Akim> * automake.in (&scan_aclocal_m4): Handle $relative_dir and special
Akim> variables.
Akim> Specify to the caller whether $regen_aclocal_m4.
Akim> (&handle_configure): Adjust.
Akim> Transform `REGEN-ACLOCAL-M4'.
Akim> * configure.am: Use it.
Akim> * tests/defs (me): New.
Akim>
I'm going to apply it tomorrow.
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&scan_aclocal_m4): Handle $relative_dir and special
variables.
Specify to the caller whether $regen_aclocal_m4.
(&handle_configure): Adjust.
Transform
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> Also, I wondered for a long time whether we should introduce
Akim> lang specific files. Most of the content of register_language
Akim> calls, should probably be in a file, and some adhoc stuff in
Akim> lang_*_finish/rewrite could be
> "David" == David Simpson <[EMAIL PROTECTED]> writes:
David> Sorry to spam all list recipients, but I'm trying to contact
David> someone administering the automake mailing list. Basically, I
David> need to unsubscribe but automated attempts to automake-request
David> have failed (I don't ha
Sorry to spam all list recipients, but I'm trying to contact someone
administering the automake mailing list. Basically, I need to
unsubscribe but automated attempts to automake-request have
failed (I don't have the password), and mails to automake-admin
have fallen on deaf ears. If someone wit
Tom Tromey <[EMAIL PROTECTED]> wrote:
| > "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes:
|
| > | Ok, I re-read your message. Why do we ever run `missing --run :'?
| > | That seems weird.
|
| Jim> I didn't know off hand either, but found this in
| Jim> automake/m4/missing.m4:
|
| Jim> if
> "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes:
> | Ok, I re-read your message. Why do we ever run `missing --run :'?
> | That seems weird.
Jim> I didn't know off hand either, but found this in
Jim> automake/m4/missing.m4:
Jim> if eval "$MISSING --run :"; then
What if we just change
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> * automake.in (Language): Add attributes `flags', `compile' and
Akim> `compiler'.
Akim> (&finish_languages, &handle_single_transform_list, &handle_dist)
Akim> (&add_depend2, ®ister_language): Use them and the `linker' and
Aki
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Tom> Objective C shouldn't be a `pure' language. My understanding is
Tom> that in this case you want to use the `C' linker. If not then
Tom> contact and ObjC hacker to get truer information.
Akim> So what should I actually do? I really
> "Lorenzo" == Lorenzo Bettini <[EMAIL PROTECTED]> writes:
Lorenzo> make dist is not a problem for me: I only wish I could have
Lorenzo> different "make install" rules, that's all :-)
If the programs are in different subdirs then there is nothing
preventing you from writing your own:
in
> "Adam" == Adam C Powell IV <[EMAIL PROTECTED]> writes:
Adam> Hmm, the "include" line seems to be copied verbatim into the
Adam> Makefile.in, so maybe when automake doesn't find the file, it
Adam> gives up and copies it?
Oh, I see -- any non-conforming `include' is passed through.
Your part
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
Tom> What happens if you run automake with this patch on autoconf
Tom> (where I understand that aclocal.m4 is not a generated file?).
Tom> Do you get an `aclocal.m4' target in Makefile.in?
You are right, thanks for the details!
I'll provide
Hm...
I can't find the message you sent Tom. I now perfectly understand the
problem, but I would like to delay the fix until the next patches are
applied, because one important point is reached later: depend2.am and
ext-compile.am are merged together (i.e., the use of the dependency
code is now
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> In particular Automake cannot override user variables with
Akim> Automake variables.
Tom> Ok, that's good. But if the user defines a variable
Tom> conditionally then automake sh
Tom Tromey wrote:
> > "Adam" == Adam C Powell IV <[EMAIL PROTECTED]> writes:
>
> Adam>* hard-code the variable I had hoped to get from the foreign
> Adam>makefile, which will break if it differs from platform to
> Adam>platform or machine to machine.
>
> Adam>* give up on cros
Tom Tromey <[EMAIL PROTECTED]> wrote:
| Jim> It was only a problem when $@ was a shell built-in, like : or cd.
|
| Ok, I re-read your message. Why do we ever run `missing --run :'?
| That seems weird.
I didn't know off hand either,
but found this in automake/m4/missing.m4:
# AM_MISSING_HAS_RU
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> * automake.in (Language): New package, temporarily in this file.
Akim> Use Class::Struct. (Automake): New package. (%languages): New.
Akim> (&handle_single_transform_list): Use
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> BUILT_SOURCES = header.h
Lars> # build $(BUILT_SOURCES) before entering subdirs
Lars> Makefile: $(BUILT_SOURCES)
Note this will only work with GNU make and is relatively bad besides.
Eg, try `make clean' and watch it rebuild the header
> "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes:
Jim> It was only a problem when $@ was a shell built-in, like : or cd.
Ok, I re-read your message. Why do we ever run `missing --run :'?
That seems weird.
Tom
Tom> My reading is that this patch always generates an aclocal.m4
Tom> target, regardless of whether one is warranted.
Akim> Hm, no. I think you missed the first lines of the patch, or I
Akim> might have misunderstood what you mean:
I saw that code.
To me it looks like handle_aclocal_m4 is cal
Tom Tromey wrote:
>
> > "Lorenzo" == Lorenzo Bettini <[EMAIL PROTECTED]> writes:
>
> Lorenzo> I mean that, e.g., java2html could be 1.7 and cpp2html 1.3,
> Lorenzo> though they are in the same package, and possibly do
> Lorenzo> something like make install_java2html...
