| | <<<Back 1 day (to 2015/08/10) | 20150811 |
| mvrhel_laptop | night all | 06:49.59 |
| kens | night Michael | 06:50.25 |
| jimgray | Could I copy a figure from mupdf? | 09:06.14 |
| kens | WHat do you mean by 'figure' and copy in what sense ? | 09:06.35 |
| jimgray | picture in pdf file | 09:07.23 |
| | I can copy text by mouse right click | 09:07.44 |
| | text VS figure | 09:08.05 |
| | kens: that's possible? | 09:11.20 |
| sebras | jimgray: I don't think you can no. | 09:24.43 |
| jimgray | sebras: hmm. I also searched and fail. No such function. | 09:25.34 |
| sebras | jimgray: it may be possible to extract it if it is a raster image, like a jpeg file. that would involve running mutool show -be file.pdf objnbr | 09:26.02 |
| | jimgray: but finding the correct objnbr it a bit technical. unfortunately I have to head out now. | 09:26.30 |
| jimgray | sebras: thanks. I will look into 'mutool'. | 09:27.06 |
| tor8 | jimgray: if you're just trying to copy an image as a one-off thing, you could use 'mutool extract' to copy the images from a PDF into separate files | 09:28.09 |
| kens | Sorry network outage | 09:29.54 |
| | jimgray, if you mean is it possible in the current MuPDF viewers, then no it isn't (as far as I know). Is it techncally poossible using MuPDF, yes certainl;y | 09:30.24 |
| tor8 | for the x11/win32 viewer it's more of a limitation on the copy/paste system integration -- copying images is a lot harder than plain text | 09:31.01 |
| jimgray | tor8: I tried 'mutool extract file.pdf'. It generated many '.cff' file. | 09:31.44 |
| | kens: mutool extract? or other command? | 09:32.06 |
| tor8 | jimgray: those are the embedded fonts, you can ignore those | 09:32.07 |
| | mutool extract pulls out both fonts and images | 09:32.16 |
| jimgray | tor8: But I cannot find image generated. | 09:32.57 |
| tor8 | you could use 'mutool info <page number>' to list the resources on a page (and get their object numbers) | 09:33.30 |
| jogux | it's possible the image is actually vector artwork, mutool will only save out bitmap format images | 09:33.37 |
| tor8 | the object number is the X in (X 0 R) on the lines | 09:33.53 |
| jimgray | tor8: thx. info is useful. I cannot find image information from that. Maybe it is due to image format. | 09:34.55 |
| tor8 | jimgray: if mutool info doesn't show an image on that page, it's not a bitmap image | 09:35.34 |
| | or it's hidden in a pattern | 09:35.46 |
| kens | Or form ? | 09:35.51 |
| jimgray | tor8: kens Maybe in form. I can find many form objects. | 09:36.11 |
| kens | Perhaps the simplest solution is to share the PDF file | 09:36.13 |
| tor8 | kens: yeah, patterns or forms. not sure how mutool info displays form objects. | 09:36.15 |
| | mutool info is sebras baby | 09:36.26 |
| jimgray | just as a examaple | 09:36.46 |
| | https://www.f5.com/pdf/solution-profiles/tcp-optimization-wireless-networks-solution-profile.pdf | 09:37.13 |
| | kens: i will paste the pdf link | 09:37.16 |
| | kens: tor8 above is the pdf link as an example, you could find image in the 2nd page. | 09:37.39 |
| tor8 | jimgray: yeah, those images are all line art | 09:39.22 |
| kens | Indeed, no bitmap images at all | 09:40.40 |
| | plenty of forms, sahdings, transparency | 09:40.51 |
| jimgray | tor8: I can see a mobile phone, an ipad image. | 09:41.42 |
| tor8 | jimgray: your best bet if you want a PNG or JPEG from that file is to take a screenshot at the zoom level you want and crop it | 09:41.43 |
| | jimgray: yeah, all line art and vector gradients | 09:42.05 |
| kens | jimgray : it may look like an image, but it appears to be constructed using vectors | 09:42.21 |
| | as tor8 says shadings and fills | 09:42.35 |
| jimgray | kens: I see. They might be Visio drawing? I don't know what is shadings or fills or line art. | 09:43.04 |
