Hi Iain,
>> On 25 Jun 2026, at 20:47, Rainer Orth <[email protected]> wrote:
>>>> On 25 Jun 2026, at 20:35, Iain Sandoe <[email protected]> wrote:
>>>>> On 25 Jun 2026, at 20:30, Rainer Orth <[email protected]>
>>>>> wrote:
>>>>
>>>>>> Current Darwin no longer supports non-pie executables, leading to
>>>>>> spurious
>>>>>> fails because of linker warnings; skip these cases.
>>>>>
>>>>> I see that quite a number of tests are already skipped on Darwin because
>>>>> of this issue. ISTM that it would be better to introduce a no_pie
>>>>> effective target rather than using such ad-hoc skips...
>>>>
>>>> seems fair, however, I’m not a big fan of double-negatives - and we’d end
>>>> up
>>>> with lots of { target { ! no_pie } }
>>>>
>>>> maybe “pie_exe_OK”?
>>>> (open to better spellings)
>>>
>>> hmm .. I blame the heat - non_pie is perhaps the right thing in this case
>>> .. because
>>> it’s gating the no-pie option; I suppose we should do a survey of which
>>> permutations
>>> are most often used.
>>
>> to avoid the double negative, we could use pde (position-dependent
>> executable). I've recently seen that use in binutils, but don't know
>> how widespread the term is.
>
> That’s a new one to me - although that’s not much of a survey :)
The binutils PR was the first instance I saw this, but it stuck somehow.
> “pie_exe_OK” and “non_pie_exe_OK” are kinda in line with the way we have
> spelt things in the testsuite to date .. even if a bit longer than we might
> like.
_exe_ is redundant with position-independent *executable*. I'm not
particularly fond of the _ok forms, though, and no uppercase, please.
> just throwing stuff out there .. not strongly attached to any solution so far
Me neither: we're getting into bikeshedding territory here.
Rainer
--
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University