>
> Version numbers a
> "tailbert" == tailbert writes:
tailbert> probably should not emit a LINK unless it's something
tailbert> automake knows absolutely nothing about. in the case of
tailbert> overriding .m.o, perhaps we should check extension_map
tailbert> first, and then *keep* the test for keys of suffix_ru
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == akim <[EMAIL PROTECTED]> writes:
Akim> I don't understand the test objc.test (which fails with one of
Akim> the patches I work on).
Akim> I don't understand why $(LINK) should be (and was) defined.
Tom> Objective C shouldn't
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> -my $c_suffix = $yacc_suffix;
Akim> -$c_suffix =~ tr/y/c/;
Akim> -push (@suffixes, $yacc_suffix, $c_suffix);
Tom> Deleting this push seems wrong.
Tom> How else does this stuff end up on suffixes?
Akim> It is obtained vi
> "Lorenzo" == Lorenzo Bettini <[EMAIL PROTECTED]> writes:
Lorenzo> I mean that, e.g., java2html could be 1.7 and cpp2html 1.3,
Lorenzo> though they are in the same package, and possibly do
Lorenzo> something like make install_java2html...
Version numbers are up to you.
Automake wants to kn
Tom> Incidentally changes of this sort are ripe for test cases which
Tom> test expected behavior.
Akim> You're hitting the limits of my English :) I don't know if you say
Akim> ``follow what the existing tests say'' or ``this is typically
Akim> something new tests should check''.
I mean that if
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> In particular Automake cannot override user variables with
Akim> Automake variables.
Ok, that's good. But if the user defines a variable conditionally
then automake should still be able to define it when the condition
does not hold
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> Isn't this beautiful? A *.am file only change! * subdirs.am:
Akim> Don't define info related recursive targets. * texinfos.am: Do.
Tom> ... and here's why you made the change
> "Pavel" == Pavel Roskin <[EMAIL PROTECTED]> writes:
Pavel> Hello! With the today's CVS automake the testsuite fails in
Pavel> subobj2.test on my RedHat Linux 7.
Hi Pavel.
Yes, I know, I have explained why several times, and I will fix this
soon. It is unlikely to happen in real conditio
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> The aim of this patch and the next is to clarify the status of
Akim> variables: who owns them, and how their defined type (+=, := or
Akim> =) can change.
Akim> I end up with a ru
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> The aim of this patch and the next is to clarify the status of
Akim> variables: who owns them, and how their defined type (+=, := or
Akim> =) can change.
Tom> Why does it matter
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> * automake.in (&handle_configure): Don't bother with optimizing
Akim> macro uses.
Tom> Ok.
Tom> I thought I ok'd this one already?
Maybe something similar? Or I think we just
Hello!
With the today's CVS automake the testsuite fails in subobj2.test on my
RedHat Linux 7.
$ make check TESTS=subobj2.test VERBOSE=1
make check-TESTS
make[1]: Entering directory `/home/proski/src/automake/tests'
=== Running test ./subobj2.test
Makefile.am:0: @AMDEP@CXXDEPMODE multiply defin
On Fri, Apr 06, 2001 at 12:17:47PM -0500, Raja R Harinath wrote:
: @PACKAGE@ is needed. It is used by Makefile.in for 'make dist'.
:
: "Lars J. Aas" <[EMAIL PROTECTED]> writes:
: > Index: Makefile.am
: > ===
: > @@ -33,6 +33,7 @@
: >
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> * automake.in (&transform): Fix an incredible bug which was
Akim> breaking the if/endif system. ($ELSE_PATTERN): Admit an
Akim> optional argument.
Tom> Ok.
Tom> Is there some ki
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
Tom> I think this patch changes the semantics.
Tom> Right now aclocal.m4 isn't necessarily generated. Ok, it almost
Tom> always is -- but it isn't actually a requirement.
Tom> My reading is that this patch always generates an aclocal.m4
Tom
| Akim> +sub condition_negate ($)
| Akim> +{
| Akim> +my ($cond) = @_;
| Akim> +
| Akim> +$cond =~ s/TRUE$/TRUEO/;
| Akim> +$cond =~ s/FALSE$/TRUE/;
| Akim> +$cond =~ s/TRUEO$/FALSE/;
| Akim> +
| Akim> +return $cond;
| Akim> +}
|
| Doesn't this rely on the user not using cond
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> So from now on, `FALSE' is used where ever a condition is
Akim> needed, and `#' where ever a Make condition is needed. Of
Akim> course `#' should not be actually output, but why
Tom Tromey wrote:
>
> > "Lorenzo" == Lorenzo Bettini <[EMAIL PROTECTED]> writes:
>
> Lorenzo> I'd like to include two of my programs java2html and cpp2html
> Lorenzo> into the same package (say src2...), but I would like to
> Lorenzo> maintain two different program names and versions, and
>
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
Akim> + &am_line_error ($varname, + "invalid variable `$varname': dist
Akim> is forbidden");
Tom> How about: "`dist' is forbidden"? Ordinarily I put `' around
Tom> words like that in messages.
OK, will do!
> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> * automake.in (&generate_makefile) Use variable_define to define
Akim> SOURCES and OBJECTS. (&variable_defined): Don't check for
Akim> $conditional{VAR}{COND} as this would make
41 matches
Mail list logo
