Hello Bruno,
Apologies for the delay.
* Bruno Haible wrote on Mon, Oct 15, 2007 at 10:51:36PM CEST:
>
> > | The contents of this variable are passed to every compilation that
> > | invokes the C preprocessor; it is a list of arguments to the
> > | preprocessor. For instance, `-I'
Bruno Haible <[EMAIL PROTECTED]> writes:
> Applying the appended patch and proposing this additional patch.
> Paul, Jim, what do you think of this?
Thanks for doing that. It looks good to me.
Bruno Haible <[EMAIL PROTECTED]> wrote:
> Colin Watson wrote:
>> I was trying to prepare a test case for you, but I find that now I
>> cannot reproduce the problem, either with VPATH or non-VPATH builds! It
>> seems there must have been some junk lying around in my tree after all,
> ...
>>
>> ht
Ralf Wildenhues wrote:
> > To me, it looks like AM_CPPFLAGS is empty by default. Automake's default
> > includes are in a separate variable called DEFAULT_INCLUDES.
>
> Only if you've used nostdinc (`info Automake "Program variables"'):
>
> | `AM_CPPFLAGS'
Then the doc of AM_CPPFLAGS is misunde
Colin Watson wrote:
> I was trying to prepare a test case for you, but I find that now I
> cannot reproduce the problem, either with VPATH or non-VPATH builds! It
> seems there must have been some junk lying around in my tree after all,
...
>
> http://people.debian.org/~cjwatson/tmp/man-db-gnuli
On Sun, Oct 14, 2007 at 11:07:28PM +0200, Bruno Haible wrote:
> Colin Watson wrote:
> > I made this change and it now works correctly.
>
> Thanks for checking. That's important to know.
>
> How exactly did you invoke gnulib-tool? With or without a --makefile-name
> option?
Without. Also, I'm ins
Hello Bruno, Colin,
* Bruno Haible wrote on Sun, Oct 14, 2007 at 11:07:28PM CEST:
> Colin Watson wrote:
>
> > Gnulib needs to put -I. in its AM_CPPFLAGS, since it
> > appears to define that variable to empty
>
> Yes, I agree that this may be the mistake.
>
> > thus cancelling Automake's default
Colin Watson wrote:
> I made this change and it now works correctly.
Thanks for checking. That's important to know.
How exactly did you invoke gnulib-tool? With or without a --makefile-name
option?
> Gnulib needs to put -I. in its AM_CPPFLAGS, since it
> appears to define that variable to empty
On Sun, Oct 14, 2007 at 07:37:35PM +0100, Colin Watson wrote:
> On Sun, Oct 14, 2007 at 05:41:06PM +0200, Bruno Haible wrote:
> > The expected behaviour is that __getwcd expands to getcwd which in turns
> > expands to rpl_getcwd. If getcwd does not expand to rpl_getcwd, you need
> > to find out why
Colin Watson wrote:
> REPLACE_GETCWD is 1 and unistd.h is up to date.
This is all as it should be.
> However, getcwd.c includes . Should this not be "unistd.h",
> in order to use the version from Gnulib in the current directory?
Good idea? Maybe your AM_CPPFLAGS are missing -I. -I$(srcdir) ?
On Sun, Oct 14, 2007 at 05:41:06PM +0200, Bruno Haible wrote:
> Colin Watson wrote:
> > I'm in the process of making my man-db package use Gnulib. Upon
> > importing the xgetcwd module and using it
>
> Have you done a "make distclean" and reconfigured after doing this?
Yes.
> > Isn't this bound
Hi,
Colin Watson wrote:
> I'm in the process of making my man-db package use Gnulib. Upon
> importing the xgetcwd module and using it
Have you done a "make distclean" and reconfigured after doing this?
> I found that calling xgetcwd caused infinite recursion:
>
> #34894 0x0805cc1e in getcwd (
12 matches
Mail list logo
