2/dmd-testsuite")
+ (add-after 'unpack-submodule-sources 'patch-dmd2
+ (lambda* (#:key inputs #:allow-other-keys)
+ (substitute* "dmd2/root/port.c"
+ ((" ::isnan") " isnan")
+
2/dmd-testsuite")
+ (add-after 'unpack-submodule-sources 'patch-dmd2
+ (lambda* (#:key inputs #:allow-other-keys)
+ (substitute* "dmd2/root/port.c"
+ ((" ::isnan") " isnan")
+
7.1, is not currently compilable
without the isnan patch... So normally, they should continue providing bug
fix releases to the 0.17 branch, to make sure it actually does compile
with a C++ compiler in a contemporary setting.
Maybe you could do some lobbying for it...
Andreas
On Wed, Aug 03, 2016 at 07:14:50PM +0200, Roel Janssen wrote:
> I prefer getting this version in first too. We need to bootstrap 1.0.0
> from the 0.1.x one, so this is a welcome upgrade :-).
Yes, the D compiler was rewritten in D (from C++). We need to
bootstrap from the older C++ version.
Reall
Leo Famulari writes:
> On Wed, Aug 03, 2016 at 11:12:34AM +0200, Andreas Enge wrote:
>> Hi Danny,
>>
>> On Wed, Aug 03, 2016 at 09:51:10AM +0200, Danny Milosavljevic wrote:
>> > ldc fails because they do
>> > ::isnan
>> > which is a bad idea if
On Wed, Aug 03, 2016 at 11:12:34AM +0200, Andreas Enge wrote:
> Hi Danny,
>
> On Wed, Aug 03, 2016 at 09:51:10AM +0200, Danny Milosavljevic wrote:
> > ldc fails because they do
> > ::isnan
> > which is a bad idea if isnan is a C macro (which it usually isn&
2/dmd-testsuite")
+ (add-after 'unpack-submodule-sources 'patch-dmd2
+ (lambda* (#:key inputs #:allow-other-keys)
+ (substitute* "dmd2/root/port.c"
+ ((" ::isnan") " isnan")
+
Hi Danny,
On Wed, Aug 03, 2016 at 09:51:10AM +0200, Danny Milosavljevic wrote:
> ldc fails because they do
> ::isnan
> which is a bad idea if isnan is a C macro (which it usually isn't - but hey
> O_o). I've substituted
> isnan
> for it and this part then comp
> And to repair your remaining, not building favourite packages - for instance
> bio packages with disappeared source...
>http://hydra.gnu.org/eval/109040#tabs-still-fail
ldc fails because they do
::isnan
which is a bad idea if isnan is a C macro (which it usually isn't
On Sun, Jul 31, 2016 at 04:53:49PM +0200, Andreas Enge wrote:
> And rapicorn:
>http://hydra.gnu.org:3000/build/1345676/nixlog/1/tail-reload
Fixed in commit 42bf34298eec19b37e276e49074113f29c494aed.
Andreas
On Sun, Jul 31, 2016 at 04:42:35PM +0200, Andreas Enge wrote:
> It finished successfully. The next package with the same problem is synfig,
And rapicorn:
http://hydra.gnu.org:3000/build/1345676/nixlog/1/tail-reload
Andreas
finished successfully. The next package with the same problem is synfig,
or probably rather etl that it depends on:
http://hydra.gnu.org:3000/build/1345394
time.cpp: In member function ‘bool synfig::Time::is_valid() const’:
time.cpp:322:10: error: ‘::isnan’ has not been declared
return !::isn
On Sun, Jul 31, 2016 at 12:44:27PM +0200, Danny Milosavljevic wrote:
> > Apparently the "std::" is missing in front of "isnan"; should this not be
> > implicit?
> No. Many people in the C++ community frowns upon even doing "using namespace
>
> Apparently the "std::" is missing in front of "isnan"; should this not be
> implicit?
No. Many people in the C++ community frowns upon even doing "using namespace
std" at all, never mind implicitly.
http://stackoverflow.com/questions/1265039/using-std-name
Hello,
has anyone followed what the deal is with isnan in core-updates? A number
of C++ packages does not build any more. For instance, dealii:
solver_control.cc:96:24: error: ‘isnan’ was not declared in this scope
isnan(check_value) ||
^
[ 30%] Building CXX object
15 matches
Mail list logo
