| <<<Back 1 day (to 2013/10/10) | 2013/10/11 |
mvrhel_laptop | hmm how did 694704 get dropped on my lap | 04:55.34 |
| this will take me a little time. I wonder if ray might be better at taking a look at this | 04:56.32 |
| will talk with him about it in the morning | 04:56.42 |
henrys | mvrhel_laptop: I gave it to you because we shouldn't see up the halftone at all, then the other part can go to ray | 04:59.11 |
mvrhel_laptop | the halftone call is coming from the interpreter | 04:59.36 |
henrys | ah so we'll let kens have a look at it first? | 05:00.20 |
mvrhel_laptop | I will see if I can figure out why | 05:00.24 |
| bascially it is the zsetscreen call | 05:00.37 |
henrys | in gs_init.ps around line 395 is where it should be setup | 05:03.24 |
mvrhel_laptop | comments says | 05:03.27 |
| % Redefine setscreen and setcolorscreen to recognize halftone dictionaries, | 05:03.31 |
| % and to insert the Frequency and Angle into Type 1 halftones, per | 05:03.33 |
| % Adobe TN 5085. | 05:03.34 |
| oops | 05:03.53 |
| wrong file | 05:03.55 |
| nevermind above | 05:04.25 |
| looks like setdefaulthalftone does it | 05:07.48 |
henrys | I know I use this in PCL - gx_device_must_halftone(dev) and don't call any set halftone stuff if it is false. I know it worked at some point but haven't looked recently. | 05:09.59 |
mvrhel_laptop | right. | 05:11.39 |
| to me, this is one of those weird interpreter things like how pdfwrite device is always appearing even when I am doing a contone device | 05:12.18 |
| I don't understand why | 05:12.23 |
| and the interpreter being in PS makes it hard for me to figure out | 05:12.37 |
henrys | what initially sets pgs->halftone the graphics state member? | 05:14.37 |
| it seems to start out with "ht_type_none" so there is precedent for not having a halftone ;-) | 05:17.35 |
moquette31 | hi everyone | 05:18.12 |
henrys | moquette31: hi | 05:19.48 |
moquette31 | I am trying to port GS to Android using the NDK/JNI | 05:20.43 |
| i spend a week on it already | 05:20.51 |
mvrhel_laptop | henrys: yes. after looking at the interpreter code, I do fear that ken or chris would be much faster at figuring out why we are installing a screen for a contone CMYK devcie | 05:20.52 |
| or how to avoid it | 05:21.03 |
moquette31 | and I made many progress | 05:21.07 |
henrys | okay I'll give it to kens | 05:21.26 |
moquette31 | but still struggling with dependencies | 05:22.02 |
| I followed the Cross Compile instruction in GS Make manual (6.6) | 05:22.22 |
| and faced some issues | 05:22.35 |
mvrhel_laptop | done for the night | 05:24.25 |
moquette31 | first the command (on machine Z) : make obj/arch.h obj/genconf obj/echogs give me errors | 05:24.33 |
henrys | good night mvrhel_laptop | 05:24.37 |
moquette31 | make: *** No rule to make target `obj/genconf'. Stop. | 05:25.02 |
| I am on 9.10 version (latest) | 05:25.10 |
| so I launched a full make so in order to compile the genconf and the echogs | 05:25.51 |
| then edited the arch.h to fit the arm arch I am on | 05:26.10 |
| and finally the make CC=: CCLD=: command | 05:26.24 |
| but as I made a full make so before I think it does not recompile everything according to my arm arch.h | 05:26.49 |
| am I right ? | 05:26.53 |
| I am talking about this procedure : http://ghostscript.com/doc/9.10/Make.htm#Cross-compiling | 05:27.40 |
| can somebody help me on this GS cross compiling ? | 05:29.33 |
henrys | I would talk to Robin_Watts he'll be in a few hours - UK business hours. | 05:30.27 |
moquette31 | another point is : procedure tells to use the ldt.tr file to know the list of source file we need to cross compile, but doing that I have many "undefined reference" when linking object | 05:30.53 |
| henrys -> should I ask all this to Robin_Watts ? | 05:31.49 |
henrys | undefined references which you can find in the gs code base or outside like c library stuff? | 05:32.01 |
moquette31 | in the GS code base, but in c file not listed by ldt.tr | 05:32.23 |
| so not compiled | 05:32.30 |
henrys | moquette31: yes, although we haven't spent any time porting to android and point folks to mupdf for android, why do you need gs? | 05:33.03 |
