On 6/7/2021 4:32 PM, Steffen Wolfrum wrote:
Hi Hans,
is there some kind of randomizer involved in line-breaking?
I have a 458 pages project, a book of humanities.
When I immediately typeset (engine=luatex) it a second time, the resulting PDF differs on 14 pages, diffpdf-compared to the original PDF when I typeset it a third time, the resulting PDF differs on 7 pages, when I typeset it a forth time, the resulting PDF differs on 12 pages, when I typeset it a fifth time, the resulting PDF differs on 0 pages.
"differs" means that there a paragraphs (mostly footnotes, rarely in the main text) that show different line-breaking: This doesn't mean wrong hyphenation, just *different* hyphenation, different each time.
As you can imagine, I can't provide a minimal example for this, only the entire project-folder and the entire context-standalone.
Not that I know of ... but you can run wiht --keeptuc and see if there are differences in the tuc file that create some oscillation. Also, as you mention luatex: there have been no (nor will be) fundamental changes to the engine because that would harm 'long term stability' as well as not play well with macro packages that depend 'no changes' (otherwise an update in engine for an existing setup e.g. older texlive, could fail). So, a change in hyphenation only can happen when there's is something different in the node lists (something that influences for instance what is seen as boundary of a character). When lua is involved a difference can be in hashing (which is random per run) but that seems unlikely here. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------