On Nov 16, 2005, at 19:24, Andreas L Delmelle wrote:

On Nov 16, 2005, at 11:51, Luca Furini wrote:

<snip />
There are other properties with a "validity range" and a fallback value: column-count, initial-page-number, column-number, number- columns-repeated, number-columns-spanned, number-rows-spanned, hyphenation-{push, remain}-character-count; only hyphenation- ladder-count does not have a fallback value in 1.0, so maybe this was just an oversight. Note that the fallback value is different, in general, from the default value, as it is derived from the illegal value by rounding.

I guess we'll have to have a look at all these properties to see how the fallbacks are dealt with.
I'll look into this as soon as I find some spare time.

I took another look, and none of the mentioned properties (with the exception of column-number) perform a validity check/provide a fallback. Haven't checked all applicable FObj's yet to see if the errors are reported there...

On another note, there are a few other number-properties for which bogus-values are currently allowed: - orphans / widows: from the definition, one would say "positive integer" (?); any value allowed
- reference-orientation: only 7 possible values; any value is allowed.
- grouping-size: zero-or-positive; negatives are allowed
- maximum-repeats: zero-or-positive; negatives are allowed

Since grouping-size and maximum-repeats are applicable to only one or two types of FO, the fallback could easily be provided when the property is bound. For reference-orientation, I'm still doubting... The values are numeric, but supposed to be treated as a sort of enum. Probably best off with a tiny custom Maker to avoid code duplication, and catch the error as early as possible. For orphans and widows, the Rec doesn't explicitly state what to do with zero-or-negative values (neither does CSS) --maybe we could suffice with a warning message here (and a common-sense substitution with the initial value?)



Reply via email to