| | <<<Back 1 day (to 2018/03/19) | 20180320 |
| enometh | in pdf/pdf-xref.c the default handler should still be the pdf handler, so pdf files without a .pdf extension (or any extension) can be opened at all | 04:08.33 |
| | the best app was w3m which let you specify it: w3m -T application/pdf | 04:09.23 |
| | i think the old `pdf_recognize' should be reinstated as the recognizer for pdf_document_handler, and the recognizer mechanism should NOT be deleted as "unsued" | 04:10.25 |
| | (it's still there, ) | 04:11.25 |
| sebras | tor8 (for the logs): 10 lines of patches in the bottom two patches on sebras/master. | 04:53.22 |
| physkets | Hi! | 09:22.13 |
| | Are animations in PDFs supported? | 09:22.29 |
| kens | I believe the answer is 'no' | 09:23.01 |
| | If you mean does MuPDF run movies | 09:23.58 |
| physkets | kens: not movies... I mean JavaScript based animations | 09:25.00 |
| kens | Are we talking about pure Java or PDF here ? | 09:25.24 |
| physkets | I'm making a pdf through latex, using its 'animate' package: | 09:25.38 |
| | http://ctan.imsc.res.in/macros/latex/contrib/animate/animate.pdf | 09:25.47 |
| kens | OK so PDF | 09:25.51 |
| physkets | yep | 09:25.56 |
| kens | I know nothing about the animate package, what kind of annotation deos it embed ? | 09:26.16 |
| physkets | In that pdf, on page 17 in an animation of the sort I am describing. | 09:26.17 |
| | I'm not too sure... | 09:26.39 |
| kens | I suspect that is etiher a Media Clip or a movie annotation | 09:26.54 |
| physkets | https://github.com/agrahn/animate | 09:27.06 |
| kens | Yeah no offence, but I'm not going to waede through reams of someone else's code.... | 09:27.32 |
| | If you can make a simple one page file with nothing else on it I'll take a look | 09:27.56 |
| | However, I suspect the answer is still 'no' | 09:28.04 |
| physkets | https://www.ctan.org/tex-archive/macros/latex/contrib/animate/ | 09:28.23 |
| kens | Because both media clips and movie annotations require external viewers | 09:28.23 |
| | Yeah that's still a 17 paghhe file with lots of stuff in it, not a simple one page file | 09:28.58 |
| physkets | In the first link, look at page 17 | 09:29.27 |
| kens | Seriously, the only way I can tell how the animation is embedded is by opening up the PDF file, decompressing it and reading it | 09:30.15 |
| physkets | The last link above is a description of the macro | 09:30.17 |
| kens | This is a 3.6MB compressed file | 09:30.23 |
| physkets | oh! | 09:30.24 |
| | okay | 09:30.26 |
| | I'll make one | 09:30.30 |
| kens | That will hel pa lot thanks | 09:31.00 |
| sebras-mobile | kens: this package seems to be for animating graphs, but it doesn't explain how it is realized in the file. The ui-buttons are obviously javascript though. | 09:41.04 |
| kens | sebras : I can look at an example file but a 17-page file is too much for me :-) | 09:41.30 |
| | Doe MuPDF support rendering to the c anvas from JS ? | 09:42.06 |
| | The buttons appear to be standard PDF buttons | 09:45.01 |
| sebras-mobile | kens: I think it will, when we implement the widgets with javascript running the event handling. But right now I'm not sure how much of this works at all. | 09:45.06 |
| kens | Umm, one for tor8 then | 09:45.27 |
| | The buttons are all Widget annotations though, not JS | 09:45.51 |
| sebras-mobile | kens: my guess is that this is all implemented in javascript. I don't expect them to render it as a movie clip. | 09:46.11 |
| | But they might depend on javascript objects we haven't implemented yet. | 09:46.35 |
| kens | There's a Screen annotation | 09:46.44 |
| | With 4 JavaScritp objects | 09:46.50 |
| | Well, 2 twice | 09:46.59 |
