[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: tack: --enable-warnings configure flag doesn't actually work
From: |
Thomas Dickey |
Subject: |
Re: tack: --enable-warnings configure flag doesn't actually work |
Date: |
Tue, 06 Mar 2012 04:56:54 -0500 |
User-agent: |
Mutt/1.5.18 (2008-05-17) |
On Mon, Mar 05, 2012 at 10:16:53PM -0500, Samuel Bronson wrote:
> On Sat, Mar 3, 2012 at 7:35 PM, Thomas Dickey <address@hidden> wrote:
> > On Fri, Mar 02, 2012 at 05:38:50AM -0500, Thomas Dickey wrote:
> >> On Thu, Mar 01, 2012 at 10:23:43PM -0500, Samuel Bronson wrote:
> >> > I feel a little silly not to have noticed this sooner, but it occurred
> >> > to me today as I was working on packaging the latest snapshot that the
> >> > warning flags that tack's configure script detects when given the
> >> > --enable-warnings flag don't actually get passed to the compiler.
> >> >
> >> > Searching through config.status reveals that it is prepared to
> >> > substitute them in wherever @EXTRA_CFLAGS@ appears in Makefile.in, but
> >> > Makefile.in doesn't actually contain that anywhere.
> >>
> >> I see (overlooked that...)
> >>
> >> > And if you do add it in, you get a whole pile of warnings about
> >> > undefined preprocessor macros...
> >>
> >> I can fix that (have just made a first draft of that, will check the
> >> various compilers that care about this).
> >
> > I uploaded a new snapshot, which fixes the various compiler warnings,
> > and was built on several platforms (should be an improvement...).
>
> Okay, I've tried that, and I'm kind of puzzled to report that it
> doesn't seem to work correctly with bash (invoked either as bash or as
> sh), failing like this:
>
> % bash ./configure
> [...]
> configure: creating ./config.status
> config.status: creating Makefile
> ./config.status: line 697: unexpected EOF while looking for matching ``'
> ./config.status: line 714: syntax error: unexpected end of file
>
> but (and this is the *really* strange part) it runs to completion with dash!
Perhaps that's in the last set of changes I was working on - which I
built (after the fix for Solaris) with Ubuntu 10 and openSUSE 11.3.
Earlier, I recall an issue where it turned out that bash's nonstandard
behavior had hidden a problem in the substitution at the end.
Will see how to reproduce this, and provide a fix...
thanks
--
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net
signature.asc
Description: Digital signature