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&variable_define): Don't lose the location if we
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&define_variable): Use &variable_define.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 12:17:17 +0100 akim (am/f/39_automake.i 1.138 755)
+++ automake.in Sat, 10 Mar 2001 12:25:06 +0100 akim (a
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&read_am_file): Keep $cond_string up to date.
Index: automake.in
--- automake.in Sat, 10 Mar 2001 12:58:42 +0100 akim (am/f/39_automake.i 1.143 755)
+++ automake.in Sat, 10 Mar 2001 13:19:15 +0100 ak
29 matches
Mail list logo