sebras2 | moquette31: just because I'm curious: why are you trying to port ghostscript to Android? :) | 05:33.12 |
moquette31 | i have to precise that I want a very light GS, so I activated only one DEVICE and disabled everything possible (freetype, cups, ...) | 05:33.26 |
| I need to rasterize a pdf (generated in the Android App) and create the PCL code (from hpijs server, also ported in my Android App) in order to generate the PCL code I should sent to a specific printer | 05:34.57 |
| my pdf, just contain graphic jpeg or png, (no text, no tiff) | 05:35.22 |
| and my printer is a HP one | 05:35.31 |
| so need the ijs/hpijs stuff to create the pcl3 code | 05:35.43 |
| before I was using a port of GS in Openwrt on a small MIPS proc (400Mhz, 16MB ram) | 05:36.19 |
| but it take 60s to rasterize and create the PCL code | 05:36.34 |
henrys | moquette31: let's keep it simple and not remove any devices - does the cross compile work if you do that? | 05:36.37 |
moquette31 | from the pdf generated by my Android App | 05:36.47 |
| so I want to move gs/hpijs into my Android App | 05:36.58 |
| even with the full device list, i didn't finish dependencies problems | 05:37.43 |
| because except libz that is included in my toolchain | 05:38.02 |
henrys | moquette31: you may also put an enhancement request into bugs.ghostscript.com, we might make the problem bountiable - we pay someone a bounty to work on it. | 05:38.37 |
moquette31 | i ported libpng, libjpeg, libtiff, libexpat, freetype, libijs, libhpijs, libcups, libcupsimage liblcms2 | 05:38.56 |
| ok good point I am doing so | 05:39.31 |
henrys | moquette31: and you can talk to Robin_Watts about it but he is terribly busy with other stuff and may not be able to give it much attention. Good luck. | 05:40.13 |
| good night sebras2 | 05:40.23 |
moquette31 | ok | 05:40.31 |
| thanks | 05:40.33 |
| good night | 05:40.37 |
| waiting for an hypothetic talk with Robin, maybe somebody can explain me more about GS structure and light compilation : | 05:42.00 |
| can I get ride of freetype dependency ? | 05:42.02 |
| in configure script there is an option --disable-freetype | 05:43.08 |
| does it really allow to not have freetype dependencies ? | 05:43.26 |
| this is for exemple a question I am wondering... | 05:43.44 |
sebras2 | henrys: good night. | 05:46.03 |
| moquette31: just for the record -- I know very little about gs so I'm not able to help you. I think Robin might wake up in an hour or two. | 05:48.21 |
moquette31 | sebras2: ok, thanks | 05:50.41 |
| i am waiting to Robin so ;) | 05:50.52 |
| I have another idea | 05:57.43 |
| MuPDF is PDF viewer ( reader) | 05:57.59 |
| i didn;t dig into its code | 05:58.11 |
| but don;t you think i could be possible to use it (as it already have a android port) to interprete my pdf, and giving some input to hpijs as GS should do... | 05:58.57 |
| my point is that GS is very complex and heavily multi platform, so it is more and more difficult to port | 05:59.37 |
| bu I really need GS to convert the PDF into raster that hpijs should need to create PCL code, right ? | 06:00.12 |
| So maybe I can use MuPDF that is much lighter | 06:00.28 |
sebras2 | moquette31: there is an source/fitz/output-pcl.c in mupdf that suggests that PCL output may be possible. I don't know if it is as feature complete as you require or how well it works from inside Android. But maybe it is worthwhile looking into. | 06:02.45 |
moquette31 | yessss, thanks you | 06:04.58 |
| i really want to use hpijs server as it should have the best support for my hp printer, and I want the photo quality/definition | 06:05.39 |
| as described here only the hpijs driver should provide it : https://www.openprinting.org/printer/HP/HP-DeskJet_5650 | 06:06.49 |
| but it is a good point for MuPDF | 06:08.23 |
| I will look into it right now | 06:08.33 |
| look like output-pcl.c is used only in mudraw, at least it is a good example of use | 07:15.38 |
| and in output-pcl.c code it is explained that the PCL code is generic, so does not fit perfectly to ANY pcl printer... | 07:16.07 |
| my plan now is to integrate HPIJS code into mupdf in order to support better my printer | 07:16.39 |
