| <<<Back 1 day (to 2014/10/13) | 20141014 |
mvrhel_laptop | henrys: I actually have not tried yet to tell you the truth :( | 02:23.33 |
| will see if I can get it working tomorrow | 02:23.42 |
rayjj | I don't know if anyone saw my posts here on the weekend. I made a "breakthough" with building for the company M board. Since then, I've gotten *both* gs and pcl6 building and running on it, with ssh access (still can't "push" files to it via scp) but can scp or wget files in and scp results out | 02:26.14 |
| but I *can* do performance tests | 02:26.41 |
mvrhel_laptop | rayjj: great! | 02:27.06 |
rayjj | I still have a problem with PDF's that I have to investigate, but PS and PCL work | 02:27.13 |
mvrhel_laptop | one would think if PS was working PDF would be working | 02:27.27 |
| rather odd | 02:27.44 |
rayjj | no PDF's run -- they all give undefined in --run-- as soon as the Page object is processed, which is AFTER a lot of the PDF has been processed (using -dPDFDEBUG | 02:28.18 |
mvrhel_laptop | strange | 02:28.27 |
rayjj | yeah, so probably something related to needing the pdfwrite device (which I did *not* include in the build) | 02:29.00 |
mvrhel_laptop | oh yes | 02:29.09 |
| pdfwrite is always started up. | 02:29.15 |
| even when it is not the output device | 02:29.19 |
rayjj | I did trip over some bitrot with the build when I am using a small subset of devices | 02:29.25 |
mvrhel_laptop | something I complained about once | 02:29.31 |
rayjj | for instance, since chrisl's commit for stream_dct_finalize we get an undefined from the link because sdcte.dev isn't included | 02:30.41 |
| since the jpeg encoder doesn't add much, I'm inclined to just fix that by adding sdcte.dev to all builds | 02:31.22 |
| but I'll leave that to chrisl :-) | 02:32.12 |
mvrhel_laptop | :) | 02:32.27 |
rayjj | I guess there's a gravity well centered where chrisl lives since stuff seems to roll downhill that way ;-) | 02:33.32 |
mvrhel_laptop | stuff.... | 02:50.33 |
ara_ | hi | 08:48.18 |
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. | 08:48.18 |
jogux | hi ara_ | 08:54.03 |
ara_ | if I want to use for our commercial app , is the library free to use | 08:55.13 |
kens | No. | 08:55.20 |
ara_ | if not then how much is the cost | 08:55.31 |
kens | Both Ghostscript and MuPDF are distributed under the AGPL, for commercial use you need a commercial licence. | 08:55.48 |
ara_ | can you give me idea regarding the expense | 08:56.07 |
kens | ara_ : you need to contact sales@artifex.com. They will send you a big questionnaire, much of which will be irrelveant to you. Just fill in what you can and return it. | 08:56.29 |
| ara_ : Only engineers here, we don't do commercial discussions | 08:56.47 |
| Bu it would help to know what product you are talking about and what your proposed usage is. | 08:57.12 |
| You should definitely pass that information to sales when you contact them | 08:57.27 |
ara_ | okay thanks .. :) | 08:58.04 |
| yaah... | 08:58.04 |
Robin_Watts | ara_: MuPDF or gs? | 09:04.01 |
ara_ | MuPDF | 09:04.14 |
Robin_Watts | ara_: As kens says, the costs vary according to what you want to do with it. If you are doing an app where 1% of it's functionality is provided by MuPDF then you'll pay less than you would for an app where 90% of its functionality is derived from MuPDF. | 09:05.28 |
| hence giving Scott as much information as you can will help. | 09:05.47 |
ara_ | Robi_Watts_: thanks for the detail clarification | 09:06.30 |
| actually I am exploring options for the pdf rendering library in android , could not found any better library that is free .. now exploring this MuPDF | 09:07.26 |
Robin_Watts | I like to think you won't find a better PDF rendering library. | 09:08.03 |
| Adobe just licensed it. | 09:08.10 |
ara_ | hii | 10:38.41 |
Robin_Watts | hello again ara_. | 10:40.17 |
ara_ | Can you suggest any good example source with android studio project .. which is using MuPDF lib | 10:42.17 |
