CPU hogging in the current master

[Linus Torvalds]

On Thu, Sep 26, 2013 at 3:38 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
>
> Not sure I understand the "faking the xml part" comment...

So I'd basically create all the same names that xml parsing creates,
and feed it to the rest of the parser.

> Your example doesn't include the notes

Yeah, I don't have the exporter done for that, and had to go lie down
for a headache..

> are we leaving it uncompressed in the file format or are we going to
> compress it by default?

Compress it, I think.

> any line starting with a number or a + is a sample? I notice that you
> don't have a "from here on out it's samples" keyword

Any line that doesn't have a tag (or a continuation marker).

> how are we creating trips?

Our trips don't actually nest even if syntactically the xml could, and
dives outside of trips have the NOTRIP thing, so we don't even need
the "trip end" marker, although maybe I'll do that just for clarity.

Basically everything is just "one line at a time", with no real
context. So a "trip" line starts a new trip, and all subsequent dives
will be part of that trip (modulo NOTRIP). Same for "dive", etc etc.

           Linus