| as it is specific to my printer, I am not sure you will be interested by it, aren't you ? | 07:17.02 |
Robin_Watts | moquette31: Morning. | 07:52.00 |
| moquette31: The expert on the gs build system and makefiles is really chrisl, but I have done some cross compiling of gs for ARM targets before. | 08:02.03 |
| Not specifically on android though. | 08:02.16 |
chrisl | Hmm, I assumed he'd moved onto mupdf....... | 08:02.41 |
Robin_Watts | Personally, if I was having to port gs to android, I would *definitely* be using the existing makefiles and not trying to roll my own. The dependencies are hard. | 08:03.12 |
| Also, I would be using (at least initially) the libraries built into gs, not attempting to share the system ones. | 08:03.43 |
chrisl | Yeh, I would start getting a complete default build working, and then start cutting out devices and stuff | 08:03.54 |
| BTW, the cross compiling stuff in the docs is out of date | 08:04.34 |
Robin_Watts | As chrisl says, the cross compiling stuff in the docs is out of date. I think it's easier nowadays: Essentially, CCAUX should be set to the host C compiler and CC should be set to the cross compiler. | 08:05.02 |
| I have to pop out between 9:45 and 11:15 or so, and I'm not really here yet, so bear with me. (9:04 local time now) | 08:05.58 |
chrisl | You also use "TARGET_ARCH_FILE=" so you can roll your own arch.h instead of genarch making it for you | 08:06.37 |
Robin_Watts | If you decide to go down the mupdf route, I am better placed to give you pointers on that. | 08:07.05 |
sebras2 | tor7: good afternoon tor7! | 08:17.34 |
chrisl | Ugh, looks like I need to download Visual Studio 2013? I think I need more coffee........ | 08:35.48 |
moquette31 | hi Robin | 08:35.54 |
| i am reading your answers | 08:36.02 |
| thanks to revert | 08:36.07 |
| BTW, the cross compiling stuff in the docs is out of date -> haha I understand better now ;) | 08:37.01 |
| "TARGET_ARCH_FILE=" so you can roll your own arch.h instead of genarch making it for you -> good to know | 08:38.07 |
| If you decide to go down the mupdf route, I am better placed to give you pointers on that. -> yes definitively, I spend some hours around MuPDF and hpijs source code to understand better | 08:38.58 |
| MuPDF with its output-pcl.c option (used in mudraw only) could be very close to what I need | 08:39.29 |
| I just need to add support for more printer langague in output-pcl.c like to one used by my printer : PCL3GUI | 08:39.59 |
| i am working on that now | 08:40.11 |
| ;) | 08:40.13 |
| will let you know | 08:40.16 |
tor7 | Robin_Watts: fix fts_01_0106 LGTM | 10:36.24 |
| Robin_Watts: and zeniko's newest fix (gel->alen = 0) looks plausible as well | 10:37.04 |
Robin_Watts | tor7: Thanks, both pushed. | 11:26.49 |
| tor7: Documentation commit on robin/master. | 12:07.36 |
| I've tried to improve the stuff that mvrhel_laptop ran into yesterday. | 12:08.00 |
defthoms | Hello there, any MuPDF people/user here? | 12:13.53 |
kens | despite the lack of response, yes. It is lunch time so replies may be slow | 12:15.47 |
defthoms | oh yeah did not think of this, cheers | 12:22.49 |
| I'll come back later it is not urgent | 12:23.53 |
tor7 | Robin_Watts: looks good. | 12:30.53 |
Robin_Watts | ta | 13:05.51 |
| tor7: 2 more small things. Again, no hurry. | 13:30.59 |
tor7 | Robin_Watts: OMG! IT WORKS!!! | 13:33.59 |
| now I just have to flip what I store in the low bits and high bits... seems like there's a slight issue with asking nvidia to do the winding calculations in any but the lowes bits :( | 13:34.47 |
| Robin_Watts: both 2 look good. | 13:39.40 |
defhoms | Hi there, does anyone knows if it's possible to change the default resolution of MuPDF ? | 13:42.48 |
kens | Robin_Watts : tor7 ^^ | 13:44.46 |
sicos | hihi | 14:10.16 |
| hi | 14:10.21 |
| somebody here that knows how to make a fontmapping file for ghostscript on windows? | 14:14.43 |
kens | Its int eh documentation | 14:14.55 |
sicos | I know but there is nothing about windows | 14:16.10 |
mvrhel_laptop | kens: how did you stumble on the segv with the xps file? I wonder why it did not show up on the regression testing | 14:16.16 |
kens | sicos, its not winows-specific | 14:16.28 |