| | OK the animation appears to be a Widget annotation | 09:47.43 |
| | Looks like its a self-rendered button | 09:48.30 |
| | Presumably the JS renders it | 09:48.37 |
| sebras-mobile | The buttons show up in mupdf on Android. But there is no animation. | 09:49.14 |
| kens | % /JS (app.focusRect=false;try{if\(a2_playing\){a2_pause\(\);}}catch\(e\){}) | 09:49.16 |
| | So the answwer is no, as I suspected | 09:50.05 |
| sebras-mobile | kens: mmm, it is that app.focusRect that I'm worried about. I don't think we've implemented that. | 09:50.06 |
| kens | GS renders the appearance, which is fair enough | 09:50.31 |
| physkets | kens: https://send.firefox.com/download/78dab04a93/#DEz-u0Ce7GdWB8FCDGPeDA | 10:14.46 |
| | ^this is supposed to be an animation over two pages, | 10:15.01 |
| | but it only displays one | 10:15.06 |
| kens | There's only one page there | 10:16.02 |
| | Two graphs though | 10:16.16 |
| | You shoudl probably read the discussion with me and sebras above | 10:16.30 |
| | This appears to be a 'ahem' clever reuse of AcroForms and the use of Java to drw a Button. | 10:17.13 |
| | The appearance of the button that is. | 10:17.22 |
| | Looks to me like the JS repeatedly redraws the button, leading to an animation | 10:17.46 |
| physkets | When I open it in Oular, it does ask me if I want to see the Forms in the document | 10:18.28 |
| kens | Yep | 10:18.34 |
| | Its not what Forms are intended for | 10:18.42 |
| physkets | although, after that, even if I say yes, it just gives me a grey block | 10:18.46 |
| kens | Not surprised | 10:18.53 |
| | Someone is being too clever by half, frankly | 10:19.09 |
| physkets | okular doesn't do Js yet | 10:19.10 |
| kens | I'd suggest very few viewers will display this as expected | 10:19.23 |
| physkets | But why, exactly? | 10:19.36 |
| | Becasue it does non-standard stuff? | 10:19.48 |
| kens | That's not how you are supposed to do media in PDF :-) | 10:19.49 |
| | AcroForms are supposed to be forms | 10:19.59 |
| | You can draw a custom button, and there's the ability to use JS for that. | 10:20.12 |
| | But you are also supposed to supply an Appearance for the benefit of those viewers that don't do JS | 10:20.28 |
| physkets | OKay, so is it that mupdf does not implement AcroForms? | 10:20.34 |
| kens | In fairness, this file does that | 10:20.36 |
| | physkets : no, it doesn't implement the bit of JS that's being used | 10:20.51 |
| | To allow you to draw on the button | 10:21.09 |
| | Like I said, you aren't really supposed to do animations like this | 10:21.20 |
| physkets | okay, and is that as intended, or a bug? | 10:21.30 |
| | I mean that fact that it does not implement that bit of Js | 10:21.48 |
| kens | What, the fact that MuPDF doesn't support that currently ? Well its not a bug, we know it doesn't. | 10:21.53 |
| | Or more accurately the MuPDF developers (which is not me) know it doesn't | 10:22.10 |
| physkets | Oh, by bug, I mean lack of the feature | 10:22.13 |
| kens | I've no idea if there's any intention to do so | 10:22.19 |
| physkets | ah, okay | 10:22.24 |
| | I really would've liked to be able to create animations without making .mp4 s | 10:23.28 |
| | But I guess I'll have to do that now | 10:23.41 |
| kens | Well that's the 'correct' way to do animations in PDF | 10:24.03 |
| | Though I don't know if MuPDF supports that either | 10:24.13 |
| physkets | ah, :( | 10:24.44 |
| | Well, I don't know about correct, but the 'animate' package al;lows you to do lots of simple things like have the animation run in a 'loop', and also in a 'palindrome', where it goes back and forth across the inages. | 10:25.56 |
