Re: More regtest differences

[address@hidden]

On 29 mars 2013, at 20:36, Janek Warchoł <address@hidden> wrote:

> On Fri, Mar 29, 2013 at 7:04 PM, Phil Holmes <address@hidden> wrote:
>> ----- Original Message ----- From: "Janek Warchoł"
>> <address@hidden>
>> 
>>> On Fri, Mar 29, 2013 at 5:48 PM, Phil Holmes <address@hidden> wrote:
>>>> it seems that all system start braces have
>>>> been pushed a little bit further from the initial system clef.
>>> 
>>> Isn't this a rounding error introduced when converting postsript to
>>> raster image?  I tried compiling this regtest with these two commits
>>> and the resulting pdfs were the same.
>> 
>> Don't reckon so.  I've now taken PDFs from Lilypond built from master and
>> from my earlier commit, and they're different - master has moved the system
>> start bar, such that the barlines protrude over it.  See attached images,
>> created by rasterizing the PDFs at 1200 dpi and then downsizing them.
> 
> Indeed, you're right.  My master wasn't up to date - my bad.
> The faulty commit is either 6e4e69f2735a764eab2e6f70f83546461da0203b
> or 8a35d5d2e856fa07c42c626ae7966c69dde61fdb (both from Mike).
> 
> best,
> Janek

Created tracker issue 3277 to fix that.  The problem is that there are some 
grobs that position themselves on an empty set of grobs. 
8a35d5d2e856fa07c42c626ae7966c69dde61fdb ends this practice, but there are some 
residual grobs whose functioning were built on this behavior like 
StanzaNumbers.  I created a function to deal with this case.  Apparently 
SystemStartBars are the same, but as they don't show up in regtests this can't 
be seen.  So, I just used the StanzaNumber function for SystemStartBars 
(renaming the function to make it more general) and it works like a charm.

Cheers,
MS