For a deleted file, invalid diff received

[pratikjain1993]

I have received the following diff object for a file which I had deleted in the commit : c9050a81069d94b9b9692589608ec8f8d6c52827 .

myFile = {Diff} webapp/META-INF/ion-config/EU/pipeline_context_data/R2L/UK_R2l_SOA_SILVER/BankAccountFragment/InputSets/DE_IBAN.ion\n=======================================================\nlhs: None\nrhs: 100644 | e69de29bb2d1d6434b8b29ae775ad8c2e48c5391\nfile added in rhs
 NULL_BIN_SHA = {str} '��������������������'
 NULL_HEX_SHA = {str} '0000000000000000000000000000000000000000'
 a_blob = {NoneType} None
 a_mode = {int} 33188
 a_path = {unicode} u'webapp/META-INF/ion-config/EU/pipeline_context_data/R2L/UK_R2l_SOA_SILVER/BankAccountFragment/InputSets/DE_IBAN.ion'
 a_rawpath = {str} 'webapp/META-INF/ion-config/EU/pipeline_context_data/R2L/UK_R2l_SOA_SILVER/BankAccountFragment/InputSets/DE_IBAN.ion'
 b_blob = {Blob} e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
 b_mode = {int} 33188
 b_path = {unicode} u'webapp/META-INF/ion-config/EU/pipeline_context_data/R2L/UK_R2l_SOA_SILVER/BankAccountFragment/InputSets/DE_IBAN.ion'
 b_rawpath = {str} 'webapp/META-INF/ion-config/EU/pipeline_context_data/R2L/UK_R2l_SOA_SILVER/BankAccountFragment/InputSets/DE_IBAN.ion'

What I understood was that I should receive b_path/b_blob/b_mode as none, but it seems not to be working fine.

[Byron]

Is this issue similar to this one: #531 ?
Otherwise I would be tempted to close this one, as it is impossible to understand what is expected, and how to reproduce the issue with GitPython.

[pratikjain1993]

My Source code:

import os
import whatthepatch
from git import Repo
import unicodedata

#workspacePath = os.getcwd()
workspacePath = '/Users/jainprat/jainprat-SWIPERTOL/src/SWIPERTOLWebStack'
repo = Repo(workspacePath)
repo.config_reader()
cln = raw_input('Enter the commit id: ')
hcommit = repo.commit(cln)
oldCommit = hcommit.parents[0]
hcommit.diff()  # diff tree against index
mydiff = hcommit.diff(oldCommit, create_patch=True)
outputFile = open('output.ion','w')


def getIonStartingEndingLineTuple(ionStartingAndEndingLineNumbersTuples, changedLineNumber):
    for entry in ionStartingAndEndingLineNumbersTuples:
        if changedLineNumber >= entry[0] and changedLineNumber <= entry[1]:
            return entry
    return None


def getStartingEndingLineNumbersForIon(allLines):
    result = []
    lineNumber = -1
    bracesCount = 0
    entry = []
    for line in allLines:
        lineNumber += 1
        if "{" in line:
            if bracesCount == 0:
                entry.append(lineNumber)
            bracesCount += 1
        if "}" in line:
            bracesCount -= 1
            if bracesCount == 0:
                entry.append(lineNumber)
                result.append(entry)
                entry = []
    return result

def getChangedIons(myFilePath, addedLineNumbers, removedLineNumbers, modifiedLineNumbers):
    changedIons = ""
    if len(removedLineNumbers) > 0 or len(modifiedLineNumbers) > 0 :
        unicodeContents = repo.git.show('{}:{}'.format(oldCommit.hexsha,myFilePath))
        oldFileContent = unicodedata.normalize('NFKD',unicodeContents).encode('ascii','ignore')
        oldFileContent = oldFileContent.split('\n')
        oldFileStartingAndEndingLineNumbersTupleList = getStartingEndingLineNumbersForIon(oldFileContent)
    unicodeContents = repo.git.show('{}:{}'.format(hcommit.hexsha,myFilePath))
    newFileContent = unicodedata.normalize('NFKD',unicodeContents).encode('ascii','ignore')
    newFileContent = newFileContent.split('\n')
    newFileStartingAndEndingLineNumbersTupleList = getStartingEndingLineNumbersForIon(newFileContent)

