Possible dmd 2.078 regression ?

[Basile B.]

On Friday, 26 January 2018 at 22:15:15 UTC, Basile B. wrote:
> On Friday, 26 January 2018 at 21:52:30 UTC, timotheecour wrote:
>> On Friday, 12 January 2018 at 18:51:31 UTC, Basile B. wrote:
>>> On Friday, 12 January 2018 at 18:50:10 UTC, Basile B. wrote:
>>>> On Friday, 12 January 2018 at 17:58:30 UTC, Stefan Koch 
>>>> wrote:
>>>>> On Friday, 12 January 2018 at 14:13:22 UTC, Basile B. wrote:
>>>>>> I have a simple program that only compiles if the 
>>>>>> dependency is not pre-compiled as a static library. It 
>>>>>> worked fine before.
>>>>> I guess a mangle problem ?
>>>>
>>>> Yes and quite old...apparently it's more a 2.074.x 
>>>> regression.
>>>> i'm digging right now.
>>>
>>> digger: Commit 1e7b526b40852e9b85df3684430e371034cdf7ec (1/1) 
>>> is untestable.
>>> digger: There are only untestable commits left to bisect.
>>> digger: The first bad commit could be any of:
>>> digger: 1e7b526b40852e9b85df3684430e371034cdf7ec
>>> digger: 6fecaa8232a427fb3ca29c5a5245e08fc43b71b1
>>> digger: f0410bea1ad2b130884964d603b34e729b3e4f69
>>> object.Exception at bisect.d(186): We cannot bisect more!
>>
>> please file a bug on d.puremagic.com/issues/
>
> Yeah i did.
>
> https://issues.dlang.org/show_bug.cgi?id=18234
>
> I saw a related (but not similar) thing yesterday. the guy has 
> to pass a dummy.d file for linking to succeed.

Ah it was you actually:

https://forum.dlang.org/thread/mailman.2616.1516919399.9493.digitalmars-d-bugs@puremagic.com