Re: [PATCH 3-5] fdo#47644 performance regression on largish .doc #2 ...

On Sun, 2012-05-13 at 00:32 +0100, Michael Meeks wrote:
>       So - I re-worked this to simplify, incrementally build the page chain
> cache which might help performance in nasty corner cases

I think it looks pretty good now anyway FWIW, I'd initially wanted to
build up an incremental chain, but knew I wouldn't get (even the simple
case) right in any reasonable length of time. Glad I didn't try given
the "busted storages, but we happen to stick to the fairly ok bits when
reading" case :-)

> It'd be worth testing any documents we know of, where previously we
> could recover some of the document content from the beginning of the
> stream, where the end was corrupt (I guess).

I don't know of any specific other examples outside of the existing
"special" ole2 cases we have in the regression tests. At least
collecting those together as cppunit tests has had some some payoff.

C.