| | These are a little difficult to do when creating movies | 10:26.10 |
| kens | Yes, but its resuing features intended fro one purpose for a totally different purpose. | 10:26.25 |
| physkets | So begin able to use this would be reall nice | 10:26.31 |
| | Is that such a bad thing? | 10:26.41 |
| kens | Well yes. | 10:26.46 |
| physkets | ok :( | 10:27.07 |
| kens | Because it opens you to the possibility of security problems, plus if you take advantage of it, and the package later restricts it, you are lost. | 10:27.15 |
| | Its like using alcohol to fuel your car, it may work, but its probably going to break emissions standards, and may blow up the engine | 10:27.48 |
| kens | can't think of a better analogy offhand | 10:28.14 |
| tor8 | mupdf doesn't support any form of animation at the moment, neither by using the Movie widget, nor by abusing javascript to animate stuff. | 10:30.22 |
| | our javascript support is very limited, there are many functions that just aren't implemented. you have to keep in mind the JS support was driven by form validation, and doesn't do much if you step outside of that comfort zone | 10:30.52 |
| | and that work was done several years ago and hasn't seen any work done on it since then | 10:31.08 |
| | velix: no, the exe files are built by the makefile and then run to generate code for the next step of the build. | 10:32.50 |
| | they are not linux binaries, unless you're trying to cross compile or something. they need to be built and able to run by the host compiler. | 10:33.09 |
| malc_ | tor8: was watching Lucas de Grot lecture, there, in passing, he mentions hinting, apparently he used it in calibri to achieve something otherwise impossible with standard curves and stuff... i guess the hinting is not all b/w there are uses for it apart from grit fitting | 10:34.18 |
| tor8 | the vista fonts were hinted specifically for use with cleartype | 10:34.50 |
| | the calibri, consolas, the set of fonts that start with c that were introduced with windows vista | 10:35.18 |
| physkets | okay, thanks for your time, kens Bye! | 10:36.39 |
| kens | NP | 10:36.44 |
| malc_ | tor8: yes sure, but from his words it seems that hinting is _essential_ to achieve the designers vision, regardless of the output medium | 10:38.11 |
| tor8 | malc_: huh. could be. | 10:38.36 |
| | malc_: have you seen what the DynaLab chinese fonts do with truetype hinting? | 10:38.48 |
| kens | But that's always true, regardless of the technology | 10:38.52 |
| | for example counter control hints in PostScript fonts | 10:39.02 |
| malc_ | tor8: heard horror stories.. :) we should ask our han friend here - sebras | 10:39.30 |
| kens | Dynalab fonts are a horror | 10:39.46 |
| | But they are not alone | 10:39.50 |
| | We've seen many fonts that do not render properly (or in some cases, at all) wihtout implementing TT hints. | 10:40.12 |
| | 'hinting' is really a misnomer in TrueType | 10:40.21 |
| malc_ | kens: what would be a "nomer"? | 10:40.39 |
| kens | minsomer = mis named | 10:40.53 |
| | Hinting in TrueType is really a programmatic way of deforming the glyph outlines | 10:41.29 |
| | You might call it deformations | 10:42.02 |
| | Or just 'weirdness' | 10:42.28 |
| malc_ | kens, tor8: https://www.youtube.com/watch?v=pWNwjNRd0tE&t=1950 | 10:49.32 |
| kens | Different stem widths is hinting, that;s reasonable | 10:50.45 |
| | I'm not sure I agree the delta hints improve the rendered character personally | 10:52.00 |
| | But anti-aliasing is always going to be a problem. | 10:52.23 |
| malc_ | all type should be fraktur anyhow | 10:52.23 |
| kens | He should use a bigger size, or a screen with more pixels, then he wouldn't have a problem | 10:53.04 |
