Remove compiler dependency: added new mapping test data format

[Niraj-Kamdar]

I have created separate python file to store mapping test data. I have done this because it would be nice to keep data separated from the code.
We can import whole array from the mapping_test_data file and use it directly in test_scanner.
@terriko, @pdxjohnny needs your opinions on this new structure.

[codecov-commenter]

Codecov Report

Merging #716 into master will increase coverage by 0.94%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #716      +/-   ##
==========================================
+ Coverage   87.71%   88.65%   +0.94%     
==========================================
  Files          76      125      +49     
  Lines        2173     2266      +93     
  Branches      264      263       -1     
==========================================
+ Hits         1906     2009     +103     
+ Misses        208      201       -7     
+ Partials       59       56       -3     

Flag	Coverage Δ
#longtests	88.65% <100.00%> (+0.94%)	⬆️

Impacted Files	Coverage Δ
cve_bin_tool/checkers/__init__.py	100.00% <ø> (ø)
cve_bin_tool/checkers/ffmpeg.py	100.00% <ø> (ø)
test/test_checkers.py	100.00% <ø> (ø)
test/test_scanner.py	95.94% <ø> (+4.51%)	⬆️
cve_bin_tool/checkers/gnutls.py	100.00% <100.00%> (ø)
cve_bin_tool/checkers/libcurl.py	100.00% <100.00%> (ø)
test/test_cli.py	100.00% <100.00%> (ø)
test/test_data/__init__.py	100.00% <100.00%> (ø)
test/test_data/bash.py	100.00% <100.00%> (ø)
test/test_data/binutils.py	100.00% <100.00%> (ø)
... and 103 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1f39bd3...2e4cada. Read the comment docs.

[pdxjohnny]

This format looks okay to me

[pdxjohnny]

• Do you mean to still have test/assets/test-curl-7.34.0.out and test/assets/test-curl-7.34.0.out?
• Where are the tests being run?
• Do are the common strings that we usually find when compiling with gcc being added to the fake binaries?

[pdxjohnny]

Looks like you're making great progress here!

[Niraj-Kamdar]

@pdxjohnny I am keeping two compiled binaries in assets because, testing our python strings module against fake binary won't be really useful. We have to make sure that our strings implementation give same result as strings command on real binaries.

No, I am not going to add those common strings in every binary file we generate. I think it would be nice to have a separate test for that so that developer get more details.

[Niraj-Kamdar]

This will close issue #638, #669 and #668.

[Niraj-Kamdar]

We don't have following test data.
no package for: bluetoothctl
no package for: gimp
no package for: kerberos
no mapping for: kerberos_5
no package or maping for: openssh-server
We should create issue for these and label it as first good issue.

[pdxjohnny]

Can you make those issues you were talking about? Then this looks good, also, see my comment above ^

[Niraj-Kamdar]

• Do you mean to still have test/assets/test-curl-7.34.0.out and test/assets/test-curl-7.34.0.out?
• Where are the tests being run?
• Do are the common strings that we usually find when compiling with gcc being added to the fake binaries?

Here is the answer for these comments: #716 (comment).
And I am making those issues :)

[SaurabhK122]

@Niraj-Kamdar, we have package test data for gimp, please see the commits in PR #726.

[Niraj-Kamdar]

@Niraj-Kamdar, we have package test data for gimp, please see the commits in PR #726.

I might have missed it. thanks!