lgtm

removed whitespace diff

fixed, break should occur if the env variable is not set OR if there was an error finding the variable, so anytime the return code is 0.

rebased to master

Planning to rebase this patch on master instead of release/9.x

clang formatted the patch

Committed revision 370390.
Committed abac219120d to svn.
git: https://github.com/llvm/llvm-project/commit/11802ccc9ff9348828128202ce91c549cc84a446

Fixed editor! Thank you for the tip :)

update to add !defined(clang)

closing, submitted as commit 08204d2610b095fbddea6959ea6a2fc8a4543ada

scrooched the includes up to fix the typedef issue with size_t

remove commented out work on GlobalAlloc :)

@mrstojo no worries! I'm very new and still learning how to be a good contributor so I appreciate feedback on what I'm doing wrong. I will remove the heapapi include and replace with just the function prototypes I needed along with those constants.

Updating for RNK comments

Just pinging again for review to land this if possible.

updated consexpr and namespace issue

Addressing comments

addressing comments from @vitalybuka
changing macros to constexpr
moving private stuff into asan namespace

Another bump for review from @rnk or @vitalybuka

@vitalybuka Heya, not marking as done but need some feedback from you.

addressing comments from @vitalybuka

Code cleaning, adding comments for clarity

Switches CMake config and environment variable to turn tests on to an ASAN_OPTIONS runtime variable. This makes the activation and testing cleaner, as well as the code itself.

@rnk @vitalybuka @mstorsjo ready for review :)

Modifying to make this option selectable at runtime. This makes it cleaner, without needing config and env var hacks

updating to mask RTL interception from x64 (not supported yet)

closed with revision 359498

fix BYTE typdef and incorrect format

Ready for review (clang-format only applied to new code)

fix call to get_usable_size so that it is not called unless HEAP_ZERO_MEMORY is passed in (the only case when it is needed).

Update diff to only apply clang-format to new code

In D61268#1482723, @vitalybuka wrote:

Let first land NFC reformatting part of this patch?
And rebase this patch on top of the NFC one.

Ready for review

adding tests for correct behaviors

Yep! 100% I made a mistake, the test should remain the same. I'm going to abandon this revision. Thank you for your quick review!

Oh wow, good catch. Thank you Reid.

Looks good! Thank you for the pointer to the mingw kit, also.

@rnk that suggestion works. I'll start testing mingw along with msvc before I commit in the future.

Added tests for interception of new functions (just in case).

Added stack alignment instruction which is common on msvc /Od executables

In D54581#1303419, @vitalybuka wrote:

In D54581#1303325, @mcgov wrote:

you can switch this patch into "Plan Changes" state

This needs some more work, I'll ping you all when I have a real fix.

Fix debug builds by defining a Windows section of the OPERATOR_NEW_BODY macro with the MemalignFromLocalPool call and if condition removed. This section is dead code on Windows, anyway.

Replaced #define with a function that asserts if called.

In D54581#1300511, @vitalybuka wrote:

And I don't see this code in the r346999

Could you please explain why you need to patch code which is not in the repo?

Fixing to account for Windows and Debug checks.