I was actually hoping to get rid of this code entirely.
Why do you need this change?

LGTM

but please update the commit message to explain why this enables us to use __GLIBC_PREREQ

The asan bot is unhappy:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/35078/steps/check-llvm%20asan/logs/stdio

LGTM with one nit, also asking Matt for a second review.

I would prefer to not introduce this complexity.
For periodic pruning we can use an empty dir, like you describe.
For stats, we can use the overal corpus size (in bytes and in files)

LGTM

With a build command line, of course

clang++ -g -std=c++17 string-view-use-after-free.cc  -fsanitize=scudo

Maybe add an example?
something like my favorite

#include <iostream>
#include <string>
#include <string_view>

Vitaly, please take a look as well.

Frankly, I would like to see a parallel effort in Musl.
I've seen the recent comment at https://github.com/google/sanitizers/issues/1080#issuecomment-506565699
suggesting that Musl is not likely to accept this approach.
But please look at the situation from our perspective:
this amount of #ifdefs in the code will increase the cost of maintenance for us, w/o giving much to anyone who doesn't use Musl.
And this is a wrong technical decision anyway, as we've known in 2011 (just didn't have other choice)
and as we've confirmed with our current implementation of HWASAN in Bionic.

Actually, the test started failing in ubsan mode right after this commit. Please fix ASAP. Reproduce: build with -DLLVM_USE_SANITIZER=Undefined, then "ninja check-llvm"

This new test is currently failing on the ubsan build.
Could you please take a look?
I don't yet know is the breaking change.

I like the simplicity of the fix!
Please add a comment explaining the fix, and a test that verifies how it works.

Thanks for doing this!

Hi, 
I think this is causing a failure on the asan bot. Please fix or revert ASAP

LGTM given a test
Thanks!

All other dfsan functions will actually continue working - most of them just won't be needed.

make total sense, but please wait for Evgenii's LGTM

LGTM, thanks!

reverted in r360190.

This change broke most of the buildbots:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/32015/steps/check-llvm%20asan/logs/stdio
lib/CodeGen/LexicalScopes.cpp:176: llvm::LexicalScope *llvm::LexicalScopes::getOrCreateRegularScope(const llvm::DILocalScope *): Assertion `cast<DISubprogram>(Scope)->describes(&MF->getFunction())' failed.