bug-autoconf
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: sun cc -pedantic


From: Ralf Wildenhues
Subject: Re: sun cc -pedantic
Date: Thu, 10 Jan 2008 09:00:15 +0100
User-agent: Mutt/1.5.13 (2006-08-11)

Hello Peter,

* Peter Bunclark wrote on Wed, Jan 09, 2008 at 01:37:11PM CET:
>       I had 3rd-part software configure fail on Solaris with
> "sizeof(char)".  It turns out that they include -pedantic; the -pedantic
> check tests that cc -pedantic returns without error. The sun compiler can
> legally use cc -p -e dantic which returns without error causes no
> executable to be formed.
> 
> Later, the sizeof(char) test tries to compile with cc -pedantic and fails.

Autoconf itself contains no test for -pedantic.  So that is probably
part of the 3rd-party software you're building.  Please report to them,
so they can fix their macro that tests for this flag.

To make life easier for them, please be more specific in what happens
and what fails.  (FWIW, I only half understood your problem after trying
out Sun cc myself.)  Namely: Sun cc understands `-pedantic' as
  -p -e dantic

where '-e dantic' is passed on to the linker (meaning the symbol
'dantic' will be the entry point address for the output file), and -p
has to do with profiling.

What I still don't understand however, is this:

>                            try_link="`$CC -pedantic -o conftest$ac_exeext \
>                                     conftest.$ac_ext 2>&1`"

If this is the test, then the output for Sun cc should most likely be
nonempty.  For me, it's something like

| ld: fatal: entry point symbol `dantic' is undefined

So I don't understand why the next line should be necessary for you:

>         try_link=1
>                            if test -z "$try_link"; then

Hope that helps.

Cheers,
Ralf




reply via email to

[Prev in Thread] Current Thread [Next in Thread]