| kens | I can confirm that by using Acrobat Pro whcih alloes me to selcet and delete the portions making up the pictures indfividually | 09:43.21 |
| jimgray | Any way, I know 'mutool extract/info' for next time try and screen capture. | 09:43.31 |
| | kens: tor8 I use screen capture. | 09:44.25 |
| sebras | tor8: I didn't know of mutool extract! but I do now. :) | 12:01.40 |
| henrys | meeting in 5 minutes | 14:27.26 |
| | or 3 | 14:27.30 |
| rayjj | or now | 14:30.51 |
| henrys | I was pausing for michael but.. | 14:31.04 |
| kens | he was up late | 14:31.14 |
| henrys | he was | 14:31.18 |
| | ears burning | 14:31.25 |
| mvrhel_laptop | uhoh | 14:31.42 |
| henrys | mvrhel_laptop: I was hoping you would give this a go this week: http://bugs.ghostscript.com/show_bug.cgi?id=694574, if it's not easy we can get rid of tests or something. | 14:32.23 |
| | tor8: thoughts about XFA? well stuff you can say on IRC. | 14:33.08 |
| mvrhel_laptop | henrys: ok I will have a look starting today | 14:33.15 |
| kens | henrys 'Its not part of PDF' | 14:33.24 |
| henrys | nor is pcl or postscript but we do those ;-) | 14:34.18 |
| kens | Gien that it involves reflow, its not a good plan to put it in GS IMO as that would imply printing the reflowed docuemnt without necessarily proofing it | 14:34.21 |
| henrys | no it's a no go for gs | 14:34.40 |
| kens | henrys if you want to add a new interpreter, that would be fine | 14:34.44 |
| henrys | I'm asking about mupdf | 14:34.47 |
| | marcosw: so the casper thing this week was rather dramatic. How about an alarm system? | 14:35.47 |
| | it's going to be one of those summer meetings ;-) | 14:36.55 |
| mvrhel_laptop | chirp chrip | 14:37.32 |
| henrys | fredross-perry: how the heck are you doing? | 14:37.47 |
| rayjj | henrys: marcosw said in the email he was going to get an automatic reboot going | 14:37.52 |
| fredross-perry | I am good, thnks. | 14:37.58 |
| | -thanks- | 14:38.04 |
| rayjj | henrys: (from marcosw's email) "'ll look into writing a script that checks if casper is up and if not force a restart of the instance when I get back next week." | 14:38.36 |
| tor8 | henrys: I'm going to need to update the linux viewer to do forms... | 14:38.39 |
| mvrhel_laptop | i need to add forms to gsview at some point... | 14:39.01 |
| fredross-perry | forms? | 14:39.13 |
| rayjj | as in form filling ? | 14:39.26 |
| fredross-perry | my question | 14:39.34 |
| rayjj | that would be *very* handy | 14:39.39 |
| tor8 | fredross-perry: rayjj: form filling in the pdf viewer, yes. | 14:39.42 |
| fredross-perry | hmmm. | 14:39.56 |
| tor8 | it barely works on win32, x11 is bust | 14:39.58 |
| henrys | mvrhel_laptop: could be a commercial option | 14:40.05 |
| rayjj | tor8: if we have it in gsview, do we need it in the 'sample' linux viewer ? | 14:40.15 |
| tor8 | so I figure a good first step would be to make that robust, then I can see about XFA details | 14:40.16 |
| mvrhel_laptop | yes maybe | 14:40.32 |
| rayjj | we want folks to have reason to migrate to gsview | 14:40.44 |
| tor8 | rayjj: it wouldn't hurt, for ease of forms development. I don't expect to have annotation creation in the sample viewers. | 14:41.11 |
| kens | Acrobat reader can do XFA forms I htnk, and its free | 14:41.17 |
| henrys | tor8: how does summatra do? | 14:41.20 |
| | have they added anything? | 14:41.34 |
| rayjj | kens: yeah, but not available on linux | 14:41.35 |
| tor8 | henrys: no idea | 14:41.36 |
| mvrhel_laptop | if AR does it free then we can't charge for it | 14:41.49 |
| rayjj | mvrhel_laptop: sure we can | 14:41.59 |
| kens | My feeling also, even if its not available on Linux | 14:42.02 |
| mvrhel_laptop | well we can... | 14:42.47 |
| | in any event I am going to have a look at adding it | 14:43.09 |
| henrys | I didn't really have a lot for the meeting anyway and chris and robin are out today. Anybody else have topics? | 14:43.25 |