| mvrhel_laptop : while working on a different bug, I wanted to see the rendered output | 14:16.40 |
sicos | what name do i have to use for the mapping file? | 14:17.02 |
kens | mvrhel_laptop : it seems its a debug-only thing because we overwrite the freed poitner in debug mode only | 14:17.20 |
| sicos, anything you like | 14:17.32 |
| I think fontmap.gs works | 14:17.44 |
sicos | And how do I need to call that file? I use ghostscript from vb6 through it API | 14:18.07 |
kens | read the documentation | 14:18.22 |
sicos | can i just add it as an input file? | 14:18.24 |
mvrhel_laptop | oh I see | 14:18.32 |
kens | No, its a library file | 14:18.34 |
mvrhel_laptop | so you already fixed the crash issue then or is there something else I need to look at kens? | 14:18.49 |
kens | moment | 14:19.02 |
sicos | I have a pdf that gives this error | 14:19.05 |
| Can't find CID font "TimesNewRoman". Attempting to substitute CID font /Adobe-Identity for /TimesNewRoman, see doc/Use.htm#CIDFontSubstitution. | 14:19.06 |
chrisl | sicos: what version of GS are you using? | 14:19.27 |
kens | sicos, then you need to map it as a CIDFOnt, not a font | 14:19.31 |
sicos | 9.10 | 14:19.49 |
kens | did you read hte docs you wre pointed at ? | 14:19.50 |
sicos | yes | 14:19.57 |
kens | so did you create a cidfmap ? | 14:20.48 |
sicos | Yes, it was created automaticly by the installer | 14:21.06 |
kens | thenhj you cna modify that | 14:21.13 |
chrisl | sicos: you'll to setup a cidfmap file, and use the "-I" command line option to tell Ghostscript what directory to search for it | 14:21.23 |
sicos | uhm ok.. didn't know that.. tried the -I function with a fontmap file :-) | 14:22.36 |
mvrhel_laptop | kens: oh you did not do a commit to fix the mixed up free. did you want to do that or do you want me to do it? | 14:22.39 |
kens | mvrhel_laptop : no I didn't change the sources. Also, whenI render it the greay area fill is missing | 14:22.59 |
mvrhel_laptop | so there is a rendering bug too that I need to look at | 14:23.26 |
| ok | 14:23.27 |
kens | If you look at it with the MS XPS viewer or convert it to pdf, and comare the rendered output you'll see what I mean | 14:23.44 |
| I was using up to date sources as well, so its not something you recently fixed | 14:24.08 |
| I have little idea what the problem really is, the fact that pdfwrite gets it 'correct' makes me suspect the pdf14 compositor, but I cannot say for sure | 14:24.55 |
tor7 | defhoms: the -r flag sets the default resolution | 14:25.30 |
kens | mvrhel_laptop : I left the mis-ordered free for you though, I figured you oculd sort that out when you look at the rendering problem | 14:25.47 |
mvrhel_laptop | yes. ok thanks | 14:26.03 |
defhoms | i thought it was just changing the current document | 14:26.05 |
tor7 | defhoms: there is no config file for mupdf, if you want to change the hard-coded default you'll have to recompile | 14:26.39 |
| or create a shell/batch script that calls mupdf with the options you want | 14:26.57 |
defhoms | ok, thanks for the answer | 14:27.15 |
tor7 | defhoms: on linux we pick the resolution from the X11 server | 14:27.47 |
henrys | chrisl: can I take the disablefapi hackery out of pcl startup now? | 14:32.21 |
chrisl | henrys: I'll do it, if we're happy with the "maturity" of the code | 14:32.46 |
henrys | chrisl: thanks, I'm happy | 14:33.07 |
chrisl | I find it really annoying that PCL has 3 "main" functions, two of which do nothing but call another one :-( | 14:36.42 |
Robin_Watts | tor7: excellent news! | 14:36.59 |
chrisl | henrys: is there any sane reason we need main(), pl_main() *and* pl_main_aux()?? | 14:39.03 |
henrys | chrisl: Did kens add the aux with the PCL viewer? | 14:40.27 |
| for MS | 14:40.31 |
kens | I don't think so | 14:40.34 |
| But I guess its possible | 14:40.55 |
henrys | chrisl: I could figure it out but I don't remember the history | 14:41.00 |
chrisl | That would seem odd: main() contains one line, a call to pl_main(), and pl_main() contains one line, which is a call to pl_main_aux() | 14:41.35 |
henrys | chrisl: seems like it should be fixed with the reorg you are doing now ;-) | 14:42.11 |
