January 02, 2017
Waiting for 1.1.0 final very very much ;)
January 04, 2017
kinke <noone@nowhere.com> writes:

> Hey all,
>
> on behalf of the LDC team I am proud to announce the new 1.1.0-beta6
> release!
> It's based on the 2.071.2 frontend and standard library and supports
> LLVM 3.5 up to current trunk (4.0).

I spent some time catching up on D and tried this 1.1.0-beta6 version on Rasperry Pi 3 with LLVM 3.9.1.

The dmd-testsuite is getting Bus Errors in 4 tests.  I haven't dug in to investigate yet.  Anybody know the last tagged version that passed dmd-testsuite on ARMv7?
-- 
Dan
January 04, 2017
Hey Dan, good to see you back.

On Wednesday, 4 January 2017 at 14:40:18 UTC, Dan Olson wrote:
> kinke <noone@nowhere.com> writes:
>
>> Hey all,
>>
>> on behalf of the LDC team I am proud to announce the new 1.1.0-beta6
>> release!
>> It's based on the 2.071.2 frontend and standard library and supports
>> LLVM 3.5 up to current trunk (4.0).
>
> I spent some time catching up on D and tried this 1.1.0-beta6 version on Rasperry Pi 3 with LLVM 3.9.1.
>
> The dmd-testsuite is getting Bus Errors in 4 tests.  I haven't dug in to investigate yet.  Anybody know the last tagged version that passed dmd-testsuite on ARMv7?


The dmd tests ran fine for me natively on Android/ARM as of ed22ff1 from master, dunno about tagged versions.  Could it be specific to armhf?

January 05, 2017
Joakim <dlang@joakim.fea.st> writes:
> Hey Dan, good to see you back.

Thanks!

> On Wednesday, 4 January 2017 at 14:40:18 UTC, Dan Olson wrote:
>
> The dmd tests ran fine for me natively on Android/ARM as of ed22ff1 from master, dunno about tagged versions.  Could it be specific to armhf?

I tried making ed22ff1 and make errors out when linking ldc.  Some duplicate druntime symbols.  So going backwards...

Built 1.1.0-beta3 and that is golden with dmd-testsuite except for lit tests which I don't think have ever been made ARM-ready.

I guess next is finding the commit that introduced Bus Errors sometime after beta3 before tackling the linker problem.  gdb shows all cases happen in code generated for synchronized blocks, something messed up in D_CRITICAL_SECTION, probably the pthread_mutex_t.
January 06, 2017
Dan Olson <gorox@comcast.net> writes:

> Joakim <dlang@joakim.fea.st> writes:
>> Hey Dan, good to see you back.
>
> Thanks!
>
>> On Wednesday, 4 January 2017 at 14:40:18 UTC, Dan Olson wrote:
>>
>> The dmd tests ran fine for me natively on Android/ARM as of ed22ff1 from master, dunno about tagged versions.  Could it be specific to armhf?
>
> I tried making ed22ff1 and make errors out when linking ldc.  Some duplicate druntime symbols.  So going backwards...
>
> Built 1.1.0-beta3 and that is golden with dmd-testsuite except for lit tests which I don't think have ever been made ARM-ready.
>
> I guess next is finding the commit that introduced Bus Errors sometime after beta3 before tackling the linker problem.  gdb shows all cases happen in code generated for synchronized blocks, something messed up in D_CRITICAL_SECTION, probably the pthread_mutex_t.

Introduced in commit 8d06f23 "Allow align(1) for global variables". I'll have to study it, but my guess is that there was an existing alignment problem and this commit just revealed it.

Added https://github.com/ldc-developers/ldc/issues/1955
Next ›   Last »
1 2