Yes, the behavior changed very recently, I would be surprised if
somebody came to depend on it. It's more likely that some clients are
depending on the old behavior.
- Marco.
Il 26/06/2018 22:43, Stephen Hines via Phabricator ha scritto:
> srhines added a comment.
>
> In https://reviews.llvm.org
Wouldn't it be better to keep compatibility with GCC and make
__gcov_flush have default visibility?
- Marco.
Il 26/06/2018 00:21, Xinliang David Li ha scritto:
> I don't have an objection having another interface which is just a
> simple wrapper to __gcov_flush but with default visibility. Also
I don't have an objection having another interface which is just a simple
wrapper to __gcov_flush but with default visibility. Also clearly document
its usage and behavior.
David
On Mon, Jun 25, 2018 at 10:12 AM, Chih-Hung Hsieh via Phabricator via
llvm-commits wrote:
> chh added a comment.
>
>
Should we have two versions of `__gcov_flush`?
One version is visible outside a .so file,
we use dlsym to find and call it to dump
profile data of .so files, like Android.
One version is hidden and must be called from another
wrapper function in a .so file. An application that
wants to flush .so
On Wed, Apr 11, 2018 at 11:31 AM, Chih-Hung Hsieh via Phabricator via
llvm-commits wrote:
> chh added a comment.
>
> Yes, calling `__gcov_flush` within .so files are different,
> but it's a revert of https://reviews.llvm.org/D38124.
> I think https://bugs.llvm.org/show_bug.cgi?id=27224
> can be
