|Re: [eigen] ICE with gcc-4.4.3|
[ Thread Index |
| More lists.tuxfamily.org/eigen Archives
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] ICE with gcc-4.4.3
- From: Benoit Jacob <jacob.benoit.1@xxxxxxxxx>
- Date: Wed, 10 Feb 2010 14:37:40 -0500
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:content-type :content-transfer-encoding; bh=fVXx8947WMrpCUII1uXv4Dw4hpRvcvxTYG7cVupoc9c=; b=YEEw8A1z0r0imnnfEQ+bJsj8i/xpBW1EvgFJzBfKCDi5AA4466tuRA7AGKBz5AgQJX 7DiDv5KkYzV4PxmIZ3aq6aVYkHaPCO2BD+IpOv/IH3zLwGbxaaG0rwA7lHX1YyPfHE53 Tb1mnuL27xEF5Rj0t7aZH8U8fVR3Knyd7AeNM=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type:content-transfer-encoding; b=LkSOuKC7y69QYDEcTTui+rL1WMmSfZ9cFgTZq78YqrdvU3OEM/gqsKnl2/m13shAnm ocgePqCR5OqUp6OlaX6R3x5Zg/DFL9RPMvcgogXhVsrloBt/BZQuAHLRyrCa6a4Fbtcw CsynBhC7VLpc47uwOUCo9GhPsSPhmKFURjy/k=
Great! Indeed, an ICE is by definition a bug in GCC.
2010/2/10 Thomas Capricelli <orzel@xxxxxxxxxxxxxxx>:
> Ok, thanks for your feedback. Here all tests compile without any problem. The piece of code that triggers the ice is related to NonLinearOptimization... and i'm afraid i've done something wrong, but i cant see anything obvious. I've narrowed the pb to this, which triggers the ice for me.
> gcc segfaults with -O2 and -O3, but not with -O1:
> g++ -c -O2 -I../../eigen2-tip -o ice.o ice.cpp
> So my conclusion is that it's more a gcc bug than eigen.. and i've reported the bug there (benoit: yes, with *.ii stuff)
> Thomas Capricelli <orzel@xxxxxxxxxxxxxxx>
> In data martedì 09 febbraio 2010 13:10:35, Benoit Jacob ha scritto:
> : > Or, if it's not easy to isolate a short test case, can you generate
>> the preprocessed sources. For that, rebuild your project with
>> VERBOSE=1 make ...
>> so you see the actual g++ command lines. Then copy and paste the
>> relevant command line and add -save-temps. This will generate a .ii
>> file. You can then reproduce by just doing:
>> g++ filename.ii
>> So can you compress this file and attach filename.ii.bz2 to an email
>> to this list...
>> 2010/2/9 Helmut Jarausch <jarausch@xxxxxxxxxxxxxxxxxxx>:
>> > On 9 Feb, Thomas Capricelli wrote:
>> >> I've updated gcc last night, and now when compiling with eigen trunk, i got an internal compiler error (aka, gcc segfaults) on
>> >> ../../eigen/Eigen/src/Core/DenseStorageBase.h:516:
>> >> It's reproduceable in the sense that several of my *cpp triggers this, through different include 'paths'....
>> >> Do you have the same problem?
>> > Could you please post a short example? (I have gcc-4.4.3 running on my
>> > Gentoo system)
>> > Helmut.
>> > --
>> > Helmut Jarausch
>> > Lehrstuhl fuer Numerische Mathematik
>> > RWTH - Aachen University
>> > D 52056 Aachen, Germany