Re: toy parser committed to cvs

["Frank Ch. Eigler" <fche at redhat dot com>]

Hi -


> [...]
> Is there any particular reason that tools such as antlr are not being 
> used to generate the parser?

I spent a week or two trying out several compiler generator tools,
including antlr.  In each case, the generated code was not up to
snuff in way or another: poor integration with hosting application,
too many files, poor class structure.  I was actually disappointed
to have to go it "alone".

> In dtr there were declarations for the original functions. What are you 
> thinking about for the syntax for the jprobes systemtap probes?  [...]

Syntactically, I'm thinking of representing jprobes-related C declarations
in some combination of strings and some elementary mapping to the systemtap
data types.  I hope we get going soon on the higher-level debuginfo library,
so we can use kprobes, and leave jprobes' unsightly crossing of C and script
behind.


- FChE