public inbox for java@gcc.gnu.org
 help / color / mirror / Atom feed
From: Andrew Haley <aph@redhat.com>
To: Herman ten Brugge <hermantenbrugge@home.nl>
Cc: java@gcc.gnu.org
Subject: Re: Making -findirect-dispatch the default?
Date: Sat, 19 Dec 2009 17:27:00 -0000	[thread overview]
Message-ID: <4B2D0CFB.5080404@redhat.com> (raw)
In-Reply-To: <4B2CBCB2.3050401@home.nl>

Herman ten Brugge wrote:
> I would like to make a suggestion to make the "-findirect-dispatch" the 
> default when compiling code with gcj.
> 
> If you do not do this you some times get the message:
> 
> "Incorrect library ABI version detected.  Aborting."
> 
> This is really puzzling the first time you see it. You compiled you 
> source code from scratch and
> still this message is given when running it.

No, you only get this if you install binaries that are not compatible
with the installed libraries.  If you build with the same version you'll
be fine.

> I found the solution on the internet and used "-findirect-dispatch" to 
> solve it.
> Why is this still not the default. I see that is was first introduced in the
> "ftp://gcc.gnu.org/pub/gcc/summit/2004/GCJ New ABI.pdf"
> because of serious ABI problems. But this was 5 years ago.

Well, these were not, exactly, problems, although they could be.
gcj defaults to the same ABI that C++ uses.

> I also see that it has some performance penalties. But for optimizing 
> code we
> always have to use some optimizing flags. We can then supply 
> "-fno-indirect-dispatch"
> is we want faster code (and ABI incompatible).

There are not many packages that depend on gcj's "old" ABI, the default.
But there are some, and these include all of the packages that use CNI.
If we change the default, we will break their build scripts.  This is a
Bad Thing.  If indirect dispatch worked with CNI, this would be much
easier, but it doesn't.

gcc is in stage 3: bug fixes only.  While there is considerable merit
in making indirect dispatch the dafault, this isn't the time to make
such an ABI-breaking change.

Andrew.

      reply	other threads:[~2009-12-19 17:27 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-12-19 11:44 Herman ten Brugge
2009-12-19 17:27 ` Andrew Haley [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4B2D0CFB.5080404@redhat.com \
    --to=aph@redhat.com \
    --cc=hermantenbrugge@home.nl \
    --cc=java@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).