Robin_Watts | with android studio, no. | 10:42.55 |
ara_ | I want to learn how to integrate, build and use this MuPDF lib within my own android app... any suggession would be helpfull | 10:43.04 |
Robin_Watts | All my android work is done at the command line - no IDE. | 10:43.26 |
| Other people here use eclipse. | 10:43.33 |
| I believe that mattchz used android studio though, but he's not here any more. | 10:43.56 |
ara_ | ohh... okkay.. I have loaded the DroidReader app sources in eclipse and exported for android studio | 10:44.10 |
| but when I am trying to ndk build it gives some error | 10:44.21 |
Robin_Watts | DroidReader? | 10:44.30 |
| Surely you want to be working from our sources? | 10:44.40 |
ara_ | actually I want to learn how to integrate and use the lib with my android app | 10:45.23 |
Robin_Watts | Work from our sources, and we can try to give you pointers. | 10:46.00 |
| Importing it into android studio shouldn't be hard really. | 10:46.17 |
| There isn't that much makefile cleverness. | 10:46.36 |
ara_ | that would be nice ... | 10:46.45 |
| where shall i start | 10:46.58 |
Robin_Watts | cloning the git repo from mupdf.com would seem sensible. | 10:47.16 |
ara_ | okay ... | 10:47.27 |
| i will clone .. | 10:47.30 |
Robin_Watts | Then all the android specifics are in platform/android | 10:47.44 |
| There are step by step build instructions in platform/android/ReadMe.txt. | 10:48.02 |
| Do not skip step 10. | 10:48.05 |
ara_ | I would really like to work there , but before that I am in need to just use the lib within my app | 10:48.47 |
| can you give me any good link example for that | 10:49.20 |
Robin_Watts | ara_: No, sorry. | 10:51.26 |
| To use the lib in your own app, the first step is to be able to build the lib. I've given you pointers on how to do that. | 10:51.52 |
ara_ | Robin_Watts : yahh.. thanks lot ... I will be lookin at those | 10:52.17 |
| :) | 10:52.21 |
tor8 | dinamic_: did you figure out how to store C pointers in MuJS yet? | 10:53.47 |
nsz | tor8: fyi i wrote a short something about the silly Math.round spec | 12:02.50 |
| http://port70.net/~nsz/48_round.html | 12:02.53 |
| (for those who are interested in floating-point arithmetics things) | 12:03.38 |
jogux | nsz: nice :-) | 12:59.59 |
kens | : Hmm, not a good flight to be on: | 13:13.16 |
| http://www.reuters.com/article/2014/10/14/us-usa-flight-california-idUSKCN0I30LN20141014 | 13:13.16 |
henrys | paulgardiner: to: joann cc:staff for vacations. | 14:00.17 |
paulgardiner | henrys: oh okay. I'd noticed others doing so, but thought that was because of vacation allocations, whereas I just invoice for the time I work. | 14:03.07 |
henrys | paulgardiner: yes but eventually I'm going to get Joann to start a shared Google calendar ... | 14:04.49 |
paulgardiner | Right. Makes sense. | 14:05.28 |
henrys | oh god all the women are talking about shopping in London... this trip is going to hurt | 14:20.23 |
kens | thinks no worse than New York..... | 14:21.13 |
| marcosw have you seen the support emails from Helmut ? | 14:30.49 |
marcosw | possibly. there were 3 or four last week. | 14:31.14 |
kens | Anopther today, he seems concerned you haven't replied | 14:31.26 |
henrys | I guess we should do the meeting... | 14:32.24 |
kens | well its time. :-) | 14:32.34 |
| No chrisl today | 14:32.39 |
henrys | kens: the firstpage lastpage business? I think marcosw should say we don't have a solution in sight unless you have a good candidate solution. | 14:33.13 |
kens | I haven't looked at that yet, but I thinkits perfectly possible, eventually | 14:33.40 |
marcosw | kens: okay, I'll send him a reply. he's having strange issues with font substitutions being different in 9.10 and 9.15 but I can't duplicate it on Linux, so either it's a windows only issue or something to do with his configuration. I'll open a bug. | 14:33.53 |
kens2 | Bad time for network to glitch | 14:35.25 |