chrisl | Ah yes, it does look like it was related to the display device - but it's pointless now..... | 14:42.47 |
kens | If its pointless, kill it :-) | 14:43.13 |
chrisl | henrys: I'm not changing any code (unless I have no choice) for the directory/build reorg | 14:43.24 |
| kens: I'll probably do so on Monday - one level of indirection is probably enough! | 14:44.08 |
Robin_Watts | mvrhel_laptop: I added stuff to docs/overview.txt this morning to hopefully make things clearer. | 14:46.52 |
mvrhel_laptop | Robin_Watts: ok I will review it | 14:48.18 |
| thanks | 14:48.20 |
defhoms | quit | 15:05.02 |
Robin_Watts | tor7: ping | 15:51.27 |
| Bah. Luminosity softmasks in SVG :( | 16:29.38 |
tor7 | Robin_Watts: pong. | 16:34.58 |
Robin_Watts | tor7: So, for the SVG device, when I meet a mask, I have to define it,and give it a name. | 16:35.56 |
| i.e. on a begin mask. | 16:36.10 |
| and on the end_mask, I then close the mask and then start a group that uses the mask I've just defined. | 16:36.32 |
| which means I need the name that I created in begin_mask in end_mask. | 16:36.51 |
| and I currently have no easy way to store such things. | 16:37.02 |
tor7 | oh. | 16:37.46 |
Robin_Watts | So... I've updated the scissor stack to not just have a rectangle for each level, but to have a void *user; that the device can store stuff in. | 16:37.50 |
| so the svg device stores an int in there in begin_mask, and can read it back in end_mask. | 16:38.20 |
| Does that sound reasonable ? | 16:38.31 |
tor7 | I'm not convinced it belongs in the null-device | 16:38.43 |
| a void* seems both generic and too limited | 16:38.53 |
sicos | hi again | 16:39.11 |
tor7 | I'd just put the mask_name_stack in the svg device | 16:39.18 |
Robin_Watts | tor7: wherever we hold it, it needs to be held in some sort of clip stack. | 16:39.38 |
| So either I need to copy the clip stack logic into the svg device (which kinda invalidates the point of having the clip stack in the null device)... | 16:40.25 |
| ... or I need to make the null devices clip stack be a bit smarter. | 16:40.39 |
| I could live with it being an int rather than a void *. | 16:41.02 |
sicos | I have a problem. When I concatenate 2 pdf files together with ghostscript the font in my second pdf gets replaced by some other font. The page looks ok. But when I select the text on the page and try to copy it to notepad I only get garbage.. | 16:41.28 |
tor7 | I think an int should be useful enough for most cases | 16:41.33 |
Robin_Watts | ok, and an int doesn't have issues about needing to be freed. | 16:41.52 |
tor7 | and it doesn't have the problem of compiler warnings when casting between pointer and int on architectures with different pointer and integer word sizes | 16:41.58 |
Robin_Watts | tor7: While I'm at it, I'm pondering adding a flags word to the clip stack. | 16:42.11 |
tor7 | containing? | 16:42.20 |
Robin_Watts | For now, the flag word would just indicate what sort of thing gave rise to that entry | 16:42.36 |
| (clip path/mask/clipped text etc) | 16:42.46 |
tor7 | right. image mask or path or text? | 16:42.47 |
| sounds reasonable enough | 16:42.53 |
| I'm going AFK sporadically, will check in now and then if there's anything else | 16:43.11 |
Robin_Watts | OK, next question... | 16:43.14 |
| if you have time... | 16:43.18 |
tor7 | (soon) | 16:43.20 |
| no rush | 16:43.30 |
Robin_Watts | luminance SMasks vs normal smasks. | 16:43.40 |
tor7 | messy :( | 16:43.48 |
Robin_Watts | normal smasks use the alpha, right ? | 16:43.51 |
tor7 | yeah, normal smasks discard the color and use only the opacity | 16:44.04 |
Robin_Watts | whereas luminance ones convert the group to grayscale and use that ? | 16:44.10 |
tor7 | and luminance smasks render to a mask, and don't forget the backdrop color. | 16:44.22 |
| I'd just render to an image mask if I wanted to retain my sanity | 16:44.33 |
Robin_Watts | You're suggesting that I output a bitmap with the smask rendering in for luminance ones? | 16:45.04 |
tor7 | yeah, the luminance one renders to a color image, initialized with a backdrop color, and then convert the grayscale value to alpha | 16:45.08 |
