how is cdepend different from the existing cross-reference tools?

... such as? Have I reinvented the wheel?

why did you find the need to read debug information from binaries, instead
of using some tool that parses the source (e.g. finding an existing
lex+yacc based tool that already parses the C code, and adding your own
parts in it? what is (are?) the advantage(s?) of parsing the binary file?

I don't parse the binary, but the assembly code. It's a text file. And the advantages are several:

1. Assembly code is so easy to parse.
2. Examining the code as compiled (including flags and command-line defines). But this is a disadvantage as well...
3. May work on other languages (g77?)



as for questions about the tool itself:


1. does it handle static functions?

Yes. It would be a mess otherwise.

2. i guess you didn't yet handle sources scattered on several directories
  - of you just presume someone will feed _all_ the '.S' files together,
  and you'll generate one gigantic directory for _all_ the functions in
  the code?

Exactly. If you want to look for unused functions, you have to check the whole thing together. So yes, it's basically a `find sourcedir -name \*.S`. And yes, you get one directory with all files together. Which is quite comfortable, because the list of files is the list of functions, so it's easy to find a function by name.

3. i presume you cannot handle function pointers passed around the code?

I will be able to tell that function A accessed the pointer to function C. But I can't tell who uses this pointer, that's right.

So Guy, please enlighten me: What other tool does the same?

Eli

--
Web: http://www.billauer.co.il



--------------------------------------------------------------------------
Haifa Linux Club Mailing List (http://www.haifux.org)
To unsub send an empty message to [EMAIL PROTECTED]




Reply via email to