henrys | marcosw: will it be long before we have xpswrite in the clusters. I know you are busy with other stuff... | 14:35.27 |
kens2 | henrys I was looking at some PCL indeterminisms with pdfwrite, I'll look at FirstPage instead | 14:35.51 |
rayjj | henrys: marcosw: I posted a comment on bug 695568 -- the gsmchunk.c change made NO difference | 14:35.51 |
henrys | paulgardiner: have you tested your office 365 subscription? | 14:36.10 |
marcosw | henrys: i have been busy with SOT testing, but to be honest working on Ghostscript testing code would be a welcome break :-) | 14:36.25 |
mvrhel_laptop | let me know then I can push my xpswrite stuff. | 14:36.52 |
henrys | marcosw: well I think SOT has to remain priority. | 14:36.58 |
paulgardiner | henrys: Oh no I haven't yet. It wont expire or anything, will it? | 14:37.48 |
mvrhel_laptop | paulgardiner: you are like me | 14:37.59 |
rayjj | tor8: dinamic_: the issue with "storing C pointers in mujs" (and I assume passing them in and out of C functions) could probably be done with %p pointers in strings | 14:38.18 |
Robin_Watts | and me, except I ran into something where I needed it :) | 14:38.22 |
henrys | paulgardiner: can you just try it, mine doesn't work so I'd like to go back to Joann with all the not working subscriptions | 14:38.33 |
marcosw | henrys: understood, but I think the ATS test system is up and running with only a couple of "minor" issues, so the remaining issue is how we do the switchover from Glasgow to Fremont. | 14:38.39 |
mvrhel_laptop | I will try min today | 14:38.41 |
paulgardiner | will do | 14:38.42 |
mvrhel_laptop | mine | 14:38.43 |
Robin_Watts | marcosw: I thought we wanted to split the VMs across more machines first? | 14:39.11 |
henrys | mvrhel_laptop, paulgardiner : she sent a login just a copy paste username and password | 14:39.17 |
Robin_Watts | or has that be done? | 14:39.22 |
| The new super turbo machine eating FBS needs more cores to cook. | 14:40.02 |
mvrhel_laptop | Robin_Watts: so the tower tickets are for after I leave right? | 14:40.11 |
Robin_Watts | It's tasted blood with Josephs macbook and wants more... | 14:40.12 |
| mvrhel_laptop: They are for the monday evening after the staff meeting. | 14:40.25 |
mvrhel_laptop | ok. darn | 14:40.31 |
Robin_Watts | Sadly I didn't get to pick dates :( | 14:41.19 |
mvrhel_laptop | no problem I understand | 14:41.40 |
Robin_Watts | You can do tours of the Tower on normal days though, just not private ones in the evening. | 14:41.41 |
henrys | rayjj: I assigned norberts performance issue to you temporarily. | 14:42.11 |
paulgardiner | henrys: Seem to work. I have a "Welcom, let's get started" screen | 14:42.15 |
henrys | paulgardiner: great thanks | 14:42.35 |
mvrhel_laptop | Robin I do have a mupdf question that has come up. Miles visited a potential customer and they may actually want to use a version of gsview. But, they need something somewhat significant which is some editing capabilities | 14:42.55 |
Robin_Watts | has disconnected... | 14:43.13 |
mvrhel_laptop | hehe | 14:43.24 |
tor8 | rayjj: there's a special JS object class in MuPDF for storing C pointers -- userdata | 14:43.38 |
Robin_Watts | actually, that'd have to be more fun than SO :) | 14:43.39 |
rayjj | henrys: yes, I saw that. that's what I posted -- I cannot reproduce the performance issue with gsmchunk.c, and FWIW, the commit f945f8b~1 (that does twice as many pages) is slower than HEAD | 14:43.49 |
Robin_Watts | mvrhel_laptop: What sort of editing? | 14:44.08 |
mvrhel_laptop | so the question is, acrobat pro allows one to replace text, all be it with font issues spacing issues etc. how hard would it be for us to do this with mupdf | 14:44.18 |
rayjj | tor8: OK. So why was dinamic having problems ? | 14:44.19 |
tor8 | rayjj: js_pushuserdata will wrap a C pointer in a javascript object, and js_touserdata will extract it | 14:44.24 |
marcosw | Robin_Watts: I've done that. there are now virtual machines running on two different physical machines. I haven't figured out what the optimal division is but that can be tuned later. | 14:44.25 |
