Any news on this? If it's just one constant whose value I would have to increase in my private copy of the code, before recompling, for testing purposes, which one would that be?
Peter On Wed, Mar 16, 2016 at 12:37 AM, Zdenek Wagner <zdenek.wag...@gmail.com> wrote: > 2016-03-16 0:06 GMT+01:00 Jonathan Kew <jfkth...@gmail.com>: >> >> On 15/3/16 18:04, Doug McKenna wrote: >>> >>> There could be some subtle problems that simply changing the character >>> count constant causes. >>> >>> In particular, the allocation size of a "whatsit" language node might >>> also need changing, which would require adjusting other code in the core >>> engine that assumes a default small size for that language node sub-type of >>> a "whatsit". >>> >>> Or not. I can't tell from the TeX source what the bit sizes of these >>> node fields are. But if they're too small to fit a pair of enhanced >>> character count limits for hyphenation, there will likely be bugs elsewhere >>> due to truncation or wraparound in the arithmetic. >>> >>> FWIW, >>> >>> Doug McKenna >>> >> >> >> AFAICS, this would only become an issue if we allow \lefthyphenmin and >> \righthyphenmin to be given larger values; currently, they're limited to the >> range 0..63 (actually, it's possible to set the parameters to larger values, >> but they'll be clamped when stored in a language node). >> >> But I don't think that's needed here; surely there's no realistic use-case >> that requires setting the *-min values greater than 63, even when >> hyphenating 1000-letter Sanskrit words. > > > The contrary is true. These long strings are not single words, several words > are glued together. For instance, Bhagavdgita contains श्रीभगवानुवाच which > are in fact three words श्री भगवान् उवाच. It makes no sense to increase > values of \lefthyphenmin and \righthyphenmin. >> >> >> >> JK >> >> > > > Zdeněk Wagner > http://ttsm.icpf.cas.cz/team/wagner.shtml > http://icebearsoft.euweb.cz > > >> >> >> -------------------------------------------------- >> Subscriptions, Archive, and List information, etc.: >> http://tug.org/mailman/listinfo/xetex > > > > > > -------------------------------------------------- > Subscriptions, Archive, and List information, etc.: > http://tug.org/mailman/listinfo/xetex > -------------------------------------------------- Subscriptions, Archive, and List information, etc.: http://tug.org/mailman/listinfo/xetex
