| <<<Back 1 day (to 2014/01/27) | 2014/01/28 |
Robin_Watts | fair enough. | 00:10.13 |
| tor8: for the logs; in this kind of thing the spec is extremely untrustworthy. | 00:30.19 |
| It has a diagram in it that explicitly shows what operators are allowed in what states - and when I first wrote my PDF interpreter I wrote it to only accept operators in the right states. And most PDF files failed :( | 00:31.16 |
| I'm thinking of the case where we have a Text rendering mode that uses stroking. In that case, it would be quite likely that you might want to change the dash pattern while doing text. | 00:31.56 |
vtorri_ | hey | 07:38.37 |
ghostbot | hi, vtorri_ | 07:38.37 |
vtorri_ | what are the main added features in the current dev tree, compared to 1.3 ? | 07:38.56 |
| oups | 07:40.06 |
| for mupdf :) | 07:40.11 |
Robin_Watts | vtorri: Let me look. | 11:17.02 |
| Lots of bugfixes | 11:17.12 |
| Initial support for digital signatures. | 11:17.55 |
| support for incremental updates. | 11:18.26 |
| Some of the plotting functions were optimised for speed. | 11:19.08 |
| Banding operation for mudraw | 11:19.43 |
| Fixed subpixel positioning that made type 3 fonts look crap at some rotations. | 11:20.09 |
| SVG output device (still a work in progress, but not bad) | 11:20.36 |
| Much improved ios port. | 11:22.33 |
| bug fixes to android. | 11:22.37 |
| lots of openjpeg/jbig2 fixes. | 11:22.50 |
| Windows 8 viewer. | 11:23.05 |
| CMYK output support | 11:23.12 |
| better repairing of broken PDF files. | 11:24.27 |
| faster rendering of mesh based shadings. | 11:24.58 |
| better handling of multiple threads/contexts. | 11:25.31 |
| I've got some JNI extensions on a branch, but I don't know that that'll make it for the next release. | 11:27.29 |
tor8 | Robin_Watts: right. well, I guess doing the allocation for an array object for every bit of well-formed BT/ET text strings might be a reasonable sacrifice to simplify the code and handle the bad case you're worrying about. | 12:26.05 |
Robin_Watts | tor8: I've run into a couple of problems with this stuff. | 12:26.57 |
| splitting different sets of pdf_run functions out, I mean. | 12:27.36 |
tor8 | okay... | 12:28.11 |
Robin_Watts | The array handling is one of them - currently I'm reconstructing a singleton array with each part so I can call TJ with it. | 12:28.51 |
| and the graphics state stuff is nastily intertwined with run_xobject etc. | 12:31.57 |
| If we straighten the array handling to the more obvious code, and I do some tweaks here, I think I have a solution. | 12:39.17 |
tor8 | go for it. | 12:59.25 |
| the xobject state balancing, should that call the pdf_run_xxx functions as well, to make it easier? | 12:59.58 |
Robin_Watts | tor8: It does now in my version. | 13:07.05 |
| well, my version is in bits all around me, but that fragment should be in place :) | 13:07.31 |
tkamppeter | I have a problem with the screen display of a .eps file, it displays correctly only if rotated to portrait orientation. See https://bugs.launchpad.net/ubuntu/+source/ghostscript/+bug/1242678, especially my comment #13. | 13:08.00 |
| Is something like that known? Has anyone an idea what is going on here? | 13:08.25 |
Robin_Watts | tkamppeter: I don't recall seeing anything like that before. | 13:09.55 |
| Can the command line be simplified any further? | 13:10.33 |
| Are the screwy resolutions important? Are the alphabits important? Does it go wrong when drawing to (say) ppm raw ? | 13:11.16 |
kens | THere appear to be commas in the resolution, is that real ? I don't think we support that. | 13:12.31 |
tkamppeter | Robin_Watts, I can simplify it to "gs -sDEVICE=x11 -g574x402 -c '<< /Orientation 3 >> setpagedevice .locksafe' -f ~/Documents/M_1_no_of_ppts.eps" | 13:15.39 |
| Robin_Watts, the problem disappears if I remove the "-g..." parameter. If I swap the two numbers in the "-g..." argument, getting a landscape-shaped frame, the problem still persists. | 13:16.50 |
