Generate a dashboard for ompparser's current capability

[chunhualiao]

Pick two sets of test files

• the official OpenMP examples: https://github.com/OpenMP/Examples/tree/master/sources
• DataRaceBech: https://github.com/LLNL/dataracebench/tree/master/micro-benchmarks

Test the standalone ompparser to generate a wiki table of this repo, similar to:
https://github.com/LLNL/dataracebench/wiki/Summary-report-Apr-2019

, with columns indicating file names and pass/fail info.

Make the process automated.

[ouankou]

The initial testing result for the official OpenMP examples is generated.
https://github.com/passlab/ompparser/blob/b05d2ee9ff1280bc37d20bbdcb8bcd5ac12380df/docs/openmp_examples_results.md

They are marked as PASS be default and will be checked manually later.

[ouankou]

The initial result for DataRaceBench examples is generated.
https://github.com/passlab/ompparser/blob/b05d2ee9ff1280bc37d20bbdcb8bcd5ac12380df/docs/dataracebench_results.md

Some of the results are recognized as FAILED correctly.

[chunhualiao]

@ouankou Please calculate the percentage of pass for two benchmarks. Manual calculation is fine (with help of grep of key words).

[yanyh15]

Anjia's work: test_folder.sh script is now used to extract OpenMP constructs from C examples and then invoke ompparser to parse and unparse. To validate, user needs to manually compare the ompparser output with the construct in the original C code. Ideally, a script to auto-validate would be useful.

[chunhualiao]

LLVM has a testing method to embed expected results into input source files and to automate compare compiler generated vs. expected results:
https://llvm.org/docs/TestingGuide.html

FileCheck - Flexible pattern matching file verifier

https://llvm.org/docs/CommandGuide/FileCheck.html