dzbarsky wrote:

> I left some more comments, but in general this change is far bigger than it 
> needs to be because it's trying to match the behavior of `cl::opt` one-to-one 
> which was never our goal. I'd recommend looking at some of the previous 
> conversions, the description of those changes has more details:
> 
>     * https://reviews.llvm.org/D105532
> 
>     * https://reviews.llvm.org/D105598
> 
>     * https://reviews.llvm.org/D105605
> 
>     * https://reviews.llvm.org/D105330
> 
>     * https://reviews.llvm.org/D104889

@petrhosek Thanks for sharing those, I've taken a stab at reworking this PR to 
align with the semantic flag changes made in the other ones, such as dropping 
the single-dash long flags for consistency, removing `=` flags, simplifying 
bool flag handling, etc. This required a lot of changes to tests but it's a 
fairly mechanical update and aligns better with other tools, so maybe worth it? 
Not sure.

Mind taking another look at the llvm-profdata main and let me know if you see 
additional opportunities to use OptTable better? Would love to get this merged!

https://github.com/llvm/llvm-project/pull/177868
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to