Re: problem with ifelse

[Vincent Torri]

On Thu, 5 Mar 2009, Eric Blake wrote:

> Most of the m4sugar macros I referred to are quite old.  The nice thing
> about git development is that it is easy to find when something was
> introduced.  For example, m4_toupper was only mentioned in two commit
> messages:

I have looked in the ChangeLog files and indeed, they were all added 
before autoconf 2.50, which is good.

> > For me, ifelse([$2], [no], [part1], [part2]) executes part1 if $2 is the
> > same than the string "no", and executes part2 otherwise. But whatever
> > the value of $2 is (in my cases, $2 is "yes" or "no", it is always part2
> > that is executed.
>
> Have you tried using the tracing utilities?

I didn't know that there were tracing utilities. I looked a bit in 
autoconf manual, and it seems that i have to call autoconf with the -t 
option, like:

autoconf -t ECORE_CHECK_MODULE

Is it the good option ?

> Could you post a snippet
> where you are actually invoking ECORE_CHECK_MODULE, to make sure there
> aren't any obvious typos in how you called it?

Here is an example of use:

want_ecore_win32="no"
ECORE_CHECK_MODULE([Win32], [${want_ecore_win32}], [yes],
  [ecore_win32_libs="-lole32 -lgdi32"])


> > and what to do with m4_pushdef if i want to transform - to _. For example, I 
> > use this:
> >
> > pushdef([UP], translit([$1], [-a-z], [_A-Z]))dnl
> > pushdef([DOWN], translit([$1], [-A-Z], [_a-z]))dnl
>
> [m4_]translit expands its result, so the safest approach is to
> double-quote the first argument in preparation for the later expansion.  I
> would write this as:
>
> m4_pushdef([UP], m4_translit([[$1]], [-a-z], [_A-Z]))dnl

ok, thank you very much

Vincent Torri