mvrhel_laptop | also, they want to replace hyperlinks. | 14:44.32 |
Robin_Watts | marcosw: Oh, excellent. | 14:44.36 |
tor8 | rayjj: probably because he didn't know where to look in the API :) | 14:44.37 |
mvrhel_laptop | and perhaps some images | 14:44.49 |
| Robin_Watts: you should ask what did they not want to edit | 14:45.04 |
Robin_Watts | mvrhel_laptop: Well, we have a new thing in mupdf that might help for this. | 14:45.07 |
mvrhel_laptop | oh great | 14:45.15 |
tor8 | rayjj: http://mujs.com/docs/examples in the "Complete userdata example" section | 14:45.16 |
| dinamic_: ^ | 14:45.21 |
Robin_Watts | We have the ability to do filtering of the pdf operator stream. | 14:45.32 |
marcosw | The remaining issue, and I will ask Joseph about at the SOT meeting, is build issues with the ios 7 version, but I'm pretty sure that's just a certificate issue. | 14:45.38 |
mvrhel_laptop | I was thinking a filter would be ideal | 14:45.43 |
henrys | tor8: I wanted to ask about epub are we still on for a partially functioning demo by end of year? | 14:46.03 |
Robin_Watts | So, potentially we can make changes inline like that, but it's a bit hairy. | 14:46.10 |
| What do you mean by "replacing hyperlinks" ? | 14:46.52 |
tor8 | henrys: I hope so :) I'm bashing on the visual box layout model at the moment | 14:47.03 |
rayjj | henrys: I'll be doing performance testing of the company M board today. Initially I'll do the files we did on the raspberry to rgb contone (what their printer pipeline probably prefers), CMYK contone, Gray and 1-bit mono | 14:47.06 |
| mvrhel_laptop: henrys: are there other files (other than J9, J11, J12 and PLRM 100 pages) ? | 14:48.05 |
henrys | mvrhel_laptop: every "miles" question should be answered with we don't have resources both Robin_Watts and I have responded that way to recent requests. | 14:48.16 |
mvrhel_laptop | Robin_Watts: well lets say I am replacing the word Joe with Bob everywhere and one of the Joe text items has a hyperlink to Joe's store. I instead want to replace it with a hyperlink to Bobs store | 14:48.31 |
henrys | I think we can all agree we are fully engaged with what we have and should not add more | 14:48.57 |
mvrhel_laptop | henrys: its hard when he is visiting the customer right there and calling. | 14:48.58 |
Robin_Watts | mvrhel_laptop: Right. That would again be handled by the filter stuff. | 14:49.05 |
mvrhel_laptop | henrys: I think he wanted a ball park figure on how long it would take. we would charge the customer. | 14:49.29 |
| they have lots of money | 14:49.33 |
| on the gsview side of things this is all easy. Its the mupdf side I was not sure of | 14:49.51 |
henrys | okay you'll see ;-) | 14:49.51 |
Robin_Watts | So it's technically possible, but it would still be a non insignificant amount of work, would require lots of fiddling, and, as henrys says, we don't currently have capacity. | 14:49.54 |
mvrhel_laptop | Robin_Watts: ok. if you did thave the time, would you guess a couple months of work? | 14:50.13 |
Robin_Watts | Potentially more. | 14:50.22 |
mvrhel_laptop | just so I can give miles a number | 14:50.23 |
| oh | 14:50.25 |
| 6 months? | 14:50.31 |
Robin_Watts | Cos everytime we do something, it'll go to the customer and they'll come back and say "but on this file we want to change X..." etc. | 14:50.50 |
mvrhel_laptop | sue | 14:50.58 |
| sure | 14:51.00 |
| yes there would be a huge tail on this | 14:51.13 |
| AR pro has issues with text replacement itself | 14:51.27 |
Robin_Watts | I'd not want to be tied down to a figure without a strong spec being hammered out. | 14:51.31 |
mvrhel_laptop | it is a ugly thing | 14:51.32 |
henrys | mvrhel_laptop: it doesn't matter how much money they have. Who's going to work on it? | 14:51.35 |
mvrhel_laptop | henrys: yes. I understand. | 14:52.00 |
Robin_Watts | henrys: Me! Me! Pick Me! | 14:52.04 |
mvrhel_laptop | :) | 14:52.11 |
