Errors introduced when using certain regex in tests

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Use regex like '\P{L}' in a //x:expect/@test
2. resultant xspec.xsl will add an extra curly bracket, e.g. '\P{{L}',
which causes transformation to fail.

What is the expected output? What do you see instead?
see above

What version of the product are you using? On what operating system?
v0.1 - tested on WindowsXP using Saxon 6.5

Original issue reported on code.google.com by tomos.hi...@gtempaccount.com on 22 Mar 2010 at 6:12

[GoogleCodeExporter]

Saxon 6.5 is XSLT 1.0 while XSpec is written in XSLT 2.0.  I've just tried to 
compile a 
simple test suite with Saxon 6, and actually it does not even succeed to 
compile the 
XSpec stylesheets.  Are you sure you use Saxon 6?  Note I tried with Saxon 9 
and I 
cannot reproduce your problem.

Original comment by fgeorges on 22 Mar 2010 at 8:40

• Changed state: Invalid

[GoogleCodeExporter]

Thanks for the speedy reply, Florent.

Firstly I should apologise - it was quite late when I reported the bug 
yesterday, and
I'm not quite sure where I got the Saxon version number from.  I'm running 
Saxon9,
specifically version 9.0.0.4J.

I'm surprised that you could not reproduce the problem, as it occurs for both 
me and
one of our freelancers.  I'm attaching the XSpec document to see if we can 
establish
if it's a document specific issue or not.

Original comment by tomos.hi...@gtempaccount.com on 23 Mar 2010 at 10:15

Attachments:

• 0850.xspec.xml

[GoogleCodeExporter]

Thanks.  Unfortunately, I still cannot reproduce the problem.  I
tested with 9.0.0.8J (9.0.0.4 is not available anymore, 9.0.0.8
is tge only one available from 9.0.*) and it seems to be ok.

As I do not have the stylesheet "0850.xsl" I cannot really test
it from end to end, but I just compiled the test suite (.xspec to
.xsl) and the '{' character was not doubled.  Could you test with
9.0.0.8?  So at least we would be sure this is not related to a
fixed bug within Saxon.  If you still have the error with
9.0.0.8, then I'm afraid you'll have to provide a full simple
test (.xspec + tested .xsl + then command you used to compile and
run the suite) that reproduce the problem.

Thanks for your bug report!

Original comment by fgeorges on 23 Mar 2010 at 1:02

• Changed state: Started

[GoogleCodeExporter]

Original comment by fgeorges on 23 Mar 2010 at 1:03

[GoogleCodeExporter]

Hi,

Here's a 'dummy' XSL document for testing.

I've tested with 9.0.0.8 (with no avail) and have confirmed that our freelancer 
is
running saxonhe9-2-0-6j with the same results.

However, she has found the following issue, which seems like a plausible 
explanation
for the behaviour:

http://markmail.org/message/hncizg37qwvancdu#query:+page:1+mid:hncizg37qwvancdu+
state:results

Sadly neither she nor I have managed to implement Mike's suggestion in such a 
way as
it fixes the problem (we're not sure if we've overrided Xerces properly); are 
you
using the version of Xerces which ships with Java?

Original comment by yamah...@gmail.com on 23 Mar 2010 at 2:40

Attachments:

• 0850.xsl

[GoogleCodeExporter]

Ouch!  I see...  I am under Mac OS X here, and java -version gives me:

    java version "1.6.0_17"
    Java(TM) SE Runtime Environment (build 1.6.0_17-b04-248-10M3025)
    Java HotSpot(TM) 64-Bit Server VM (build 14.3-b01-101, mixed mode)

What do you use?  Can you try to use another Java version or to try to use a 
Xerces 
directly from Apache, as suggested by Mike?

Original comment by fgeorges on 23 Mar 2010 at 3:18

[GoogleCodeExporter]

I keep on trying to get my manager make the switch at work, but to no avail!  
Alas,
both of us are on Windows XP.

java -version gives me the same result:

java version "1.6.0_17"
Java(TM) SE Runtime Environment (build 1.6.0_17-b04)
Java HotSpot(TM) Client VM (build 14.3-b01, mixed mode)

I take it that you've not done anything to over-ride the Xerces libraries?

I will give the XSpec a shot on my system at home tonight, and in the mean-time
continue to try to get Apache's Xerces classes working here.

Original comment by yamah...@gmail.com on 23 Mar 2010 at 3:40

[GoogleCodeExporter]

Hi Florent,

I've just tried to run the same xspec file here at home in my mac environment 
with the same error (my java -
version is exactly the same as yours in this case).

Original comment by yamah...@gmail.com on 23 Mar 2010 at 7:32

[GoogleCodeExporter]

Oh, I completely missed you were using 0.1!  Sorry, that was clearly stated.  
So yes, 
that's a bug in XSpec 0.1 that's been fixed in 0.2.  I strongly suggest you try 
to move to 
0.2.

Sorry for the misunderstanding...

Original comment by fgeorges on 24 Mar 2010 at 11:27

• Changed state: Fixed

[GoogleCodeExporter]

Arg, so simple!

Thanks for your help.

Original comment by yamah...@gmail.com on 24 Mar 2010 at 11:57