Re: Bison XML

[Joel E. Denny]

Hi Wojciech,

On Tue, 9 Oct 2007, Wojciech Polak wrote:

> Good work with today's XSLT commits :).

Thanks.  By the way, like many others here, I do really appreciate your 
contribution.

> But, can you send me all patches connected with
> this subject (Bison XML/XSLT) before committing
> them? I'd like help you at least by testing it
> before committing or by giving my comments on it
> (of course you can ignore them :).

Absolutely.  I hadn't heard any response to the discussion between Akim 
and me for a week.  In my experience here, that usually means there won't 
be any response, and I assumed you had moved on to other things.  I 
should've been more cautious in this case.  I apologize, but I can always 
back out the patch if it's too problematic.

> I'm just very interested in this area as I spent
> a lot of hours working on it and I just feel like
> it's my baby, you know ;))

I know how you feel, and I do not mean to take this out of your hands.  
Of course, at the same time, I want to do what's right for Bison, but I'm 
happy to work with you on that.  I'm glad to know you want to stay 
involved.

> And I'm asking about this, because I wrote the
> patch which tests less redundancy and I was just about
> to send it to bison-patches when you commited your
> stuff. I like it, but I just thought it's better
> to put patches before committing them, so other
> Bison developers can comment on them.

Yes, that's true.  I try to do that if I sense there's something 
controversial or there are other developers actively engaged in the issue.  
Otherwise, I try to move forward so that Bison has some hope of making 
progress.  Unfortunately, my radar is obviously wrong sometimes.  I don't 
find it easy to strike the right balance.

> My personal opinion is that I would stay for a while
> with the current solution, and not try to remove
> all redundancy too fast, because XSLT is not the only
> way to process XML (I just sent email about it to you
> and bison-patches).
> 
> What if one wants to process only a selected part
> of this XML? For instance with SAX? After removing
> all redundancy it's just very hard. My biggest .xml
> from Bison was C grammar - 2,1MB with --report-all.
> And I just think it's not too big.

Ok, I'll pick this discussion up in the other thread.

> BTW, I'm looking for some real world examples
> of the biggest grammars for testing.
> Do you know any?

The ISO 2003 C++ grammar is 4.9 MB when the lexical section is flattened 
into just token declarations.  Gpic (Groff 1.18.1) is about 3.2 MB.  The 
latter is in the Bison test suite.  Of course, you have to purchase the 
former.  I used --report=all and --xml as well.