| I wonder if Fred-Ross-Perry could do something no this. just a wild thought | 14:52.43 |
Robin_Watts | It's more internal PDF fiddling than he might be comfortable with. | 14:53.36 |
mvrhel_laptop | yes | 14:53.40 |
henrys | mvrhel_laptop: yup that's a wild one alright | 14:54.12 |
mvrhel_laptop | :) | 14:54.17 |
| you guys are funny this morning | 14:54.30 |
| I knew I would get blasted bringing this all up :) | 14:54.49 |
jogux | henrys: emobix should/may have some spare capacity again in 2 months or thereabouts, so some potential there to shuffle things around to free up Robin if it helps. | 14:55.28 |
mvrhel_laptop | rayjj: other files to test timing with on the board you mean? | 14:56.10 |
henrys | jogux: we can talk at the other meeting, but we need to get a little more under control before taking on new stuff. | 14:56.31 |
jogux | nods. | 14:56.50 |
rayjj | mvrhel_laptop: yes. I know those are probably of interest to the Japanese market, but did company M mention any during your meetings ? | 14:56.57 |
mvrhel_laptop | rayjj: no they did not | 14:57.07 |
rayjj | mvrhel_laptop: I'm going to ask MQ for files that they may have timings for (with the *other* PDL that we know they have). | 14:57.57 |
mvrhel_laptop | rayjj: oh good idea | 14:58.09 |
| and get their timing data too :) | 14:58.17 |
rayjj | henrys: now that we can get timings, should it be Miles or me that asks them for performance data from the other PDL ? | 14:58.37 |
| henrys: I don't want to tweak any noses | 14:58.55 |
henrys | rayjj: it's fine for you to ask them I don't know if he'll tell you. | 14:59.59 |
rayjj | henrys: OK. | 15:00.13 |
henrys | well we are at the 1/2 hour let's call it. | 15:01.18 |
| mvrhel_laptop: both robin and I have responded to miles' request with the short of resources. It is worthwhile if hears that on many fronts ... | 15:08.14 |
mvrhel_laptop | henrys: I understand | 15:10.18 |
| will do | 15:10.20 |
jogux | paulgardiner: ping ;-) | 15:19.28 |
rayjj | henrys: mvrhel_laptop: email sent with status and questions to MQ and the rest of the people he cc'ed in his latest email to me (includes Miles, Scott and tech) | 15:33.17 |
henrys | rayjj: great | 15:40.32 |
rayjj | cool. I can ssh to and from between my raspberry and the company M board (had a problem with keys but that's sorted out) | 16:07.40 |
dinamic_ | evning folks | 16:07.53 |
rayjj | morning for me, dinamic_ | 16:08.16 |
dinamic_ | :) | 16:08.21 |
| passing pointer as strings sound like a bad idea :) | 16:08.35 |
rayjj | dinamic_: did you see tor's comments earlier about the correct method ? and the link to the documentation ? | 16:09.00 |
| http://mujs.com/docs/examples in the "Complete userdata example" section | 16:09.16 |
dinamic_ | rayjj, i use userdata and it works as intended | 16:09.53 |
rayjj | dinamic_: pointers as strings is a (slight) performance hit, but it's no worse than passing pointers in general -- all the usual problems about stuff moving and the pointer being 'stale' | 16:10.36 |
| dinamic_: great | 16:10.57 |
| enjoy mujs :-) | 16:11.12 |
dinamic_ | i doo.. | 16:11.16 |
| i ditch spidermoneky | 16:11.21 |
| for the nice and clean stack approach | 16:11.28 |
kens | OK I'm off, will try and make some progress on the FirstPage stuff tomorrow and the rest of the week. Goodnight all | 16:11.29 |
dinamic_ | i love it. | 16:11.33 |
rayjj | bye, kens | 16:11.47 |
dinamic_ | using it for plugins for my new project.. | 16:11.48 |
| rayjj, do you know how i could get the error string from a failed js_ploadstring() | 17:37.59 |
| rayjj, solved it :) | 17:46.25 |
rayjj | dinamic_: solved it "got the error string" or solved it "got rid of the error" ? | 17:50.48 |
dinamic_ | rayjj, got hold on the exception error string | 17:51.04 |
rayjj | dinamic_: I was just curious | 17:51.19 |
dinamic_ | :) | 17:51.48 |
| rayjj, i have another question | 23:04.18 |
| nevermind.. | 23:05.04 |
| :) | 23:05.06 |
| Forward 1 day (to 2014/10/15)>>> | |