Re: Can Ebrowse or ECB give me a list of functions called? Or something else?

[Alan Mackenzie]

Eli Zaretskii <eliz@gnu.org> wrote on Fri, 06 Aug 2004 12:19:58 +0300:
>> From: Alan Mackenzie <acm@muc.de>
>> Newsgroups: gnu.emacs.help
>> Date: Thu, 5 Aug 2004 19:55:35 +0000

>> We have a source tree of mixed C and C++.  Management is considering
>> "just" switching to a different unix-like OS.  I have been tasked with
>> discovering exactly which functions (from the OS and standard
>> libraries) are called.

>> Before embarking on the writing of a script to parse our source files,
>> I was wondering if I could somehow extract the information from an
>> Ebrowse or ECB database file.  Clearly a TAGS file doesn't contain the
>> requisite info.

> I'd begin with GNU ID-Utils.  They generate a database of all the
> symbols in a source tree, and let you run queries on that adatabase.
> There's a nice grep-like Emacs inteface to the most frequently-used
> query (find all the lines where a certain symbol or a symbol-name
> pattern is mentioned), but even if you use the less frequent queries
> outside Emacs, you could then add some simple scripts on top of that.

> Look at the GNU FTP site for the latest version of ID-Utils.  I find
> it a must when browsing a large source tree.

Thanks for this tip, Eli.  I had quite a bit of fun with ID-Utils.  Then
Management's priorities suddenly changed, and the task is now non-urgent.
Such is life.  ;-(

But it was interesting to see how ID-Utils identified _all_ "ID"s - even
numbers and keywords (like "while").  It's not quite what I needed, but I
might be able (_should_ be able) to extract what I need from it.

I'll look more closely at ID-Utils.

Thanks again, E.

-- 
Alan Mackenzie (Munich, Germany)
Email: aacm@muuc.dee; to decode, wherever there is a repeated letter
(like "aa"), remove half of them (leaving, say, "a").