Crash on init

[Bug Importer]

Cocinella crash on init. versions 0.95.12 0.95.13 and
0.95.14cvs

Hook earlyInitHook failed: 1
divide by zero
    while executing
"expr int( ($refHtmlSizeToLength($htmlSize) - $m)/$k +
0.9)"
    ("foreach" body line 2)
    invoked from within
"foreach htmlSize {1 2 3 4 5 6} {
  set fontSize2Points($htmlSize) [expr int(
($refHtmlSizeToLength($htmlSize) - $m)/$k + 0.9)]
     }"
    ("default" arm line 2)
    invoked from within
"switch -- $this(platform) {
 macintosh - macosx {
     array set fontSize2Points {1 10 2 12 3 14 4 18 5
24 6 36}
 }
 windows {
     array set fontSize..."
    (procedure "::CanvasUtils::CreateFontSizeMapping"
line 32)
    invoked from within
"::CanvasUtils::CreateFontSizeMapping"
    (procedure "::WB::Init" line 25)
    invoked from within
"::WB::Init"
    (procedure "::WB::EarlyInitHook" line 3)
    invoked from within
"::WB::EarlyInitHook"
    ("eval" body line 1)
    invoked from within
"eval $func $args"
divide by zero
    while executing
"expr int( ($refHtmlSizeToLength($htmlSize) - $m)/$k +
0.9)"
    ("foreach" body line 2)
    invoked from within
"foreach htmlSize {1 2 3 4 5 6} {
  set fontSize2Points($htmlSize) [expr int(
($refHtmlSizeToLength($htmlSize) - $m)/$k + 0.9)]
     }"
    ("default" arm line 2)
    invoked from within
"switch -- $this(platform) {
 macintosh - macosx {
     array set fontSize2Points {1 10 2 12 3 14 4 18 5
24 6 36}
 }
 windows {
     array set fontSize..."
    (procedure "::CanvasUtils::CreateFontSizeMapping"
line 32)
    invoked from within
"::CanvasUtils::CreateFontSizeMapping"
    (procedure "::WB::Init" line 25)
    invoked from within
"::WB::Init"
    (procedure "::WB::EarlyInitHook" line 3)
    invoked from within
"::WB::EarlyInitHook"
    ("eval" body line 1)
    invoked from within
"eval $func $args"

[Mats (matsben)]

The only way I see this could happen was if 'font measure Times 12 xxx'
returned zero which it shall never do. However, there is currently a
real nasty bug related to threads in the voip solution that can result
in very weird results like this. So far untraced.

[sander (s-devrieze)]

Any update on this?

[Mats (matsben)]

The likely problem here is that the 'font measure ...' returns a zero value for some unknown reason.
Since the iaxclient bug has been fixed after this bug was reported, there is a tiny probability
that it was the cause, but I don't think so. Why 'font measure ...' gives zero I have absolutely no idea.
With SVG we will switch from point to pixel font sizes why we don't rely on
some kind of add platform mapping.