git diff adding a new empty file is not parsed correctly

[wbolster]

a git diff adding a new empty file is not parsed correctly. the patch set contains 0 items in that case, while it should contain one.

sample:

diff --git a/empty b/empty
new file mode 100644
index 0000000..e69de29

[matiasb]

I see, you're right, will take a look.
Thanks for your report!

[archen]

This is a problem for any changes to binary files. Capturing these changes in the metadata and being able to iterate over them as part of the PatchSet would be great.

[wbolster]

fwiw a while ago i made a start on a diff/patch parser. i just published that w-i-p code here:

https://github.com/wbolster/diffparse

it tries to get metadata from svn and git diffs

[wbolster]

sample output from that parser (see __main__ module):

$ python -m diffparse some-git-commit.diff 
patch set: PatchSet(preamble=['commit c91d3dbeff576c3f0ea8aab5848d1b95af290544 (HEAD -> master, origin/master)', 'Author: Wouter Bolsterlee <wouter@bolsterl.ee>', 'Date:   2015-12-28 21:55:47 +0100', '', '    Ignore leading/trailing whitespace', ''], patched_files=[PatchedFile(source_file='diffparse/__main__.py', target_file='diffparse/__main__.py', source_timestamp=None, target_timestamp=None, git_header=GitHeader(source_file='diffparse/__main__.py', target_file='diffparse/__main__.py', old_mode=None, new_mode=None, deleted_file_mode=None, new_file_mode=None, copy_from=None, copy_to=None, rename_from=None, rename_to=None, similarity_index=None, dissimilarity_index=None, index_from_hash='f211439', index_to_hash='8209424', index_mode='100644'), svn_header=None), PatchedFile(source_file='diffparse/parser.py', target_file='diffparse/parser.py', source_timestamp=None, target_timestamp=None, git_header=GitHeader(source_file='diffparse/parser.py', target_file='diffparse/parser.py', old_mode=None, new_mode=None, deleted_file_mode=None, new_file_mode=None, copy_from=None, copy_to=None, rename_from=None, rename_to=None, similarity_index=None, dissimilarity_index=None, index_from_hash='007ebf6', index_to_hash='e899075', index_mode='100644'), svn_header=None)])
patched file: PatchedFile(source_file='diffparse/__main__.py', target_file='diffparse/__main__.py', source_timestamp=None, target_timestamp=None, git_header=GitHeader(source_file='diffparse/__main__.py', target_file='diffparse/__main__.py', old_mode=None, new_mode=None, deleted_file_mode=None, new_file_mode=None, copy_from=None, copy_to=None, rename_from=None, rename_to=None, similarity_index=None, dissimilarity_index=None, index_from_hash='f211439', index_to_hash='8209424', index_mode='100644'), svn_header=None)
hunk: Hunk(source_start=5, source_length=7, target_start=5, target_length=7, section='from . import parse_patch_sets')
line: Line(type=' ', value='', source_line=5, target_line=5)
line: Line(type=' ', value='def main():', source_line=6, target_line=6)
line: Line(type=' ', value='    with open(sys.argv[1]) as fp:', source_line=7, target_line=7)
line: Line(type='-', value='        for patch_set in parse_patch_sets(fp):', source_line=8, target_line=None)
line: Line(type='+', value='        for patch_set in parse_patch_sets(fp, allow_preamble=True):', source_line=None, target_line=8)
line: Line(type=' ', value="            print('patch set:', repr(patch_set))", source_line=9, target_line=9)
line: Line(type=' ', value='            for patched_file in patch_set.patched_files:', source_line=10, target_line=10)
line: Line(type=' ', value="                print('patched file:', repr(patched_file))", source_line=11, target_line=11)
patched file: PatchedFile(source_file='diffparse/parser.py', target_file='diffparse/parser.py', source_timestamp=None, target_timestamp=None, git_header=GitHeader(source_file='diffparse/parser.py', target_file='diffparse/parser.py', old_mode=None, new_mode=None, deleted_file_mode=None, new_file_mode=None, copy_from=None, copy_to=None, rename_from=None, rename_to=None, similarity_index=None, dissimilarity_index=None, index_from_hash='007ebf6', index_to_hash='e899075', index_mode='100644'), svn_header=None)
hunk: Hunk(source_start=408, source_length=6, target_start=408, target_length=14, section='class SubversionHeader(object):')
line: Line(type=' ', value='', source_line=408, target_line=408)
line: Line(type=' ', value='def parse_patch_sets(fp, allow_preamble=False):', source_line=409, target_line=409)
line: Line(type=' ', value='    it = PeekableFile(fp)', source_line=410, target_line=410)
line: Line(type='-', value='    while it.peek() is not None:', source_line=411, target_line=None)
line: Line(type='-', value='        patch_set = PatchSet._from_peekable(it, allow_preamble=allow_preamble)', source_line=412, target_line=None)
line: Line(type='-', value='        yield patch_set', source_line=413, target_line=None)
line: Line(type='+', value='    while True:', source_line=None, target_line=411)
line: Line(type='+', value='        next_line = it.peek()', source_line=None, target_line=412)
line: Line(type='+', value='        if next_line is None:', source_line=None, target_line=413)
line: Line(type='+', value='            break', source_line=None, target_line=414)
line: Line(type='+', value='        elif not next_line.strip():', source_line=None, target_line=415)
line: Line(type='+', value='            # Silently ignore leading and trailing white-space only lines.', source_line=None, target_line=416)
line: Line(type='+', value='            next(it)', source_line=None, target_line=417)
line: Line(type='+', value='        else:', source_line=None, target_line=418)
line: Line(type='+', value='            patch_set = PatchSet._from_peekable(', source_line=None, target_line=419)
line: Line(type='+', value='                it, allow_preamble=allow_preamble)', source_line=None, target_line=420)
line: Line(type='+', value='            yield patch_set', source_line=None, target_line=421)