| <<<Back 1 day (to 2016/04/04) | 20160405 |
Fennec-kun | Anyone can give me a hand here? | 00:06.43 |
mvrhel_laptop | rayjj you there? | 05:37.38 |
Maori | Hi,there,I'm using the mupdf as a lib in a Project.Now I want to highlight some special words in the pdf. But I do not know how to make it. Can some one help ? | 08:15.11 |
kens | At the moment none of the MuPDF developers is online, if you wait for another hour or so there may be someone who can help | 08:16.58 |
Maori | Thanks,kens. | 08:17.39 |
sebras | tor8: I think there might be something wrong with how we do super/subscripting. | 10:59.34 |
| tor8: in epub. | 10:59.37 |
tor8 | sebras: okay. I am not surprised. | 10:59.50 |
sebras | tor8: today we do the vertical align based on node->em * constant where the constant is either 0.2 or -0.3 | 11:00.00 |
tor8 | I just grabbed some values out of a hat that looked good | 11:00.18 |
sebras | tor8: but consider the case where you have <span class="super-huge-font">hello<sup class="very-small-font">world</sup></span> then "world" will be offset, but in reference to its own ->em, not in reference to the preceeding words ->em. | 11:01.20 |
| tor8: I have been thinking about using the parent container's ->em instead, but then there might be a difference between that and the preceeding spans ->em. | 11:02.13 |
| tor8: so I think in this case we really ought to look at the preceeding word to determine how to offset super/subscript. | 11:02.42 |
| tor8: then we have text-top/text-bottom. these are more straightforward, but here it matters a lot what measurements we use for node->h. | 11:03.16 |
tor8 | sebras: right. | 11:03.36 |
sebras | tor8: I think that ascenders are not taken into account into that value, but I'm not really sure. | 11:03.51 |
tor8 | node->h = max(any inline images, max(node->em * style->line_height)) | 11:04.35 |
| which is also wrong, according to the html spec | 11:04.46 |
| but following the spec there makes all use of footnotes butt ugly | 11:04.56 |
| since a <sup>*</sup> for the footnote will add half a line of extra line spacing above it | 11:05.12 |
| sebras: what if you just take a percentage of the node->h rather than node->em for VA_SUB and VA_SUPER? | 11:06.24 |
sebras | actually this half a ling of extra line spacing is exactly what I see when experimenting. and it just seems wrong. to me it seems as if I would want the midline of the superscript to be at the top of the x-height of the preceeding word. | 11:07.44 |
| does that make sense? | 11:07.58 |
| tor8: but node->h is also related to the superscript node itself. I'd rather take the node->h of the preceeding word. | 11:08.28 |
| tor8: for VA_SUB/VA_SUPER. | 11:08.39 |
| tor8: and just guess something if there is no preceeding word. | 11:08.53 |
| btw, do we prevent newlines between a word and a superscript? we probably should. :) | 11:09.15 |
tor8 | sebras: that depends on our usual line breaking rules, that you wrote ;) | 11:09.57 |
| a<sup>b</sup> would not allow a line break | 11:10.07 |
| a <sup>b</sup> would | 11:10.18 |
| sebras: sorry, I meant the line->h if such a thing is available there | 11:10.58 |
| sebras: it looks like we have a line_h there | 11:11.21 |
| but then you have the question of nested superscripts | 11:11.30 |
| let me finish off this bidi stuff I'm working on and then I can dig into the specs for superscript | 11:12.02 |
sebras | tor8: no worries. | 11:16.57 |
tor8 | så, nu har jag fixat det jag tänkte fixa idag. | 11:58.50 |
| hur dags vill du komma f örbi? | 11:58.54 |
| (bah, wrong channel, sorry) | 11:59.03 |
| sebras: ^ | 11:59.09 |
sebras | switches channel | 12:01.26 |
| tor8: d'oh! | 14:08.04 |
| tor8: just spent half an hour on trying to find where the issue was when trying to render 'abc <img src="test.jpg"> xyz" and not seeing the "xyz" after the image... | 14:08.51 |
| tor8: the img "block" ate my text. :-P | 14:09.27 |
| I thought I had made a horrendous mistake in the line breaking algo or something. | 14:10.17 |
tor8 | sebras: lol :) | 14:12.33 |
sebras | learns xhtml the hard way. | 14:16.34 |
chrisl | I didn't think there was an easy way...... | 14:17.03 |
| Forward 1 day (to 2016/04/06)>>> | |