On Thu, 6 Jun 2013 14:39:58 +0200 (CEST)
Johannes Meixner <[email protected]> wrote:

> 
> Hello,
> 
> On Jun 6 11:16 Ladislav Slezak wrote (excerpt and rearranged):
> > Dne 6.6.2013 10:19, Jiri Srain napsal(a):
> >>
> >> the comments are one of the tasks, which is currently being worked
> >> on. So you re right, we intend to translate the comments as well.
> ...
> > Of course, comments are essential part of the code so we switch to
> > ruby only after this is solved. This is a blocker for us.
> 
> Then I also appreciate your switch from YCP to Ruby very much!
> 
> I look forward to having openSUSE 13.1 (and then also SLE12) without
> YCP!
> 

Thanks for support. I hope result will satisfy you.

> 
> > Yes, Josef is working on it, the problem is that the YCP parser
> > does not care about comments and throws them away. The major
> > problem is how to bind the comments properly with the code, you
> > need to know the context around the comment to decide correctly and
> > that's a problem during parsing (you do not know what will be the
> > next node until it's actually parsed...)
> 
> I am a total layman regarding such issues nevertheless I dare
> to make a comment:
> 
> What about changing YCP comments into actual YCP code (e.g. change
> YCP comments into YCP strings using special string variable names)
> so that it becomes translated into actual Ruby code (e.g. into Ruby
> strings) and then change the Ruby strings with those special string
> variable names back to Ruby comments?

Currently we experiment with two different approach. One is based on
changed grammar where comments are regular token and the second one is
based on enhancing tokens from lexer.
Your proposal sound interesting, but there is problem, that you cannot
easily change comments into statement like
string ___comment = "something";

because comments occur on various places like
a && /* comment */ b
so replacing such comment into string statement will be quite
problematic.
We plan to make decision about approach and start with early attempts
to translate comments next week.

> 
> I assume this might result a quoting hell when YCP comments
> contain YCP code (e.g. YCP code that is commented out).
> 

Yes, dead code is not nice in comments, but we agreed that we don't
investigate semantics of current code and just transform it into ruby
comments.


Josef

> 
> Kind Regards
> Johannes Meixner

-- 
To unsubscribe, e-mail: [email protected]
To contact the owner, e-mail: [email protected]

Reply via email to