| yeah. or the opposite for regular ones, depending on what SVG does natively. | 16:45.31 |
| but the background color seeding of smasks is madness | 16:46.01 |
Robin_Watts | SVG seems to do masking using the alpha. | 16:46.05 |
| yeah, I'm just going to punt on difficult smask cases for now. I don't think we can really be expected to cope with that. | 16:46.24 |
tor7 | you *could* go to a lot of trouble by setting some sort of "i'm in alpha mode" in the svg device and convert all colors to alpha values | 16:46.37 |
Robin_Watts | tor7: That would mean fiddling images etc too. | 16:47.10 |
tor7 | yeah. it's messy alright. | 16:47.16 |
Robin_Watts | I think I'll skip it for now at least. | 16:47.39 |
| Thanks. | 16:47.50 |
tor7 | but it might be worthwhile to have some sort of way of dealing with it; we're going to need it for XPS output if we do that | 16:47.54 |
| or better yet, do it in the pdf interpreter :) | 16:48.05 |
Robin_Watts | outputting a rendered image as part of the output is never going to be a good solution. | 16:52.21 |
| oh, sorry, the greyscale version... | 16:52.31 |
| Hmm. Some sort of filter device? | 16:52.56 |
sicos | Hi | 17:06.19 |
| I got font substitution working.. my pdf look ok now. But when I select the text from a pdf that is substituted and past that in notepad or word I only get wingding like character. | 17:07.18 |
| Is this normal or is this an encoding problem? | 17:07.30 |
Robin_Watts | sounds like an encoding problem to me. | 17:09.43 |
| but you probably needs kens, and he's gone for the night. | 17:09.53 |
sicos | well ok.. a problem for another day then :-) | 17:11.33 |
| where is kens from.. europa? | 17:11.50 |
Robin_Watts | UK | 17:17.13 |
sicos | netherlands overhere | 17:19.07 |
Robin_Watts | tor7: another one on robin/master then. | 17:41.39 |
| ray_laptop: ping. | 18:11.20 |
| Is this Len thing rumbling on still? | 18:11.30 |
| Can I help? | 18:11.33 |
ray_laptop | Robin_Watts: got your text. It's probably something they did in integrating changes. I'm going to look at the diffs vs. the color sim (Alphard2) that I have that works. for obvious stuff | 18:18.55 |
| I _do_ wish these guys had a sccs we could acces | 18:19.17 |
| access | 18:19.21 |
| I know it is not the "selective transparency" since that isn't even enabled with a color printer | 18:19.59 |
| They've changed 56 files since the version I have that worked :-( | 18:21.08 |
Robin_Watts | ray_laptop: Joy. | 18:21.31 |
ray_laptop | what a mess of conditional compiles. :-( These guys left out a lot of changes because they were needed for the "selective transparency", but rather than just "zap" it as I had suggested (in a single place), they put #ifdef around all of the changes on the color product :-( | 18:41.05 |
| so they probably dropped one of the fixes that was needed | 18:41.36 |
| well, at least there are *only* 14 files conditional on monochrome mode | 18:42.58 |
| Robin_Watts: ~Joy | 18:43.19 |
Robin_Watts | urgh. | 18:43.53 |
| tor7: 2 on robin/master. Again, no hurry. | 18:53.14 |
Gigs- | Does anyone have a minute to help me understand a tint transform function? | 21:04.13 |
| http://paste.ubuntu.com/6224223/ | 21:04.14 |
| so what I understand, functiontype 0 means it's a "sampled function" | 21:05.12 |
| the length on the next line is just the stream length which for no apparent reason is stored as a reference | 21:05.33 |
| really what I need help with it the size/range/domain and how it applies to the sample stream | 21:06.12 |
| this is a DeviceCMYK transform btw | 21:07.37 |
| ok so domain is input and range is output | 21:18.40 |
| heh this is a really powerful 2d transfer mechanism | 21:23.04 |
| ok so I change the tint transform and GS tiffsep renders exactly the same thing :( | 21:41.30 |
| It does use deviceCMYK mappings right? | 21:41.55 |
| oh wtf this source file has multiple tint transforms for the same separations | 21:56.28 |
| ok I'm logging off, maybe I'll catch one of you guys some other time for some chat on the nuances | 22:01.17 |
tor7 | Robin_Watts: fz_device_container_stack_in_mask -> fz_device_container_stack_is_in_mask perhaps? | 23:27.16 |
| Forward 1 day (to 2013/10/12)>>> | |