Create Documentation using Doxygen and non-Doxygen commented source code - c++

Question

We got here some like 10 years old C and C++ code, without documents or manual. However the source is documented in the header files quite good, but it is a lot of working going through all the files looking for an information. It looks like this:
// Description description ....
//
// #param parameter 1 name:
// description of parameter 1
//
// #param parameter 2 name:
// description of parameter 2
//
Returntype Functionname(parameter1, parameter2);
Using doxygen wizard a documentation can be created but all the comments are lost, because they are not formated in a way the parser understands.
So is that a format i dont know? Can i teach the parser what to do? Or is it a special format that is used by another software?

Answer 1

I wrote a python script, to convert the comments to a format the parser understands. It's not pretty, it's not safe, but it works for us.
import re
import time
import os
import shutil
def convertHeaderDocumentation(file):
with open(file) as f:
lines = f.readlines()
lines = [line.rstrip('\n') for line in lines]
scanning = False
commentLines = []
convertedDocument = ""
declaration = ""
for line in lines:
if line == "" or \
line.strip().startswith("#"):
if len(commentLines) > 0:
convertedDocument += ''.join(el + "\n" for el in commentLines)
commentLines.clear()
convertedDocument += line + "\n"
continue
if line.strip().startswith('//'):
if not scanning:
commentLines.clear()
scanning = True
commentLines.append(line)
else:
if scanning:
if line.strip() != "":
declaration = line.strip()
match = re.search('\s*\w*\s*(\w+)\s+(\w+).*\((.*)[^)].*;', declaration)
if not match is None:
# check for function description
description = ""
for line in commentLines:
if line[2:].strip().startswith("#") or \
line[2:].strip() == "":
break
else:
description += line[2:].strip()
# scan for parameter description
parameters = []
parameter = ""
scanning = False
for line in commentLines:
# start scanning, if line starts with #
if line[2:].strip().startswith("#") and \
scanning == False :
# if present add to parameter lst
if parameter != "":
parameters.append(parameter)
scanning = True
parameter = line[2:].strip() + " "
continue
# stop scanning if an empty line is read
if line[2:].strip() == "":
scanning = False
# save if parameter is in buffer
if parameter != "":
parameters.append(parameter)
parameter = ""
if scanning == True and line[2:].strip() != "":
parameter += line[2:].strip()
convertedDocument += "/**\n"
convertedDocument += " * #fn " + declaration[:-1] + "\n"
convertedDocument += " *\n"
convertedDocument += " * #brief "
restLine = 80 - len(" * #brief ")
for index in range(0, len(description), restLine):
convertedDocument += description[index:index + restLine] + "\n * "
convertedDocument += "\n"
for parameter in parameters:
convertedDocument += " * " + parameter + "\n *\n"
convertedDocument += " * #return " + match.group(1) + "\n"
convertedDocument += " *\n"
convertedDocument += " * #date " + time.strftime("%d.%m.%Y") + "<br> parsed using python\n"
convertedDocument += " */\n"
convertedDocument += declaration + "\n\n"
commentLines.clear()
else :
convertedDocument += ''.join(el + "\n" for el in commentLines)
commentLines.clear();
return convertedDocument
projectDir = "path to source files goes here"
projectDir = os.abspath(projectDir)
parentProjectDir, projectDirName = os.path.split(projectDir)
convertedDir = os.path.join(parentProjectDir, "converted")
print(convertedDir)
for root, dirs, files in os.walk(projectDir):
# create directory structure if not present
tmpConvertedDir = os.path.join(convertedDir, root[len(projectDir) + 1:])
if not os.path.exists(tmpConvertedDir):
os.makedirs(tmpConvertedDir)
for file in files:
filename, fileextension = os.path.splitext(file)
# only copy/convert c++/c source files
if fileextension in {'.h','.c','.cpp'} :
newPath = os.path.join(tmpConvertedDir, file)
print(newPath)
# convert header files
if fileextension in {'.h'}:
#print("convert ", os.path.join(root, file), " to ", newPath)
converted = convertHeaderDocumentation(os.path.join(root, file))
with open(newPath, 'w') as f:
f.write(converted)
# copy source files
else:
print("copy ", os.path.join(root, file), " to ", newPath)
shutil.copyfile(os.path.join(root, file), newPath)
The function declaration was a bit tricky to catch with regex. For my case
\s*\w*\s*(\w+)\s+(\w+).*\((.*)[^)].*; works just fine, but without any extra keywords a lazy quantifier is going to be more acurate for keywords before the returntype \s*\w*?\s*(\w+)\s+(\w+).*\((.*)[^)].*;
All C and C++ files in a given projectDir directory and subfolders are converted if they are header files or just copy if they are source files. Therefore a directory ..\converted is created which contains the copied/converted files.
With the resulting files doxygen wizard created a sufficient documentation. Maybe this is going to help someone :-)