Tool added to extract network traffic annotations.

tools/clang/traffic_annotation_extractor/tests/README.md

Index: tools/clang/traffic_annotation_extractor/tests/README.md
diff --git a/tools/clang/traffic_annotation_extractor/tests/README.md b/tools/clang/traffic_annotation_extractor/tests/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..4ebd15b95d0241d4060e870b4cbefb4953eb3486
--- /dev/null
+++ b/tools/clang/traffic_annotation_extractor/tests/README.md
@@ -0,0 +1,13 @@
+# Manual tests for Traffic Annotation Extractor

[battre, 2017/04/20 13:41:03]

This testing makes no sense to me. What happens if those two files that are
completely out of control of code change?

Here is my proposal:

1) Create a file tools/clang/traffic_annotation_extractor/tests/test.cc with all
kinds of annotations you can think of.
2) Create a file
tools/clang/traffic_annotation_extractor/tests/expected_outputs/test.txt
3) Create a file tools/clang/traffic_annotation_extractor/unittests.py that runs
the extraction and verifies that the output matches the expected output.

Does that work?

[Ramin Halavati, 2017/04/20 13:53:46]

Acknowledged.

+This file describes manual sanity tests for traffic annotation extractor. A bug
+is filed (will be if you LGTM on delay) to add automatic tests in future.
+
+## Test:
+1. Run `third_party/llvm-build/Release+Asserts/bin/traffic_annotation_extractor 
+-p out/Debug components/autofill/core/browser/autofill_download_manager.cc > 
+~/Desktop/autofill_core_browser.txt`
+2. Run `third_party/llvm-build/Release+Asserts/bin/traffic_annotation_extractor
+-p out/Debug chrome/browser/bitmap_fetcher/bitmap_fetcher_browsertest.cc >
+~/Desktop/bitmap_fetcher_browsertest.txt`
+3. Expect the outputs to be similar to the content of 'expected_outputs'
+