Re: Source code review package?

On Sun, 2008-10-12 at 09:58 -0500, Joe Pluta wrote:

Terrence Enger wrote:

It is sad that the system allows a program to construct names of
programs and commands a run-time. Necessary, perhaps, but still sad.

Is there a system that doesn't allow this? Every non-i machine i know
allows you to create a script and then execute that script, which means
you can do anything.

Indeed, and it can be a lot worse than what we commonly face. Consider
the C/C++ preprocessor, whereby any word in the source code can
represent any sequence of tokens, some of which may be words which ...

And there's good reason for it: just about every OS has some sort of
menuing tool as well, which allows you to enter some arbitrary line of
code and execute it when the user selects an option. Otherwise, you'd
have to write hardcoded menus for each and every user.

Indeed.

Not sure why this is sad.

Only for those who want to be able to say confidently and with little
effort that some body of code does not, for example, GRTOBJAUT at all.

In my own work, I tend just to *assume* that any program name or command
name or procedure name will appear explicitly somewhere in the body of
code.

The ability to split a token across source lines also confounds scanning
source code.

It only confounds scanning code that can't handle line continuation
characters.

FNDSTRPDM being one such tool, unless there is something I do not know
about it. Is there?

Cheers,
Terry.

Joe