This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: hpux specific installation notes for GCC

  In message <>you write:
  > This is a multi-part message in MIME format.
  > --------------818D9389139A407132D69130
  > Content-Type: text/plain; charset=us-ascii
  > Content-Transfer-Encoding: 7bit
  > Gerald Pfeifer wrote:
  > > Would you mind submitting a patch against the current
  > > web page (or the version in our wwwdocs CVS module) instead of the whole
  > > file?
  > Output of
  > diff -U 3 specific.html.orig specific.html
  > attached. Is this OK?
NO.  I disagree strongly with most of the statements.

First, GCC should be bootstrappable with GCC or the HP compilers (bundled
or unbundled).

Second, I do *not* recommend using the binaries at

Third, you should be able to use hpux11 (32 bit mode) without using a
binutils snapshot.

Fourth, you do not need bison, if you ever need bison to build a release,
then there's a bug somewhere.

There shouldn't be any problems with alloca.  I've been developing on this
platform for years and it should just work.

I do not encourage an explicit path to the assembler.  Never have, never
will.  There are better ways to solve that problem.

  > --------------818D9389139A407132D69130
  > Content-Type: text/plain; charset=us-ascii;
  >  name="patch"
  > Content-Transfer-Encoding: 7bit
  > Content-Disposition: inline;
  >  filename="patch"
  > --- specific.html.orig	Wed Sep 27 08:55:18 2000
  > +++ specific.html	Wed Sep 27 09:09:14 2000
  > @@ -26,6 +26,7 @@
  >    <li><a href="#hppa*-hp-hpux*">hppa*-hp-hpux*</a></li>
  >    <li><a href="#hppa*-hp-hpux9">hppa*-hp-hpux9</a></li>
  >    <li><a href="#hppa*-hp-hpux10">hppa*-hp-hpux10</a></li>
  > +  <li><a href="#hppa*-hp-hpux11">hppa*-hp-hpux11</a></li>
  >    <li><a href="#ix86-*-linux*">i?86-*-linux*</a></li>
  >    <li><a href="#ix86-*-sco3.2v5*">i?86-*-sco3.2v5*</a></li>
  >    <li><a href="#ix86-*-solaris*">i?86-*-solaris*</a></li>
  > @@ -188,8 +189,69 @@
  >  but still has some problems.  Most notably the assembler inserts timestamp
  > s
  >  into each object file it creates, causing the 3-stage comparison test to f
  > ail
  >  during a `<code>make bootstrap</code>'.  You should be able to continue by
  > -saying `<code>make all</code>' after getting the failure from `<code>make 
  > -bootstrap</code>'.</p>
  > +saying `<code>make all</code>' after getting the failure 
  > +from `<code>make bootstrap</code>'.</p>
  > +
  > +<p>
  > +Some users reported (rare) linkage errors with gcc-2.95.2, which where
  > +replaced by other errors in recent snapshots.
  > +The latest release of binutils had bugs in the handling of weak symbols
  > +under hpux. Recent snapshots of gcc and binutils might even fail to build.
  > +</p>
  > +<p>
  > +Quintessence: gcc-users under hpux-10 should use gcc-2.95.2 until they rea
  > lly
  > +have problems with it. Precompiled packages of gnu software may be found a
  > t <a
  > +href="">the Software Porting And Archive Centre fo
  > r
  > +HP-UX</a>.  Everybody is welcome to test newer snapshots and report errors
  > +to the appropriate mailing lists, but should be aware of the fact that the
  > re
  > +might still be things yet to be fixed.
  > +</p>
  > +
  > +
  > +<hr>
  > +<h3><a name="hppa*-hp-hpux11">hppa*-hp-hpux11</a></h3> <p>Regarding hpux-1
  > 1,
  > +support for it is new in binutils and gcc.  There is no official release o
  > f gcc
  > +that supports it. If You wish to experiment with a recent snapshot of gcc,
  >  the
  > +use of a recent snapshot of binutils is obligatory. The latest release of
  > +binutils had bugs in the handling of weak sysmbols under hpux.  See the <a
  > +href="">binutils home page</a> for more
  > +information.
  > +</p>
  > +<p>
  > +It is recommended to install <code>bison</code>, <code>gmake</code> and to
  >  use
  > +<code>cvs</code> as described at <a
  > +href=""></a>, in or
  > der to
  > +obtain the most recent versions of gcc/binutils.<br> Precompiled packages 
  > of gnu
  > +software may be found at <a href="">the Software P
  > orting
  > +And Archive Centre for HP-UX</a>.
  > +
  > +If You encounter problems in building binutils due to missing 
  > +symbol <code>alloca(code)</code> a possible workaround is to 
  > +set environment variables <code>ac_cv_header_alloca_h</code> and 
  > +<code>ac_cv_func_alloca_works</code> to <code>no</code> 
  > +and retry with <code>make distclean; ./configure ; make</code>
  > +</p>
  > +<p>
  > +When configuring gcc we encourage the explicit specification of the
  > +gnu-assembler and its absolute path. Example:<br>
  > +<code>[build-dir] # ../gcc/configure --prefix=/opt/gcc-today \<br> 
  > +--with-gnu-as \<br> --without-gnu-ld \<br> --with-as=/opt/binutils-today/b
  > in/as</code>
  > +</p>
  > +<p>
  > +Like on hpux-10, the HP assembler inserts timestamps
  > +into each object file it creates, causing the 3-stage comparison test to f
  > ail
  > +during a `<code>gmake bootstrap</code>'.  You should be able to continue b
  > y 
  > +saying `<code>gmake all</code>' after getting the failure 
  > +from `<code>gmake bootstrap</code>'.
  > +</p>
  > +<p>
  > +Situation for gcc on hpux-11 changes daily at the moment, so most of the
  > +problems may resolve in near future. Nonetheless every helpful 
  > +contribution like patches or explanations for build or runtime failures
  > +is always welcome.
  > +</p>
  > +
  > +
  >  <hr>
  > --------------818D9389139A407132D69130--

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]