[PATCH] Add new debug kind LocTrackingOnly.

David Blaikie dblaikie at gmail.com
Fri Jun 20 12:44:31 PDT 2014


On Fri, Jun 20, 2014 at 12:01 PM, Diego Novillo <dnovillo at google.com> wrote:
>
>
>
> On Fri, Jun 20, 2014 at 2:55 PM, David Blaikie <dblaikie at gmail.com> wrote:
>>
>> Probably worth wiring up the command line support in the same patch &
>> including test coverage.
>
>
> Not sure what you are asking for here.

Sorry, misread and assumed this was a clang patch.

I'm not sure whether adding the value to this enum is the right
option, since we'll never see it in the backend (since the attribute
is added to the CU, which won't be used by LLVM whenever this
enumerator is used), but it does fit with the other values here.

Eric - what do you reckon? Use a separate flag just for DIBuilder to
know that it should omit the dbg.cu node, or add it to this existing
flag even though it'll never end up in an actually useful CU metadata
node (unlike the other two values)?

>
>>
>> I'd probably start with the LLVM change (using a local Clang change to
>> generate IR, or just hand-hacking the removal of the llvm.dbg.cu node
>> in an IR sample) to make sure this strategy is plausible and that just
>> dropping the llvm.dbg.cu node is all that's required from the
>> front-end.
>
>
> Eh, this is exactly how I know this strategy works. I've added a new test in
> the backend for this... gah, i forgot to add it to the patch. Coming up.

Ah - OK. Thanks! :)

- David



More information about the llvm-commits mailing list