On 11/6/2012 5:22 PM, Otared Kavian wrote:
Dear Dalyoung,
I can confirm that there is still a serious issue with the way mkiv handles the bounds of integral signs, and this has been reported several times by Aditya, myself and others. In particular the latest beta (version 2012.11.06 09:56) handles the bounds somwhat inconsistently, depending on the fonts used: • with the default LM fonts, the bounds of the integral sign are correct in displayed maths, but are too close whith inline maths; • with the default xits fonts, the bounds of the integral sign are too far away in displayed maths, but are a little bit far whith inline maths.
Here are the source file for my testing and the resulting PDF file. PLease typeset the file both with mkii and mkiv to see the differences. Hans did some changes recently, which resulted in a better behaviour of the bug, but still the problem is not completely solved.
The problem with integrals is that we need to support new fonts as well as old ones. One issue is that the width of an integral sign is not the real with but has some interference with italic corrections. Also, older versions of luatex has issues with this and for that context has some mechanisms that could deal with it. The current version of luatex has improved math so those hacks can go away. Also, math fonts have been adapted to the latest interpretations of the open type math specification (and those that don't ... well, lets forget about using them). In due time all math fonts will be real otf fonts, but I like to keep the virtual variant(s) around (as I've too much time invested in it, they server as example of defining virtual fonts, and as benchmark for traditional fonts). At the same time we need to deal with italic corrections, in text, where open type does not have such a property, as well as in math, where fonts can have italic corrections but (as with more cases) the implementation of them can be limited. Anyhow, I now trust luatex do some of that better than before (let's assume recent engines and fonts). However, some automagic italic correction between italic shapes and non italic shapes as well as math / text boundaries are still dealt with. Yet another issue is that we have double and triple integrals so from now on that's also done differently for the virtual fonts (read: more mkiv like) as we had already the mechanisms for that in place. So ... another attempt. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------