    changedEntries = set([])
    entry = None
    for lineNumber in addedLineNumbers:
        if entry != None and lineNumber >= entry[0] and lineNumber <= entry[1]:
            continue
        entry = getIonStartingEndingLineTuple(newFileStartingAndEndingLineNumbersTupleList,lineNumber)
        if entry != None:
            changedEntries.add(tuple(entry))
    entry = None
    for lineNumber in modifiedLineNumbers:
        if entry != None and lineNumber >= entry[0] and lineNumber <= entry[1]:
            continue
        entry = getIonStartingEndingLineTuple(newFileStartingAndEndingLineNumbersTupleList, lineNumber)
        if entry != None:
            changedEntries.add(tuple(entry))
    entry = None
    for lineNumber in removedLineNumbers:
        if entry != None and lineNumber >= entry[0] and lineNumber <= entry[1]:
            continue
        entry = getIonStartingEndingLineTuple(oldFileStartingAndEndingLineNumbersTupleList, lineNumber)
        if entry == None:
            continue
        flag = True
        for i in range(entry[0],entry[1]):
            if i not in removedLineNumbers:
                flag = False
                break;
        if flag:
            changedEntries.add(tuple(entry))
    for changedEntry in changedEntries:
        for i in range(changedEntry[0],changedEntry[1]+1):
            changedIons = changedIons+"\n"+newFileContent[i]
    return changedIons

for myFile in mydiff:
    **myFileAPath = myFile.a_path
    myFileBPath = myFile.b_path**
    if myFileBPath != None and myFileBPath.endswith('.ion'):
        print myFileBPath
        modifiedLineNumbers = set([])
        addedLineNumbers = set([])
        removedLineNumbers = set([])
        parsedDiff = whatthepatch.parse_patch(myFile.diff)
        for content in parsedDiff:
            for change in content.changes:
                if change[1] != None and change[0] != None:
                    modifiedLineNumbers.add(change[1])
                elif change[0] == None:
                    removedLineNumbers.add(change[0])
                else:
                    addedLineNumbers.add(change[1])
        outputFile.write(getChangedIons(myFileBPath, addedLineNumbers, removedLineNumbers, modifiedLineNumbers))
    elif myFileBPath == None and myFile.a_path != None:
        myFileBPath = myFile.a_path
        print myFileBPath
        unicodeContents = repo.git.show('{}:{}'.format(hcommit.hexsha, myFileBPath))
        newFileContent = unicodedata.normalize('NFKD', unicodeContents).encode('ascii', 'ignore')
        outputFile.write(newFileContent)
outputFile.close()

I believe that this could be related to #531 since this seems to be an issue with deleted/added file too.
If you see the bold font code, for a deleted file in the diff, both myFileAPath and myFileBPath are not None.

[Byron]

Thank you! Could you please reduce this to the smallest amount of self-contained code that still shows the issue you are describing ?

[pratikjain1993]

Sorry for giving such a large piece of code:
Here is the meaningful code required:

workspacePath = '/Users/jainprat/jainprat-SWIPERTOL/src/SWIPERTOLWebStack'
repo = Repo(workspacePath)
repo.config_reader()
cln = raw_input('Enter the commit id: ')
hcommit = repo.commit(cln)
oldCommit = hcommit.parents[0]
hcommit.diff()  # diff tree against index
mydiff = hcommit.diff(oldCommit, create_patch=True)

for myFile in mydiff:
    myFileAPath = myFile.a_path
    myFileBPath = myFile.b_path

here , both myFileAPath and myFileBPath are not none in case of a
deleted file for me.

On Sun, Oct 16, 2016 at 11:47 PM, Sebastian Thiel [email protected]
wrote:

Thank you! Could you please reduce this to the smallest amount of
self-contained code that still shows the issue you are describing ?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#534 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AVz5LoHUMCvSwPHMR2EXL9eE5lYDkTiKks5q0mpPgaJpZM4KX_Xp
.

[Byron]

Unfortunately, this is still not an example that can make the issue reproducible on a machine that is not yours. It requires input, and a repository that is not publicly accessible. You can have a look at this extraordinary example, which can be found in this issue.

What I wonder is if the suggestion of #531 of just inverting the code would work for you.

[pratikjain1993]

Does inverting not imply that b*** will be replaced by a*** ? If yes, my issue will still persist.
Do not have a github public repo yet. Will create one and try to reproduce and revert back to you here.

[pratikjain1993]

Not able to repro on a new repo.
Need to look what could be different for my use case.

[Byron]

I am closing this issue as it had no interaction for more than 3 months. Please feel free to comment in case you need it to be reopened.