Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Jun 2010 11:39:48 +0000
From:      "b. f." <bf1783@googlemail.com>
To:        Doug Barton <dougb@freebsd.org>
Cc:        freebsd-ports@freebsd.org
Subject:   Re: mail/thunderbird3 does not build with gcc 4.5.1
Message-ID:  <AANLkTimGZBtrAC-zg3Y9YzwJn9_eP8Zik_xehQRGDlHt@mail.gmail.com>
In-Reply-To: <4C206505.2060601@FreeBSD.org>
References:  <AANLkTikhO6G4GIQee4dNZdTwN4pDR3BkKbWVBN2ReiuK@mail.gmail.com> <4C206505.2060601@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 6/22/10, Doug Barton <dougb@freebsd.org> wrote:
> On 06/21/10 23:25, b. f. wrote:
>> Doug Barton wrote:
>>> On to the next victim. :)  In my ongoing campaign to build my ports with
>>> gcc 4.5.1 thunderbird was the next to fall. Full log is at
>>> http://people.freebsd.org/~dougb/tbird.txt
>>
>> Before you embark on this campaign, remember that others have been
>> experimenting with building ports with later versions of gcc for
>> months or even years now, and there are suggestions on how to solve
>> some of the problems that arise in the FreeBSD forums and in the open
>> PRs.
>
> I certainly mean no disrespect to those who've already been working on
> this problem. I'm really interested in the idea of having a "ports
> compiler," and I'm trying to do what I can from more of a typical user
> perspective. If bringing more visibility to the issue helps get more
> ports fixed, that's a good thing, right? (I'm also trying to fix _my_
> ports as I go along as well.)
>
> The ultimate goal (in my mind anyway) would be for both src AND ports to
> be in better shape to be "compiler agnostic" so that newer versions of
> gcc, clang, or whatever else can be more of a drop-in replacement. I'm
> not naive enough to think that it will be easy, or even 100% possible.
> But the more things we _can_ fix the better.
>

Yes, of course.  I didn't mean to suggest that I thought you were
being disrespectful, but only to let you know that you may solve some
of these problems more quickly by skimming through the open PRs,
mailing lists archives, and forum listings first.  Also, because many
GNU/Linux distros, NetBSD, and DragonflyBSD have switched to later
versions of gcc and/or binutils than we have in our base system, other
packaging systems (especially Red Hat Fedora's Rawhide packages,
Gentoo Portage, and NetBSD pkgsrc) may have relevant patches,
especially for older software.  Also,

http://gcc.gnu.org/gcc-4.3/porting_to.html
http://gcc.gnu.org/gcc-4.4/porting_to.html
http://gcc.gnu.org/gcc-4.3/changes.html
http://gcc.gnu.org/gcc-4.4/changes.html
http://gcc.gnu.org/gcc-4.5/changes.html
http://gcc.gnu.org/gcc-4.6/changes.html

provide useful hints for some problems related to re-factoring of
compiler headers, and changes in compiler defaults (those for earlier
versions of gcc than the one you have mentioned may contain
information that is also applicable to 4.5 and 4.6).

I think it would be good if someone posted tinderbox results for
exp-runs of ports built with the latest stable branch of gcc, and with
clang, to show which ports need to be patched.  Perhaps that fellow
who has a GSoC project to make it possible to use clang with the Ports
system may be willing to do this, at least for clang.

I can tell you right now from my own experiences that the ports
infrastructure and many individual ports do not respect the necessary
compiler and toolchain-related variables.

b.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTimGZBtrAC-zg3Y9YzwJn9_eP8Zik_xehQRGDlHt>