Re: Lilypond 2.17.26 coredumps on tuplet with only skips

[David Kastrup]

Rutger Hofman <address@hidden> writes:

> Good morning bug list,
>
> I get a core dump with the current development version(s) 2.17.24-1,
> 2.17.26-1, Ubuntu 12.04 64-bit.
>
> A very minimal example that triggers the bug is attached.
>
> many, many instructions. Empty tuplets are required sometimes to bind
> the Spanners to the correct time. Or, dynamics start somewhere after
> the start of a note, and the dynamic must be bound to a skip.
>
> This used to work up to 2.16 at least.
>
> Rutger Hofman
> Amsterdam
>
> \version "2.17.20"
>
> \score {
>     \tuplet 3/2 { s8 s8 s8 }
> }

Traceback starts out
#0  Item::break_status_dir (this=0x0) at item.cc:138
#1  0x0830fead in Tuplet_bracket::parallel_beam (address@hidden
    0x854d978, cols=..., address@hidden)
    at tuplet-bracket.cc:95
#2  0x08310209 in Tuplet_bracket::calc_cross_staff (smob=0xf6f3f318)
    at tuplet-bracket.cc:801
#3  0xf7efac10 in scm_dapply () from /usr/lib/libguile.so.17
#4  0xf7f01618 in scm_apply () from /usr/lib/libguile.so.17
#5  0xf7f02ed0 in scm_call_1 () from /usr/lib/libguile.so.17
#6  0x0813a589 in Grob::try_callback_on_alist (address@hidden, 
    address@hidden, address@hidden, 
    address@hidden) at grob-property.cc:232
#7  0x0813a786 in Grob::internal_get_property (address@hidden, 
    sym=0xf6a745c0) at grob-property.cc:189
#8  0x08080975 in Axis_group_interface::adjacent_pure_heights (smob=0xf6f3eb38)
    at axis-group-interface.cc:243
#9  0xf7efac10 in scm_dapply () from /usr/lib/libguile.so.17
#10 0xf7f01618 in scm_apply () from /usr/lib/libguile.so.17
#11 0xf7f02ed0 in scm_call_1 () from /usr/lib/libguile.so.17
#12 0x0813a589 in Grob::try_callback_on_alist (address@hidden, 
    address@hidden, address@hidden, 
    address@hidden) at grob-property.cc:232
#13 0x0813a786 in Grob::internal_get_property (address@hidden, 
    sym=0xf6a6e420) at grob-property.cc:189
#14 0x0807ddad in Axis_group_interface::part_of_line_pure_height (
    me=0x85cee70, begin=true, start=0, end=1) at axis-group-interface.cc:160
#15 0x0807e7b8 in begin_of_line_pure_height (start=0, me=0x85cee70)
    at axis-group-interface.cc:184
#16 sum_partial_pure_heights (end=2147483647, start=0, me=0x85cee70)
    at axis-group-interface.cc:143
#17 Axis_group_interface::relative_pure_height (address@hidden, 
    address@hidden, address@hidden)
    at axis-group-interface.cc:332
#18 0x0807ef85 in Axis_group_interface::pure_group_height (
    address@hidden, address@hidden, address@hidden)
    at axis-group-interface.cc:580
#19 0x0814f365 in Hara_kiri_group_spanner::pure_height (smob=0xf6f3eb38, 
    start_scm=0x2, end_scm=0xf678c740) at hara-kiri-group-spanner.cc:58

So it looks like Item::break_status_dir is called on a null pointer
Item.

-- 
David Kastrup