[DRE-maint] Bug#534721: Bug#534721: Patch

[T Chan]

Gunnar Wolf wrote:
> Oh, I just now realized this bug has been reopened... I do think (but
> will not single-handedly demote it) the bug's severity has been
> inflated - I use hpricot on a frequent basis, and it's far from
> "renders the package useless". It does not work, right, but  for very
> specific use cases.

I don't think I "inflated" the severity — the original bug made hpricot unusable for XML parsing (with sufficiently long tags) on x86. It probably should've been downgraded when reopened (My "The bug in Message #25 is still present in 0.8.1-1 although not as grave" may not have been entirely clear).

For completeness, the remaining bugs:
* "<afuaf></zhgaa>" is treated as "<afuaf></afuaf>", since h("afuaf") == h("zhgaa").
* <a></b> is invalid. A compliant XML parser should return an error, hpricot returns <a></b></a> (which is also invalid XML, but at least it's a fixed point).

I don't personally care all that much, since I've since stopped using appstore_reviews.

- T Chan