Re: Lilypond error behaviour

[David Kastrup]

address@hidden writes:

> On Mon, 18 Apr 2016, David Wright wrote:
>> may be a decision that a machine can make, but whether a PDF is
>> suitable for further use (to debug the source code, give to a player,
>> or admire on the coffee table) is not.
>
> If compiling the file has failed to such a degree that we are choosing
> to call it a "fatal error," that means the PDF is *not* suitable for
> further use.  Conversely, if you think the generated PDF is still
> useful, then don't call what happened a "fatal error" - it must have
> really been a warning.

With that definition, any error that does not terminally destroy the
computer is not a fatal error since otherwise you can still debug the
problem.

While this definition has something going for it, its usefulness for
LilyPond may be limited.

-- 
David Kastrup