Re: Issue 1471: Invalidate alterations upon key change rather than forgetting them. (issue4384050)

[dak]

On 2011/04/12 03:39:17, Keith wrote:

> There is already a regtest 'accidental-clef-change.ly' that you could
just
> update.  Instead of
> texidoc = "Accidentals are reset for clef changes.", we want
> texidoc = "Accidentals are repeated after clef changes."
> You could put a c-natural in the final chord, and a key signature of
g\major to
> prove that you get right two accidentals that 2.12 got wrong.
>      \key g\major
>      cis dis f!
>      \clef bass
>      <c dis fis>

I intend to add an additional regtest file accidental-clef-change2.ly
with appropriately modified content.  Changing the existing regtest
in-place would not be a good idea I think since that would cause regtest
differences for the same file when actually nothing except the test
changes.

In order not to have to add yet another test or change it: is there a
particular reason you write f! above?  f would get a natural anyway.

So is this intentional and actually testing something, or is it just an
accident?  If the latter, I'd tend to drop it.  Should be perhaps write
e\major and have a dis! instead before the clef change?  That would
appear to actually check for something.

So the proposed regtest would be
\version "2.12.0"
\new Staff \relative c' {
  \key e\major
    \clef treble
    cis dis! f
    \clef bass
    <c dis fis>
}

I'd check this in before checking in the patch.  Am I overlooking
something?

http://codereview.appspot.com/4384050/