Saxon

On the first one, we think this might be a spec problem. I think the intent of the spec is to disallow a grouping separator at position zero.

On the second one, we are having trouble reproducing it. We think the intended character is x0660 but that character appears to work fine in Saxon. Is it possible that some encoding issue has been lost in transit?

On the second one, we are having trouble reproducing it. We think the intended character is x0660 but that character appears to work fine in Saxon. Is it possible that some encoding issue has been lost in transit?

Yes, it seems that my entity was resolved in the HTML view. The query I tried should look as follows:

declare default decimal-format zero-digit = "٠";

format-number(1, '٠')

Once again it's not how it was supposed to look like. I discovered the "Preview" link, now it should be alright:

declare default decimal-format zero-digit = "٠";

format-number(1, '٠')

For the first issue, bug 28567 has been raised against the spec, to clarify whether a grouping-separator-sign is allowed at the end of the integer part.

I have now reproduced the second issue, when running the query directly from the command line. (But note that QT3 test format-number/numberformat155 containing this query passes.)

It seems that the XQuery parser is not applying unescaping (of &x..; sequences) to StringLiterals that appear other than in expression contexts. The problem apparently applies to the version and encoding declaration, option declarations, and the name in a processing-instruction node test as well as to values of decimal format declarations.

Patch committed for the 9.6 and 9.7 branches; test cases added to QT3.