Re: Lilypond error behaviour

[David Kastrup]

Andrew Bernard <address@hidden> writes:

> If you look at the MWE example I provided to illustrate what I believe
> is misleading information about a ‘FATAL’ error, then it has plain
> text which it ignores, claims a fatal error and proceeds to make a
> perfectly good PDF.

LilyPond wouldn't know that, and in any way it returns an ERROR message
and an ERROR code.

I don't know why you are obsessed about LilyPond not touching the PDF
(in which case you still would not have an idea about whether LilyPond
was successful unless you checked the PDF's modification date) or trying
to leave behind an invalid PDF file (in which case you need to call a
PDF verification program in order to figure out whether there was an
error).

Also, some runs of LilyPond may produce a PDF with a different name or
indeed no PDF at all (you can run LilyPond for indexing purposes or
information gathering and a number of other things).

> Based on this it means one can have blocks of text as comments or
> documentation with no syntax, happily ignore the error, and have a
> nice new way of annotating lilypoind source code files. I hope people
> can see that this is clearly absurd.

It is not LilyPond's task to make it harder for people to ignore errors
when they really try hard.

> It’s the FATAL error message that I am questioning. That terminology
> in my IT world means the program cannot go on. It’s a simply
> confusing. adjective. One could say given the behaviour that it should
> be a warning.

A warning is appropriate for something which is not an error: namely
LilyPond has a well-specified task but the results will not likely make
sense.  LilyPond does not return an error code for a warning.

-- 
David Kastrup