[Top][All Lists]

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

Re: test failures building 5.93 and hpux compiler patch

From: Peter O'Gorman
Subject: Re: test failures building 5.93 and hpux compiler patch
Date: Thu, 17 Nov 2005 08:03:38 -0600
User-agent: Mutt/1.5.6+20040907i

On Wed, Nov 16, 2005 at 01:44:12AM -0800, Paul Eggert wrote:
> "Peter O'Gorman" <address@hidden> writes:

The HP-UX 10.20 compiler is documented to fail on switch statements with
64 bit values, so yes, that test causes it to crash. However, the
problem only occurs with re_search which returns a regoff_t which is an
off_t which is, correctly, 64 bits. Please don't use this test in

> > The HP-UX-11.00 compiler crashed when it saw *copy_into_self |=
> > local_copy_into_self and other perfectly valid constructs.
> Perhaps all we need to do is substitute our on bool type for the
> compiler's broken one.  How was the 'bool' type defined on your
> platform?  Was lib/stdbool.h created, or the compiler's stdbool.h
> used?  What is the value of HAVE_STDBOOL_H in config.h?  HAVE__BOOL?
> Are these values correct?

Ah! That may be so, haven't tested, but look at this:
cc: "conftest.c", line 123: error 1639: Size of object being initialized
is too small to hold an address.
This line, assuming I counted right, is:
|         bool e = &s;

Hope this helps,

reply via email to

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