Abandon this because we want to keep the ability to work with not-instrumented code.

Asan is suppose to work with prebuild non-instrumented code, so we are not as aggressive here.

Hi @MaskRay, I don't have commit access, could you help land patch this for me? Thanks.

Hi @MaskRay, I don't have commit access, could you help land patch this for me? Thanks.

test/fuzzer for Apple

Resolve merge

How does the change affect the output?

Update suffix

Kindly ping @MaskRay , could you help land this patch? Thanks.

Hi @MaskRay, I don't have commit access, could you please help commit this patch?

• test2

We don't want this patch to break things. I will abandon this patch then.

Yes, I would do that. Thank you.

merge

Hi @MaskRay, I don't have commit access, could you please help commit this patch?

Thank @MaskRay for your comment, I have updated the description of the patch.

Kindly ping @vitalybuka and @MaskRay.

Hi @MaskRay , could you please take a look at this patch and help land it?

Ping @vitalybuka, could you please review this patch?

• add some testes for x_aliases

Oh, I get your idea now, will add some access to n_aliases in a later commit.

Hi @vitalybuka, I think compiler-rt/test/sanitizer_common/TestCases/Linux/netent.cpp does not need updates. This patch does not change the functionality since sizeof(char**) == sizeof(char*), it just makes more sence using sizeof(char *) in the calculation of x_aliases's size.

In D149242#4301021, @vitalybuka wrote:

Would you like to try to update compiler-rt/test/sanitizer_common/TestCases/Linux/netent.cpp ?

I guess check-msan should trigger if you try to print n_aliases

• avoid nesting

In D145904#4218381, @vitalybuka wrote:

In D145904#4215399, @yingcong-wu wrote:

Yes, for now, this class is safe to do self-assign. But I think it is best to avoid doing that for it has no benefit doing so.

Here the reasoning. This one is not performance critical, but I don't see a point to break the rule.
https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#c62-make-copy-assignment-safe-for-self-assignment

Yes, for now, this class is safe to do self-assign. But I think it is best to avoid doing that for it has no benefit doing so.

Thank you. I will remember to mention if something is added to fix testes.

simplify the LastFailure condition

• set init value of HaveFtMarker to ture, for it's ok to not have any STARTED

Hi all, kindly ping.

Hi all, kindly ping.

add a comment to test

Hi @fmayer, I don’t have commit access, can you land this patch for me? Thanks.

another try to resolve merge conflict

resolve merge conflict

In D145758#4183561, @fmayer wrote:

Could you please rebase to fix pre-merge check patch application?

add const to references

In D145758#4183509, @fmayer wrote:

Can those be const auto&?

In D145035#4176999, @vitalybuka wrote:

Fuzzer is never deleted. I added ~Fuzzer() = delete; to be sure

formatting

formatting

In D145024#4163082, @snehasish wrote:

LGTM. Let me know if you like this to be committed by proxy.

add const to pod

In D130479#3782886, @sunho wrote:

Turned out the different library name was made by a bug in compiler-rt cmake file -- my local build was generating ".a" library files. It should be fixed in https://reviews.llvm.org/D133669. I'll update the code to use just ".lib" I'll invastigate the symbol not found failure shortly.

check-all failed with those tests: