| <<<Back 1 day (to 2016/02/04) | 20160205 |
Robin_Watts | I may try it tomorrow to see how it works out. | 00:00.56 |
tor8 | can we put the java files in platform/java/com/artifex/mupdf/fitz without the intervening 'src' directory? | 00:01.29 |
Robin_Watts | We could. It wouldn't be my preference, but we could. | 00:02.12 |
| I did debate about using 'libsrc' and 'viewersrc' | 00:02.36 |
| or src/lib/ and src/viewer | 00:02.50 |
tor8 | getMetaData can use Document.META_FORMAT instead of "format" | 00:02.52 |
Robin_Watts | D'Oh. Will do that. Much nicer. | 00:03.02 |
| Could have a java Document.getFileFormat method that calls Document.getMetaData. | 00:03.36 |
tor8 | yeah, I don't see why not | 00:03.54 |
| the directory structure is deep enough as it is, annoying java | 00:05.51 |
| we could shorten it by using com.mupdf as the package... | 00:06.22 |
| sebras: that approach looks good. I'll have to review it in detail later when I'm not as tired. | 00:16.15 |
| Robin_Watts: anyway, I'd rather just keep pointer as protected than add setPointer. we're exposing access to it one way or the other. | 00:22.33 |
mvrhel_laptop | need to do a bit of stress testing, but I think I have pdfwrite using the image resource handler with the fz_hash table that is used with pdfcreate | 01:01.42 |
| what is nice, is that the resources are now shared across the document. | 01:01.57 |
| that is, previously if the same image occurred on two different pages I think it was getting inserted twice | 01:02.53 |
| that should not happen now | 01:03.00 |
marcosw | I need to take casper down again, this time for an hour or so. | 02:54.05 |
dalerank | question for mujs: how i can save js function with c-object and call it later? | 07:55.45 |
tor8 | dalerank: (for the logs) you can use the registry to save js objects for use with c. js_setregistry/js_getregistry. | 10:02.32 |
| the helper functions js_ref/js_unref will store the top js object in the registry and return a key for you to remember its location with | 10:05.08 |
RobinWattsLenovo | BT turned up to fit fibre broadband this morning. | 12:12.12 |
| Plugged a modem in and buggered off, saying it should be working within an hour. | 12:12.28 |
| The modem indeed has a connection to the exchange now, but the LAN port on it is fucked. | 12:12.44 |
| And BT won't replace broken modems. | 12:12.52 |
| So I've got to go buy and all in 1 modem/router. | 12:13.11 |
| Fuck BT. | 12:13.14 |
jogux | RobinWattsLenovo: wtf :( I thought they weren't due to stop replacing faulty modems until next year. | 12:18.03 |
chrisl | RobinWattsLenovo: Surely, if it was supplied broken, they should be obliged to replace it | 12:22.35 |
RobinWattsLenovo | chrisl: You'd think. | 12:22.48 |
| Supposedly they don't supply modems at all most of the time. | 12:23.00 |
chrisl | But, nevertheless, fitness for purpose etc etc | 12:23.21 |
RobinWattsLenovo | It's BT. A shower of aunts. | 12:23.24 |
chrisl | To be honest, my (limited) experience has been that their modems aren't great anyway, so you may be better off with your own | 12:24.15 |
jogux | RobinWattsLenovo: It depends what option the ISP order. The options that include modems are still available till 31st March 2016 ( https://www.openreach.co.uk/orpg/home/products/super-fastfibreaccess/cpeenablement/cpeenablement.do ) | 12:24.43 |
| but you would presume if the BT guy supplied a modem it's because the option that included a modem was ordered and hence BT are required to fix it. | 12:25.46 |
RobinWattsLenovo | jogux: I *could* kick up a fuss and try to get it sorted, but it would eat my time, and I would be without a solution until next week at least. | 12:26.28 |
jogux | mm | 12:26.44 |
jogux | would do like you're doing then send eclipse the bill for the new modem ;) | 12:26.55 |
RobinWattsLenovo | so I'm taking the 50 quid apology that Eclipse are offering, and I'm going to go buy a tp-link from banbury. | 12:27.03 |
jogux | ah, that's pretty reasonable :) | 12:27.15 |
| I seem to have been lucky with the BT engineers I've had. | 12:33.04 |
dalerank | hi all | 14:17.37 |
tor8 | hi | 14:22.20 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 14:22.20 |
tor8 | dalerank: to answer the question in your email, you have a C callback function that takes a JS function in an argument and want to save it for calling back later? | 14:24.21 |
| use js_ref to store the js function object in the registry and create a unique key for it that you can later use to retrieve the js object | 14:24.45 |
| js_copy(J, 1); /* copy argument 1 to the top of the stack */ | 14:25.17 |
| key = js_ref(J); /* store it in the registry and return the key */ | 14:25.30 |
| then you can js_getregistry(J, key); to retrieve the function for calling back into javascript | 14:25.46 |
dalerank | thank you tor8 | 14:54.24 |
Robin_Watts | Finally connected again. | 15:15.20 |
dalerank | tor8... after js_ploadfile() do i neec call specific name function? | 15:37.26 |
HenryStiles | super bowl madness fast approaching ... I don't think this is going to much better than the Seattle debacle for Denver... | 15:38.44 |
tor8 | dalerank: js_ploadfile will push a function object that you can call | 15:44.56 |
| dalerank: http://www.mujs.com/docs/examples see the "Callbacks from C to JS" example | 15:45.03 |
kens | HenryStiles : I was amused by the 32 mile (!) drone exclusion zone round the venue. Are they going to be using Eagles (presumably American Bald Eagles) to police that ? | 15:46.00 |
dalerank | you are very fast update docs :) | 15:46.04 |
HenryStiles | kens: an elite task force of rednecks has been assembled: http://www.nydailynews.com/news/national/colorado-town-vote-license-shoot-drones-article-1.1543030 | 15:51.23 |
kens | I thought Colorado had already voted to issue hunting licences for drones..... | 15:51.54 |
| Ah right, that report is from 2013 | 15:52.10 |
| You didn't see the news report about the Dutch using eagles to bring down drones ? | 15:52.31 |
| Its probably safer than letting hillbillies shoot them down, but perhaps not so much fun | 15:52.54 |
| Drat Robin_Watts has modified devices.htm, I wonder if Git will merge that..... | 15:54.29 |
HenryStiles | kens: amazing of course drones could outfitted to defend against that pretty easily | 15:54.30 |
kens | I'm doubtful an eagle vcould bring down a big drone, but it looks like fun | 15:55.46 |
| Boo merge conflict in devices.htm :-( | 15:56.06 |
Robin_Watts | kens: sorry. Want to me resolve it for you? | 15:56.28 |
kens | I'm doing it now, its the stuff on my branch where I rewrote teh high level devices docs | 15:57.03 |
HenryStiles | conflict resolution builds character ;-) | 15:57.08 |
kens | And pushed back up to my repo | 15:58.34 |
ferter_terter | hi guys | 16:21.45 |
| could you please help with a bit of a guidance? fz_bound_glyph function takes a fz_rect pointer as an input parameter. but what kind of rectangle does it return? i mean, after debugging my app, i'm sure, that it's not character's bbox. it is a lot bigger. so i'm wondering, if you could help me? | 16:25.26 |
tor8 | ferter_terter: it returns a rectangle that is guaranteed to contain the glyph. it may or may not be the smallest possible such rectangle. | 16:27.03 |
| ferter_terter: it would be prohibitively expensive to store the bounding rectangles for all of the characters in CJK fonts for instance. | 16:28.04 |
Robin_Watts | ferter_terter: Is your question "why are the returned bounds slightly bigger?" or "what the hell are the units on the box?" | 16:30.09 |
| i.e. if your expecting values like (0,0)->(2,2) is the problem that you're getting values like (-1,-1)->(3,3) or (-100,-100) -> (300, 400) ? | 16:31.16 |
tor8 | Robin_Watts: fz_bound_glyph takes the text transform matrix as an argument, so the units used should be plain (but then again, I'm blind to stuff like this) | 16:31.20 |
Robin_Watts | tor8: Pfft. I don't even guess unless I've got it in front of me in a debugger :) | 16:31.41 |
ferter_terter | well, right now it is, - "is there a way to get character's bbox - or the smallest possible" | 16:31.46 |
| ? | 16:31.48 |
Robin_Watts | ferter_terter: So you're after the "minimal bbox" for a glyph (or "accurate bbox"). | 16:32.16 |
ferter_terter | yea | 16:35.34 |
| that's exactly what i need in order to use a font atlas on rendering | 16:36.32 |
tor8 | ferter_terter: you can't get that bbox from mupdf, not trivially. you can extract the outline to a path and measure the bbox for that. | 16:36.45 |
Robin_Watts | tor8: Except not all glyphs can be extracted to paths. | 16:37.13 |
| (Type 3's) | 16:37.16 |
tor8 | Robin_Watts: yeah, there is that. | 16:37.26 |
Robin_Watts | We *could* add an fz_bound_glyph_accurately. | 16:37.52 |
ferter_terter | i don't care about type3 text | 16:38.18 |
tor8 | ferter_terter: Robin_Watts: fz_render_glyph_pixmap will render the glyph and return a fz_pixmap | 16:38.31 |
| ferter_terter: that should be enough to get it into a texture atlas | 16:38.59 |
Robin_Watts | but that fz_pixmap will be the same size as the fz_bound_glyph rectangle, I think. | 16:39.01 |
tor8 | Robin_Watts: no, it'll get whatever freetype decides it needs | 16:39.31 |
Robin_Watts | ah, ok. | 16:39.40 |
tor8 | Robin_Watts: or the fz_bound_glyph rectangle for type3's | 16:39.48 |
| ferter_terter: you should be aware though that fz_render_glyph_pixmap will render the glyph differently depending on the x and y translation in the matrix | 16:41.24 |
| to allow for different rendering based on fractional positions; not every instance of a character looks the same | 16:41.47 |
| ferter_terter: see the use of fz_subpixel_adjust in fz_render_glyph_bitmap for the quantization we use | 16:42.57 |
ferter_terter | ok, thank you | 16:52.42 |
rayjj | chrisl: It is sort of a pain (at least on slow machines) that gsromfs1.c rebuild every time. I've been thinking that we _might_ be able to create a dummy file that tells us that the gsromfs1.c file is up to date by having mkromfs run in a "test" mode that touches a marker file only if any files t would assemble were "newer" than the marker file (if it didn't exist it would be created). Then... | 21:41.00 |
| ...the gsromfs1.c would depend on the marker file | 21:41.01 |
| chrisl: if it _did_ discover a file that it determined was newer than the marker file, it would create the new gsromfs1.c | 22:30.34 |
| unfortunately, that doesn't get rid of the annoying "out of date" pop-up dialog from Visual Studio. I get tired of saying OK or Cancel | 22:31.50 |
| yeah, I know, Poor me :-/ | 22:32.42 |
rayjj | has to have something to gripe about | 22:32.50 |
| I guess it's better than having to maintain the project files manually | 22:33.14 |
| Forward 1 day (to 2016/02/06)>>> | |