Thank you for your feedback!
For options which have `NoDriverOption` Flags (such as -mrelocation-model),
I agree that we should expose these flags to users only when `-cc1` is also
passed.
However, as to `-mrelocation-model [tab]`, I think it is fine to provide
possible values for this option (st
On 27/06/17 07:17, Saleem Abdulrasool via cfe-commits wrote:
I think that we shouldn't be providing completion for `-cc1` options.
`-cc1as` options are fine as the IAS serves as a replacement for the
traditional unix `as`. But, the `NoDriverOption` values shouldn't be
exposed to users. They
On Mon, Jun 26, 2017 at 10:17 PM, Saleem Abdulrasool via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> I think that we shouldn't be providing completion for `-cc1` options.
> `-cc1as` options are fine as the IAS serves as a replacement for the
> traditional unix `as`. But, the `NoDriverOpti
I think that we shouldn't be providing completion for `-cc1` options.
`-cc1as` options are fine as the IAS serves as a replacement for the
traditional unix `as`. But, the `NoDriverOption` values shouldn't be
exposed to users. They are internal details, with no compatibility. If
users start usin
Author: yamaguchi
Date: Fri Jun 23 12:05:50 2017
New Revision: 306127
URL: http://llvm.org/viewvc/llvm-project?rev=306127&view=rev
Log:
[GSoC] Add support for CC1 options.
Summary:
Add value completion support for options which are defined in
CC1Options.td, because we only handled options in Opti
