Re: [K3d-development] gcc 3.3 - [was:] getting k3d on solaris
Brought to you by:
barche
From: Riccardo <ro...@ti...> - 2004-01-07 22:54:26
|
"Timothy M. Shead" wrote: > Well, the good news is that this isn't a Solaris-specific issue :) The > bad news is that you're using gcc 3.3, right? I didn't notice this in > your earlier post of I'd have mentioned it - this is a longstanding > issue that we have with gcc 3.3. Yes I'm using gcc 3.3.2. gcc 3.2.2 was able to produce some totally broken code on sparc and sometimes segfaulted itself (!!) during compile. For now I compiled many things successfully with it and this is the first project that fails. This is realy bad news. Seems that my latest porting effort are a bit doomed lately between k3d and aqsis! damn! > I've yet to find any useful information on the cause of this problem - I > am convinced that the code is correct C++ (works fine under gcc 3.0, > 3.1, 3.2, probably still works under gcc 2.95), and > __class_type_info_pseudo seems to be some internal implementation detail > for either the compiler or the runtime libraries. this is good, I have gcc 2.95 almost everywhere. But the only computer with mesa3d, gtk and enough pwoer to be useful for interactive modeling is this sun box. A headless box would not be useful :) > To ensure that everyone knows where they stand, I'm going to do some > arbitrary policy-making here: until Gentoo marks gcc 3.3 stable on x86, > I'm not gonna touch it. If you're using gcc 3.3 now and don't want to > wait, the most useful thing you could do would be to file a bug report > with the gcc team, and attach the preprocessed code that fails. bug GCC, pleasant activity. how can I preprocess the incriminated file? -Ric |