| kens | PCL->PDF/A | 14:43.34 |
| | There's no simple way to do this, I need a new option whcih will only be understood by the PCL interpreter, does that seem reasonable ? | 14:43.59 |
| rayjj | kens: what would that option do ? | 14:44.26 |
| henrys | kens: sure | 14:44.27 |
| kens | Is it reasonable to limit the production of such PDFs to RGB ? | 14:44.44 |
| | rayjj it would need to pass in an OutputIntentProfile | 14:44.53 |
| rayjj | kens: or gray | 14:44.58 |
| kens | Or more accurately, teh locaton of a profile | 14:45.04 |
| | rayjj, no | 14:45.07 |
| | RGB or CMYK for PDF/A | 14:45.14 |
| rayjj | kens: I see. So a monochrome PCL turns into an RGB PDF/A | 14:45.37 |
| kens | No..... | 14:45.44 |
| | You don't need to specify an OutputIntent for Gray | 14:45.56 |
| | So Gray is unchanged | 14:46.08 |
| | RGB would be unchanged if going to an RGB PDF | 14:46.17 |
| | RGB would be converted if goint to a CMYK PDF | 14:46.29 |
| henrys | kens: rayjj is saying there is a lot of 1 bit raster in PCL, what happens with that? | 14:46.49 |
| mvrhel_laptop | do you have an option to do a gray output intent? just curious | 14:46.56 |
| kens | Its stays in 1-bit | 14:46.57 |
| | mvrhel_laptop : not really | 14:47.06 |
| mvrhel_laptop | ok | 14:47.14 |
| kens | I believe there can be only one output intent in PDF/A-1 | 14:47.22 |
| | And it has to be either RGB or CMYK | 14:47.32 |
| | You are allowed DeviceIndependent colours (ICCBased) | 14:47.47 |
| mvrhel_laptop | ok | 14:47.57 |
| henrys | kens: is it possible to set up a cluster regression test for this say comparing against regular PDF or would there be too many differences. | 14:47.59 |
| kens | Depends if we allow CMYK PDF/A-1 to be produced | 14:48.28 |
| henrys | is worried he's not considering everything. | 14:48.30 |
| kens | If not, then the differences should be nil or minor (I think) | 14:48.40 |
| | If we do permit it, then there would be many, many differences | 14:48.53 |
| | Because we would convert all the RGB to CMYK | 14:49.10 |
| henrys | let's not worry about CMYK and let the cluster tell us what's broken | 14:49.33 |
| | I'm sure it'll just be a small matter of programming ;-) | 14:49.55 |
| kens | We need to think about CYK | 14:50.07 |
| | CMYK* | 14:50.10 |
| henrys | for pcl? | 14:50.19 |
| kens | Shoudl we permit the production of a CMYK PDF/A-1 file from PCL, or limit it to RGB only | 14:50.25 |
| henrys | RBG only is fine | 14:50.35 |
| kens | Well that makes life much easier for me. | 14:50.44 |
| | Much of the configuraton is then fixed. | 14:50.51 |
| | Just need to supply an OutputIntent profile. | 14:51.04 |
| | henrys this would mean a change to the PCL interpreter though, because it would need to see the switch and then create a bunch of device params and send them to the device. | 14:51.50 |
| | And it needs to happen only once | 14:52.03 |
| | Sp I was thnking of adding it to pl_main_process_options | 14:52.43 |
| henrys | why can't you just say: here's how you do PDF/A1 and give folks a command line? | 14:52.51 |
| kens | henrys I will do so, they would need to supply 3 switches, and one of those would be a special | 14:53.16 |
| henrys | I'll just add it to the docs if you give it to me. | 14:53.27 |
| kens | docs aren't a problem, I'm thinking of the coding at the moment | 14:53.57 |
| | I was going to add ths into pl_main_process_options just below where the DefaultyCMYKProfile and ICCProfileDir are handled | 14:54.40 |
| rayjj | kens: seems like a reasonable place -- keep all the color stuff together | 14:55.04 |
| henrys | I was hoping to avoid special changes in pl_main_process_options which is already a mess | 14:55.04 |
| kens | In there I need to create a lod of device params and send them to the device (also need to check the device *is* pdfwrite) | 14:55.07 |
