Hi again,
On 25/02/2016 01:12, Matthias Klose wrote:
> On 22.02.2016 11:15, Matthias Klose wrote:
>> Control: tags -1 + moreinfo
>> Control: reassign -1 g++-5
>>
>> please provide the preprocessed source (from one of these platforms).
>>
>> On 22.02.2016 10:39, Giulio Paci wrote:
>>> Package: g++
>>> Version: 4:5.3.1-8
>>> Severity: important
>>>
>>> --- Please enter the report below this line. ---
>>>
>>> Recently I updated phonetisaurus package and it failed to compile on a few
>>> architectures, where it used to compile (See
>>> https://buildd.debian.org/status/package.php?p=phonetisaurus&suite=experimental).
>>>
>>> On hppa, x32 and sparc64 I obtain the following error:
>>>
>>> "/usr/include/fst/interval-set.h: In member function 'bool
>>> fst::IntervalSet<T>::Contains(const fst::IntervalSet<T>&) const':
>>> /usr/include/fst/interval-set.h:351:54: error: expected ')' before 'it1'
>>> } else if (it2->begin < it1->begin || it2->end > it1->end) { // no C"
>>>
>>> that seems a bug in g++.
>>>
>>> According to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10200 it also
>>> seems
>>> to be a recurrent bug, apparently solved upstream.
>
> this works for me at least on x32.
You are right... Indeed the issue is with g++ 4:6-20160101-2 and g++-6
6-20160205-1.
I got confused by the fact that buildd installs both g++-5_5.3.1-8 and
g++-6_6-20160205-1 (phonetisaurus has been uploaded to experimental), and
g++-5_5.3.1-8 comes first.
And I have g++ 4:5.3.1-1 on my system (amd64).
I did not notice that during the x32 chroot setup g++ 4:6-20160101-2 was
installed.
So I confirm the bug, but it affects g++-6_6-20160205-1 and not g++-5_5.3.1-8.
Sorry for the confusion.
Bests,
Giulio
