Re: a few more comments ...

On Mon, 2017-06-05 at 21:14 +0100, Michael Meeks wrote:
> Hi Kohei,
>
> On 05/06/17 20:54, Kohei Yoshida wrote:
> > On Mon, 2017-06-05 at 19:40 +0100, Michael Meeks wrote:
> > >         Suggests (input from Kohei appreciated) - that we may
> > > need to
> > > do a new
> > > 'seek' each time we come to reading some bytes from the
> > > underlying
> > > package/ stream - to ensure that no other thread is messing with
> > > our
> > > underlying stream position while we are reading compressed data
> > > from
> > > it =)
> > Sorry, but can someone give me a brief summary of what you guys are
> > discussing?  What exactly is the problem, and what type of input
> > are
> > you expecting from me?
>       The patch is gerrit:
>
>       https://gerrit.libreoffice.org/#/c/38135/
>
>       And the basic conflict is that your fix solved the bug by doing
> all the
> unzipping ahead of time, and that's rather in conflict with threading
> and parallelizing the unzipping - which is the goal of this work =)
>
>       So - advice on re-doing the fix to improve package to cope with
> multiple threads doing seeking etc. would be great if possible =)

I'm afraid I don't have any advice here that I can give off the top of
my head.  I have no clue what work would involve threading and
parallelizing the unzip to begin with, much less advice to fix what I
fixed while achieving your goal...

I would probably have to spend some considerable amount of time
researching before I could even come up with something tangible. :-(

Kohei