This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: LEX vs FLEX; Was: [PATCH] Basic Ada files
- From: Michael Snyder <msnyder at redhat dot com>
- To: Andrew Cagney <ac131313 at cygnus dot com>
- Cc: Hilfinger at gnat dot com, Aidan Skinner <aidan at velvet dot net>, gdb-patches at sources dot redhat dot com
- Date: Fri, 24 May 2002 10:59:49 -0700
- Subject: Re: LEX vs FLEX; Was: [PATCH] Basic Ada files
- Organization: Red Hat, Inc.
- References: <200205240300.UAA23495@tully.CS.Berkeley.EDU> <3CEDB860.9060308@cygnus.com>
Andrew Cagney wrote:
>
> > (I note you're not the author of this) I'm not exactly comfortable with
> >> making FLEX a condition of being able to build GDB - while the above
> >> tries to hide it, the dependency still exists. I guess we'll need to
> >> come back to that later.
> >
> >
> > Andrew,
> >
> > I am the author of that, so I suppose I should jump in. I'm not quite clear
> > on your objection here. Is it
>
> Just FYI, the current objective is to get the files under CVS, and then
> get all relevant CORE-gdb changes and other problems resolved.
>
> My request to Aidan was to just fix the Makefile.in patch submition so
> we can table it. The Makefile.in change will then be committed last.
>
> Any way for reference:
>
> > * the dependence on flex as opposed to lex?
>
> yes (but as a reservation, not an objection, I don't even know how
> pratical it is to get the code build using lex).
>
> > * the dependence on either lex or flex (unlikely given the
> > dependencies on yacc)?
>
> no
>
> > * the option to use the .c code and NOT depend on (f)lex at all?
>
> I suspect it will need changes. Compare it to
>
> c-exp.tab.c: c-exp.y
> $(SHELL) $(YLWRAP) "$(YACC)" $(srcdir)/c-exp.y y.tab.c
> c-exp.tmp -- $(Y
> FLAGS)
> ....
>
> However, like I said, all of this can be returned to later. I'm not
> asking Aidan to fix it now. I'm just looking for a patch sufficent to
> build/test the Ada files as Aidan commits them.
It's a sort of an interesting conversation.
GDB depends on several tools (autoconf, byacc) that the ordinary
user is not required to have, because we include their output in
the distribution. Really, only a developer is required to have
those tools, and only if he/she needs to change the source files
that those tools operate on.
So you might reasonably say that only the developer of ada-gdb
is actually dependent on having lex/flex. And by precident, you
might say that that's OK. A user will still be able to build
ada-gdb.