[PATCH] Make the driver accept all four variants of the target option
richard at metafoo.co.uk
Fri Feb 20 15:57:15 PST 2015
On Fri, Feb 20, 2015 at 12:37 AM, James Molloy <james at jamesmolloy.co.uk>
> Hi Richard,
> My belief, from having looked at this code quite some time ago, is that we
> deliberately have mechanisms for handling --option=foo and --option foo
> differently, as in some circumstances GCC has different semantics for each.
> It just seemed to me that about a million aliases were missing for the
> cases where semantics didn't differ.
*sigh* Why am I not surprised... =)
Nonetheless, I think those cases should be the unfortunate exception rather
than the rule.
> On Fri Feb 20 2015 at 1:41:04 AM Richard Smith <richard at metafoo.co.uk>
>> By coincidence, I happened to run into the fact that we don't support
>> --target earlier today. We have a lot of truly weird baggage in our
>> command-line syntax, but I think we should be striving to minimize it. Is
>> there a justification for supporting one of "--foo bar" and "--foo=bar" but
>> not the other, for *any* of our options with arguments?
>> (I find it especially weird that our TableGen option mechanism has native
>> support for handling "-Ifoo" and "-I foo" as the same option, but not for
>> the more common case of "--blah foo" and "--blah=foo".)
>> On Thu, Feb 19, 2015 at 9:34 AM, Eric Christopher <echristo at gmail.com>
>>> Let's go with it. I don't see a reason not to. The only (somewhat silly)
>>> objection I was thinking was that --target <triple> feels like the
>>> configure option. But that's not necessarily a bad thing and I like it more
>>> than --target= anyhow :)
>>> rL LLVM
>>> EMAIL PREFERENCES
>>> cfe-commits mailing list
>>> cfe-commits at cs.uiuc.edu
>> cfe-commits mailing list
>> cfe-commits at cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cfe-commits