| | henrys, no choice. | 14:55.14 |
| | Since PCL cannot handle PostScript | 14:55.20 |
| | The PDF and PostScript interprters handle ths by hacky PostScript, but we obviousluy can't do that for PCL | 14:55.46 |
| henrys | oh sigh | 14:56.01 |
| kens | So PCL needs to do 'soemthign different' | 14:56.02 |
| | XPS would be the same | 14:56.15 |
| henrys | can you point to where this happens for PostScript so I can see what's going on? | 14:56.45 |
| kens | You can read PDFA_def.ps in ghostpdl/lib | 14:56.59 |
| henrys | thanks | 14:57.05 |
| kens | weneed to mimic all the pdfmark operations | 14:57.20 |
| | NB PDFA_def.ps is passed to the interpreer as an input *before* any input jobs are run | 14:58.00 |
| henrys | wonders if we really want to do this, it's getting scary. | 14:58.03 |
| kens | Well, its a customer request. | 14:58.20 |
| henrys | right | 14:58.23 |
| kens | Its not that hard to do I thnk, but I can't see any way to avoid the PCL interpreter having to do something | 14:58.47 |
| | In essence it needs to duplicate the 'PDFA_def.ps' setup code | 14:59.30 |
| henrys | I'll look too but I'm sure you've considered everything kens | 15:00.15 |
| kens | I'd be happy if you could see a way around it | 15:00.25 |
| | We could do a spec_op instead of creating device params and sending those, but it still means the PCL interpreter dealing with a command line option | 15:01.08 |
| henrys | rayjj: I skipped the account stuff (reboot from aws) but it's on my list. | 15:01.14 |
| | actually I'll put it on the agenda | 15:02.07 |
| | next meeting off to skype... | 15:02.49 |
| | I bet Miles is island shopping ;-) | 15:06.47 |
| | rich guys have all the fun... | 15:07.03 |
| | oh fredross-perry are you set up with work? Most folks on the staff own parts of the technology and have a continuous stream (backlog) of stuff to do. | 15:17.58 |
| fredross-perry | I have android gsproof on my plate. Plus it sounds like forms might be a thing? | 15:18.56 |
| henrys | fredross-perry: I should have brought that up at the meeting, we need that from Chicage. | 15:23.31 |
| | fredross-perry: does it look to be on schedule | 15:23.48 |
| | ? | 15:23.50 |
| | s/from Chicage/for Chicago/ | 15:24.06 |
| fredross-perry | Iâve been stalled on it for a while, but Iâll make some time ghis week and next. I also need some code form Robin when he gets back. | 15:25.01 |
| henrys | fredross-perry: thanks for reminding me about that. | 15:28.54 |
| rayjj | I thought Robin's code for displaying gsproof files was committed | 15:32.36 |
| mvrhel_laptop | I am going to get the gproof stuff working in gsview hopefully this week to make sure everything is working so that it will hopefully just be the ui portion on the android side that you need to worry about fredross-perry | 15:50.23 |
| fredross-perry | ok | 15:50.43 |
| | thanks | 15:50.45 |
| mvrhel_laptop | I was distracted the past couple days fixing some serious bit rot in the winrt application | 15:51.02 |
| rayjj | hmm, a problem with the new-improved build on windows. Since the project now has all the targets, if you have more than one selected, it tries to build them in separate threads, but the .sbr files might collide (I got a cannot open gxclutil.sbr and had to retry the build) | 17:26.31 |
| vkk | Hi, Plan to test ghostscript as RIP in CTP. How to calibrate dot % for individual CMYK colors for accurate results. Is there any method to select screening (AM/FM). Thanks. | 20:14.17 |
| | and also dot shape | 20:15.11 |
| mvrhel_laptop | henrys: are you there | 23:03.45 |
| | so with that indeterministic bug, I have caught where it is processing memory that has not been initialized. the alpha device and its memory allocations etc is quite confusing though | 23:05.20 |
| | taking a break on this for a bit | 23:05.35 |
| | Forward 1 day (to 2015/08/12)>>> | |
IRC Logs