Regtest Issues

From LilyPond Wiki

[edit] Issues for regtests, not documentation

[edit] General note

do we want to get rid of the

\paper { ragged-right = ##t }

everywhere?

[edit] alignment-order.ly

-> line 25: is the TextScript 'padding override still needed? Looks like a pre 2.8 thing to me...

Looking at the snippet in the 2.6 docs, it doesn't look as if it's ever been necessary; perhaps it was added to test something else.NeilP 23:12, 9 March 2010 (UTC)

This looks to me like a defect in the regtest, not the documentation? I thought this page was for features that are only documented in the regtests - so this needs to be moved to another bug tracker? --PhilHolmes 16:42, 19 May 2010 (UTC)

[edit] alignment-vertical-manual-setting.ly

-> is \book needed? If so, why? Ditto for NR4.3.7 and 4.4.3.

Probably not, since there's no explicit positioning of the systems (unlike page-layout-manual-position.ly or NR 4.4.3 where Y-offset is used).NeilP 23:12, 9 March 2010 (UTC)

Ditto my comment above. --PhilHolmes 16:42, 19 May 2010 (UTC)

[edit] fill-line-test.ly

-> Why the commented `abcdefgh...' lines? -> Do we still need \book here? This looks like a pre-2.10 workaround to me.

This is a potential bug in the regtest, not a problem with documentation. --PhilHolmes 16:53, 21 May 2010 (UTC)

[edit] fingering-cross-staff.ly

-> bounding box problem with the lower stem/beam.

This looks like a bug in either LilyPond or the regtest, not a documentation problem. All the other entries under F seem to fall into this camp, so I've done no snippets for them. --PhilHolmes 16:53, 21 May 2010 (UTC)

[edit] font-* regtests:

do we really want to use "Time New Roman" everywhere?

[edit] font-name-font-size.ly

-> The first and the second word do not look exactly the same (wrt kerning, etc.). Should they? If so, the regtest description should state it.

[edit] font-name.ly

-> nitpick: the "Rest in LuxiMono" text could be more appropriately named "MMRestText in LuxiMono" (since the Rest grob itself isn't affected).

[edit] fret-boards.ly

states the "the results will be best"; however the first and second diagram look exactly the same. Should they? If so, the regtest description should mention it.

[edit] general-scheme-bindings.ly

-> Do we want a visible output?

[edit] generic-output-property.ly

-> tastes like a blast from the past. I can't determine whether extra-offset still does anything in this snippet (nor *what* it should do, for that matter). Could the description be more specific?

This is basically the regression test for 'extra-offset, but it's complicated slightly by the change in override syntax (it's an old test which has been run through convert-ly to retain compatibility with the old \property syntax).

It'd make sense to split this into two separate tests: one for explicit \overrides with 'extra-offset and this one updated slightly, verifying that the conversion code for the old syntax works properly (though if it breaks, it won't show up anyway, since 'extra-offset doesn't change bounding boxes.) --NeilP 21:55, 10 June 2010 (UTC)

[edit] glissando-broken.ly, glissando-no-break.ly

-> It seems that the unbreakable-by-default behavior was decided when we used to make broken glissandos look ugly. Since it's no longer the case, perhaps we should rethink this choice?

[edit] glissando.ly

-> the description mentions a limitation of the engraver, but I fail to see where said limitation is demonstrated in the regtest. Perhaps it's just me...

[edit] grace-staff-length.ly

-> input/trip.ly doesn't seem to exist anymore. Update descr.?

[edit] lyric-hyphen.ly

-> This looks a mess. The main feature lines are commented out and the aim of the regtest isn't clear.

If it can be sorted out, I'd do a snippet. --PhilHolmes 17:46, 22 May 2010 (UTC)

Are you looking at a different test? I can't see any issues with this. --NeilP 21:24, 10 June 2010 (UTC)

Well it seems to me that 'half' the example is missing compared to the lilypond ly file. That is the 'x - - y' part is not showing on the reg tests. Is that significant?:Jlowe 15:48, 14 July 2010 (UTC)

Not really. It just reflects the history of the snippet: when it was added (in 1999), slurs didn't generate automatic melismata. That was still the case up to at least version 1.8: lyric-hyphen.ly --NeilP 20:59, 30 July 2010 (UTC)

There is nothing in this regtest that needs to be added to the manuals. Both the LM and NR contain adequate documentation of hyphens in lyrics. But the regtest itself should be tidied up. All it has to do is demonstrate that "--" works correctly in lyrics, so the easiest way of restoring the x -- y part is to delete the two slurs. TrevorD 23:04, 1 August 2010 (UTC)

[edit] pedal-ped.ly

-> Uses \override Staff.SustainPedal #'padding = #-2 which no longer has any effect. Would also be better if all the possible glyphs were used.

See my snippet with all glyphs. --PhilHolmes 17:45, 23 May 2010 (UTC)

[edit] Fixed regtest issues

[edit] backend-svg.ly

-> texidoc description is not printed in HTML output

The entire regtest is commented out, so that's why you don't see the texidoc description. We don't want to uncomment the regtest yet, but I'll reenable it once we set up SVG regtest infrastructure. Pnorcks 07:02, 7 March 2010 (UTC)

Silly me! Since the online source code isn't syntax-highlighted, I missed the big multi-line comment! VVillenave 17:17, 7 March 2010 (UTC) 17:06, 7 March 2010 (UTC)

[edit] beam-beat-grouping.ly

-> \setBeatGrouping is undocumented.

CS -- \setBeatGrouping is deprecated. In current development, BeatGrouping is not used to determine beaming.

[edit] metronome-text.ly

-> The description has a bug - it refers to hideTempoNote and should refer to tempoHideNote --PhilHolmes 17:57, 22 May 2010 (UTC)

Thanks, fixed. --NeilP 21:42, 10 June 2010 (UTC)