| malc_ | that's what she said | 10:53.43 |
| | https://www.youtube.com/watch?v=pWNwjNRd0tE&t=2924 == https://xkcd.com/1015/ | 10:57.40 |
| kens | is getting bored wiht the Russian jokes.... | 10:57.45 |
| | TYor liked that xkcd as I recall | 10:58.11 |
| malc_ | kens: russian jokes = jokes involving russia/russians or jokes told/mediated by russians, if the latter, appologies... | 11:00.00 |
| kens | Too many jokes, not enough content in the lecture | 11:00.22 |
| kens | went back to work | 11:00.34 |
| malc_ | ah | 11:01.08 |
| | tor8: mupdf's html renderer unconditionally uses ligatures (at least fi) (if present) even in <pre> blocks [alignment is thus gone] | 12:32.24 |
| tor8 | sebras: do you think we hsould throw an error instead in mutool create if the %%Font and %%Image directives are missing arguments? | 13:12.25 |
| vtorri | tor8 have you seen my latest error with msys2 ? | 15:29.04 |
| sebras | tor8: fixed, now I throw instead. this results in an uncaught exception, but that is in line with how the tools normally work. | 16:03.29 |
| tor8 | sebras: yes. at least it will exit with a failure error code than fail mysteriously now. | 16:33.10 |
| sebras | tor8: ok. | 16:38.05 |
| Tamir_Evan | vtorri: To resolve that, ' $(PKCS7_LIB)' needs to be added to the end of $(MUVIEW_WIN32_EXE)'s rule, in the Makefile (currently line 421), similar to how it was done for $(MUVIEW_X11_EXE)'s rule (on line 397), probably. | 17:14.53 |
| vtorri | Tamir_Evan but the error mentions that it's an x11 app that is built, and not a windows one | 17:16.22 |
| Tamir_Evan | vtorri: From what you showed yesterday the error was happened to a file in the x11 subdirectory, where both the (old) x11 and win32 viewers code reside. Is there more to the error that you gave yesterday? | 17:20.56 |
| fredross-perry | currently gtting a build error with master code: | 17:41.19 |
| | platform/gl/gl-main.c:1677:16: error: use of undeclared identifier 'GLUT_ACTION_ON_WINDOW_CLOSE' | 17:41.21 |
| velix | tor8: I thought, mutools are a linux project and EXE is crosscompiled :D | 17:51.04 |
| | Can I make mupdf to view the object number? | 17:51.14 |
| sebras | velix: mutool can show object numbers, but mupdf does not. | 17:51.43 |
| velix | sebras: The problems is: I don't know where my target xobj is located ;) | 17:52.02 |
| | It would be nice to have such a feature. | 17:52.09 |
| sebras | velix: well, what I usually do is mutool my-doc.pdf t to see the trailer, then walk to the Root object mutool mu-doc.pdf t 42 or if is 42, then to the Pages object, then the Kids array, then the desired Page object, and finally the Contents array which contains graphical operators that utilize the Resource dictionary in the Page object. | 17:54.40 |
| | velix: if the page is complicated the last step might require quite a bit of searching. | 17:55.08 |
| | but sometimes you can glean some hints from the XObjects themselves present in the Resource dictionary. | 17:55.33 |
| velix | sebras: yeah okay. | 17:56.25 |
| | I just checked Acrobat. It also doesn't show the ID ;D | 17:56.41 |
| | Does mutool -z also touch images? I don't want to recompress images ;) | 18:06.49 |
| | Mutools seems to recompress my PDF and destroys the images. Interesting problem. | 18:42.43 |
| fredross-perry | Ah. If I use HAVE_GLUT=no, the problem is gone. | 19:26.12 |
| vtorri | it goes further | 19:31.01 |
| | next error : thirdparty/freeglut/src/fg_init.c:370:22: error: 'WidthValue' undeclared (first use in this function); did you mean 'GetGValue'? | 19:31.07 |
| | and other similar errors | 19:31.19 |
| | Forward 1 day (to 2018/03/21)>>> | |
Log of #mupdf at irc.freenode.net.