As static memory is implicitly zero'd, the defaults are 0 and false.
This is rather useful.

Previously we were setting the maximum buffer size. For some soundcards
this could allow the device to selecta much smaller buffer, leading to
overruns.

Be more specific by using _near, which should match the user's expectations
and perhaps will fix problems with some soundcards.

The term "vinyl emulation software" is rather dated and whilst probably
more accurate, potential xwax users will be looking for a DVS.

In some edge cases of the layout the caller has decided there is enough
room to draw the library, but then there isn't an zero rows can be
fitted.

This is an indication that the whole layout functionality needs a re-think
in the way it deals with making areas invisible as they become too small.

A possible solution for later is separation into a layout and then a redraw
step, whereas at the moment the two are tightly coupled.

I tested with

  xwax -g /2.4

which prior to this patch crashes instantly.

Do memory allocatin within the index as a separate step.
It means that calling code can reserve memory up front and then
the insert/add functions cannot fail (and therefore no tricky
recovery code needs to be written.)

It also provides a simple optimistion in the index_copy() case.

Fixes the FIXME.

Some old left over whitespace from a long time ago.

Clean up where we have removed low-level code, and no longer require
the header files.

Be extra assertive. There is always the "All records" crate so for
now delete any code for dealing with the zero-crates case.

The loss of accuracy needs to be in rounding the row height to pixels,
before calculating the number of rows; the rows are to be drawn with
pixel (not sub-pixel) positioning.

Some refactoring so that this makes logical sense, the library
provides a function to rescan a crate, which may or may not be
acknowledged.

Happy New Year!

Some basic profiling on a non FPU processor such as ARM seems to suggest
that the interpolation takes most of the time.

This patch is expected to give a minor performance improvement as the
large number of adds/subtract are now done in integers, but still to
do is the multiplacation by mu, mu^2, mu^3.

Mainly it's here to provide clearer input/output to the interpolation
function.

Ubuntu called again, it's moved its fonts, again :)

Thanks to Daniel Holbach for pointing this out.

Reluctantly specify that the scan script is bash, but without this
we have no way to catch errors in eg. the cat command (like file
not found)

We no longer blank listings (as before); instead a new one is
initialised and used.

Restore the display of the current crate in the GUI by attaching
the listing to the one given to us by the crate.

Then the listing itself is actually reference counted. The previous
design was wrong and has many concurrency issues associated. This
is expected to be a much easier and well-structured way to solve them.

This commit begins to show some cracks in the ways we manage the
handles to the excrates.