Fix: duplicate scancode files being reported

[rnjudge]

When running scancode with an empty cache, Tern was reporting the same
file twice due to a logic bug in the add_file_data() function. This fix
adds a flag to ensure that a file is only added to the layer object
file list if it does not already exist there.

Resolves #1000

Signed-off-by: Rose Judge rjudge@vmware.com

[nishakm]

Can you check my logic here? I think this can be achieved this way:

for f in layer_obj.files <-- cycle through existing files
    merged = f.merge(checkfile) <-- try to update the file
    if not merged: <-- the file didn't get updated, means this is a unique file
        layer_obj.files.append(checkfile)

[rnjudge]

I think we still need the break and to update the file only when the for loop has been broken or completed. Otherwise, with your proposal, say we have three files in the layer [A, B, C] and two files from scancode that needs updating: [A, C]. We go through this for loop and compare A to A, merged = True so file gets updated but then we move to the next f in layer files and compare B from the layer files to the same A from scancode... merged = False and so we add the A file, even though it already exists and has been updated... and we are facing the same problem we have now. I think we could do:

merged = False <-- variable needs to exist outside of for loop so we can check it once the loop completes.
for f in layer_obj.files:
   merged = f.merge(checkfile)
   if merged:
      break
if not merged: <-- if we are here it's because we broke out of the loop (in which case the file has been updated and merged=True or because we never found a match and merged=False and file needs to be updated.
   layer_obj.files.append(checkfile)

Thoughts?

[nishakm]

Yup! This makes sense.