[plt-scheme] Re: ssax-code.ss appears broken under 359.200

John Clements <clements at ...> writes:

> 
> 
> On Nov 18, 2006, at 1:30 PM, Kyle Smith wrote:
> 
> > I'm using (planet "ssax.ss" ("lizorkin" "ssax.plt" 1 3)) version of  
> > Lizorkin's ssax package.  I just got around to revisiting a rather  
> > lengthy  piece  of  regression code that I had built up from his  
> > xml-tools-tutorial.  All of it had srfi-78 check statements that  
> > verified its proper functioning.   However, when I ran it again,  
> > for the first time under 359.200 I get:
> >
> > Error at position 3044
> > ([GIMatch] broken for  (END . syntax)  while expecting  END source)
> > uncaught exception: -1
> >
> > So I did a grep -i --recursive "GIMatch"  and got back the  
> > structure for this error message in the file lizorkin/ssax.plt/1/3/ 
> > ssax-code.ss.
> >
> > This entire regression file ran with no errors prior to installing  
> > 359.200.  I thought maybe you could take a look at the code and see  
> > if you could surmise what change in mzscheme might have broken  
> > ssax.plt.  It may affect other packages as well.
> 
> One thought: could this be related to the recent changes to the  
> meaning of for-each?
> 
> John Clements
> 
> 
> Attachment (smime.p7s): application/pkcs7-signature, 2484 bytes
> 
> _________________________________________________
>   For list-related administrative tasks:
>   http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>
Hi John,

Actually it turned out to be my bad.  For some unknown reason, my test XML 
document was missing an opening XML element within the body of the source.  
You see, I document my Scheme within block quotes with raw XML.  That way all 
I need to do is save it out to source.xml and pop it into Stylus Studio to 
manipulate and transform the document into whatever I want.  I use a rather 
large source.xml in my regression test for ssax/sxml, but I don't update it, 
to keep the regression test valid.  Well, it got changed somehow; probably 
inadvertently during what should have been a read-only Style Studio session.  
So, the cryptic error message makes a little more sense, but not exactly user 
friendly.  This got communicated rather quickly to Matthew, through e-mail, 
who was working on it immediately, as I wanted to save him as much effort as 
possible.  Sorry I didn't follow up on the post to the list.

Thanks for your reply,

--kyle
airfoil at bellsouth dot net
schemekeys.blogspot.com