[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-core-devel
Subject: Re: SUBMISSION: New Library for kdesupport: indexlib
From: "Malte S. Stretz" <msquadrat.nospamplease () gmx ! net>
Date: 2005-06-12 20:26:41
Message-ID: 200506122226.46145 () malte ! stretz ! eu ! org
[Download RAW message or body]
On Sunday 12 June 2005 18:11 CET Thiago Macieira wrote:
> Malte S. Stretz wrote:
> >https://bugs.kde.org/show_bug.cgi?id=101542
> >https://bugs.gentoo.org/show_bug.cgi?id=86898
>
> These are a bit different. g++ isn't at fault there: we are.
>
> We turned on visibility for code that isn't supposed to use that. We
> should have used the pragmas in our headers and source files, instead of
> using -fvisibility=hidden.
Ah, k. I didn't read the thread very thorough, just saw that bero asked for
3.4.x with -fvisibility=hidden and what might be broken :)
> In special, the hidden visibility was applied to Qt3 code, which doesn't
> support that in Unix. That meant the virtual tables and typeinfos for
> polymorphic classes without a destructor were emitted in the program and
> in libqt-mt. The end result is that dynamic_cast would operate wrongly.
>
> The bug we're alluding to here is a real bug, affecting the g++ 3.4.x
> patched versions. See my other emails for details.
Jepp, interesting stuff though I'm probably a bit masochistic calling that
"interesting" :)
Cheers,
Malte
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic