Re: [RFA 2/5] Darwin: Handle unrelocated dyld.

[Simon Marchi <simon.marchi@polymtl.ca>]

On 2018-09-19 10:36, Tom Tromey wrote:
>>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:
> 
>>> +	 (Apparently dyld doesn't need to relocate itself on x86-64 darwin,
>>> +	 but don't assume that).
> 
> Joel> I am wondering whether the difference in what you are seeing
> Joel> might be explained by a difference in MacOS X version; if I were
> Joel> to guess, I would say that Xavier was running on Mac OS X Sierra.
> Joel> What version were you running on?
> 
> High Sierra.
> 
> Joel> Or perhaps the intent is to be extra careful meaning that while
> Joel> today the relocation is not necessary, we still handle it so that
> Joel> it continues working the day it becomes so?
> 
> I couldn't really say, I just concluded what I did based on the mention
> of the specific architecture there.
> 
> Joel> If the comment above is confusing, I would vote for removing it.
> 
> I think we need more information.
> 
> If we remove the comment because we think it is untrue, then Simon's
> original critique of the patch -- that there is no reason for a second
> solib breakpoint -- seems correct.  And in this case we should remove a
> chunk of the patch.
> 
> Based on my testing thus far, this would be fine.  But I don't know 
> what
> testing Xavier and Tristan did, or with what architectures.
> 
> Also my testing hasn't been exactly exhaustive.  I just try simple
> things because today those break.
> 
> Now, for my purposes, it would be fine to land the more minimal patch.
> That would not preclude adding this code back later.  If that seems ok,
> I'm happy to do it.  I'd like to get this working so I can make some
> progress on other patches, and anyway un-break the Mac port.
> 
> Joel> That makes me realize (again) that, for MacOS X, we should be 
> more
> Joel> proactive at specificying which version a patch we are submitting
> Joel> was tested on, and some information about which versions of MacOS 
> X
> Joel> a given patch helps. A fair amount of work that Tristan did once
> Joel> the initial port was created was to adapt it to subsequent 
> versions
> Joel> of Darwin. Nearly every new version of Darwin introduced its new
> Joel> set of changes requiring additional adaptations.
> 
> Wholly agreed.
> 
> Tom

I would vote for only checking in the code you know is necessary for 
now, otherwise it will just be more confusing in the future, trying to 
figure out what is needed and what isn't.

Simon