Thread: Understanding a 40,000 lines C source code

Greetings Everyone,

I am trying to understand 40,000+ lines source code (12 different source files) so as to make modifications at various points in the code. I was the end user of that program and found many improvement areas.

Can you (experienced programmers) suggest some useful tools to understand the source code of this size written by someone else?

I did extensive search in this forum to find similar questions before posting this one. I tried DDD (debugger) but not really successful.

Thanks in advance for your valuable recommendations.

Sen

1) What is it in the code that you want to change / improve?

Looking for way s to improve code without understanding it's deficiencies, is like trying to cure a patient, before you know what's the problem(s)

After you know the problems, consider whether it would be better to write a new version, or to fix up the current version.

2) Look to the specifics that you identified in step #1, above. Don't try to "improve the code", try to improve one part of one function, of the code, that you have identified as the problem area.

Does it need a change in the algorithm of that function, or just a little extension of logic? Sometimes the best code you can write is the code you can delete and/or simplify.

Forget that it has 40,000 lines of code. Focus *only* on the area's that you want to fix, if you have decided that "to fix it" is the right way to go.

Be sure to keep an original, and document and comment, whatever you change. When you think you have improved it - put that aside, and test it. Nothing's worse than having made changes to 8 different functions, with a program this big, and *then* notice that the program no longer even works correctly - or runs in anything like the run-time that it used to.

What does this program do, and why do you need / want to improve it? (just curious).

Thanks Brewbuck and Ultraking !

I use both static and dynamic (ddd) methods as you mentioned, but in 4 months I just made progress upto 1300 lines in the main file, going back and forth to other files, looking what those functions do. The problem is that each function needs huge number of arguments of different complexity, so I find it difficult running them as individual pieces.


I am very novice in programming and unfortunately I don't have any direct contact with any experienced programmers. But I was pretty sure that there must be some better and efficient ways of doing it ----> prompting me to shoot the question here.

thanks again for your valuable time and suggestions.

It might be blatantly obvious, but after backing up the original I'd run the whole lot through some kind of code formatter/beautifier to get the source into a consistent and readable format (if it isn't); using something like astyle (Artistic Style - Index).

Salem, yes both tools sound great. I will explore them. Thanks everyone for your suggestions. I appreciate very much.

Sen

Originally Posted by brewbuck

40,000 lines in 12 files? So there must be at least one file longer than 3333 lines. Ugh...

The first application I worked on in my current job probably had a function with 3333 lines.
That may be a slight exaggeration, but one function had a seemingly endless series of multiple nested if-else-if statements.

At my old job I worked on a DOS based application where the original programmer passed no parameters to his functions. ALL variables were global. And these global variables were named flag1, flag2, flag3...flag1028...
His functions names were also cryptic. You had no idea what the function did by reading the name.

I hope my code is better.

well, in my case, of course the code has some useful comments, but may be because of the complexity of the algorithms, I find it difficult to grasp it faster.

The code I'm working in has TONS of functions > 1000 lines. Not sure what the longest is. When I read the title I thought perhaps the OP was one of our new hires