On 22-10-2012 21:41, Otared Kavian wrote:
On 22 oct. 2012, at 18:47, Marco Pessotto
wrote: […] Should the attached (compiled with the latest beta, as fixed by Hans) be used as referenced?
Hi Marco and Hans,
Thanks to everyone for your attention, and many thanks to Hans for having tried to fix the bug in mkiv. Unfortunately the bug is still there, as it can be seen from the comparison of the two PDF files I am sending with this message: they are both the result of typesetting the same source file I sent you. The one typeset with mkii has the correct output, while the one with mkiv is not correct (the spacing of the bounds of the integrals and the spacing between the integrals when they are invoked with \iint or \iiint are incorrect). (I tried this with ConTeXt version 2012-10-22, 11:01). The « reference » output should be the file « math-integral-mkii-2012-10-22.pdf »
Keep in mind that mkiv and mkii will never produce the same results (e.g.: more granularity in font metrics in mkiv, less juggling with combined shapes in mkiv, etc). SO, it can never be the pixel wise reference. The issue with italic correction is complex. First of all italic corrections are a tex thing while in otf advanced kerning mechanism is used. This means that the engine has to deal with traditional as well ad new mechanisms. In the process of developing luatex and mkiv, we went through several phases, one is that the italic correction had to be added to the width (a handicap is that luatex has to make some guesses, and the mkiv vf is a wide font on the one hand but still a traditional one). Anyhow, already quite a while ago, mkiv got two mechanisms to deal with italic correction, one for text and one for math (otf text fonts have no italic correction). Both mechanism can be configured to adapt themselves to the situation. For math we have 4 variants. \setupmathematics[italics=1] % fontitalics \setupmathematics[italics=2] % fontdata \setupmathematics[italics=3] % quad based \setupmathematics[italics=4] % combination of 1 and 3 \setupmathematics[italics=none] Depending on available italic info or the lack of it, correction is applied. Of course this can be improved. Once they are complete, we can make the new gyre math fonts default and some of the problems will go away. The virtuals will then be kept as an example of defining virtual fonts. Concerning the issue: i removed adding the italic correction to the width (something that was needed in a previous version of luatex). 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 -----------------------------------------------------------------
Hans Hagen
Keep in mind that mkiv and mkii will never produce the same results (e.g.: more granularity in font metrics in mkiv, less juggling with combined shapes in mkiv, etc). SO, it can never be the pixel wise reference.
I'm well aware of this, but having something at least similar which can be compared easily is better than nothing. Once you say the issue can be considered fixed, I'll replace the reference with the new, good one. Having a failing test, reminding that the issue is open, makes sense to me. Best wishes -- Marco
participants (2)
-
Hans Hagen
-
Marco Pessotto