kens | The original command line doesn't seem to have a -g | 13:18.11 |
| Actrually yes it does | 13:18.22 |
| I kissed it | 13:18.26 |
| missed * | 13:18.30 |
tkamppeter | Robin_Watts, kens, with ppmraw as output device the problem still persists. | 13:18.55 |
| So it reduces to not displaying in Landscape orientation when "-g..." is supplied, independent of the output device. | 13:20.16 |
Robin_Watts | tkamppeter: ok, that narrows it down nicely to the rotation code. | 13:22.19 |
tkamppeter | Robin_Watts, is an upstream bug report needed? | 13:26.51 |
Robin_Watts | tkamppeter: It would help ensure it was not forgotten. | 13:27.04 |
| I'm busy in other code at the moment, so can't look immediately (and I might not be best placed to look anyway as I have no experience of the orientation code) | 13:27.30 |
tkamppeter | Robin_Watts, kens, I have reported http://bugs.ghostscript.com/show_bug.cgi?id=694979 now. | 13:46.48 |
kens | The file does a '90 rotate' so if you set the Oritnation to 3, the content will indeed be rendered portrait | 13:49.56 |
| (assuming Orientation rotates the media content, I would have to go read the PLRM and I'm busy with something else atm) | 13:50.28 |
tkamppeter | kens, I tried all 4 values for the orientation and always got Portrait, 2 times with the Y axis at the bottom and 2 times with the Y axis at the top. | 13:55.05 |
kens | tkamppeter : its possible the EPS program is doing that, it wilil take some time to figure it out, and I don;t have the time at the moment | 13:55.36 |
paulgardiner | Robin_Watts, tor8: did you see Marcos's tech email about my commit causing compiler warnings? The commit just adds one line to one of the MuPDF source files. No intentional change to thirdparty. Any idea what I've done? | 13:56.45 |
| I wondered if I'd forgotten to submodule update, but if so I'd expect some sign of that in the commit | 13:57.34 |
Dirrk | Hi. I'm trying to trace a bug that causes wrong margins while printing. Could anyone explain to me what /.HWMargins actually does, or point me to some documentation? Is it driver dependent? The driver in question is "raw". | 14:04.26 |
kens | Arrr Dirrk you want to read the PostScript Language Reference Manual for information on all things PostScript. | 14:04.59 |
| Anything beginning with a '.' is usually a Ghostscript extension and not generally intended for end-user messing about with | 14:06.14 |
Dirrk | That may be so, but doesn't help me solving this problem :-) | 14:06.36 |
| The driver injects a .HWMargin directive, and I want to find out | 14:06.51 |
| if this is maybe the cause | 14:07.01 |
kens | Well since you haven't described the problem...... | 14:07.16 |
Dirrk | Problem: | 14:07.24 |
| I'm printing PDF documents | 14:07.31 |
| on two printers, both using gs somewhere in the pipeline | 14:07.44 |
| the top of the page is cut off | 14:07.53 |
| and the margins look weird | 14:07.58 |
| same effect on both printers | 14:08.10 |
| one printer is a brother with drivers downloaded from the brother site | 14:08.26 |
| this driver is REALLY weird, but I'm trying to follow what it does | 14:08.45 |
| so I can fix the problem | 14:08.49 |
kens | You will need to be *much* more explicit. 'somewhere in the pipeline' doesn't tell us much, you haven't said what OS you are using, nor what device in GS, nor what PD: you are sending to the printers | 14:08.52 |
| PDL* | 14:09.09 |
Dirrk | Debian Linux, I'm sending a PDF file to measure margins | 14:09.13 |
kens | SO you are using CUPS ? | 14:09.20 |
tor8 | paulgardiner: I recognize many of them as old, known, warnings | 14:09.24 |
Dirrk | yes, and cups | 14:09.24 |
kens | THen you should talk to tkamppeter here | 14:09.32 |
paulgardiner | tor8: oh good. So maybe it was the test suite getting confused | 14:10.02 |
Dirrk | just knowing what .HWMargins actually does when rasterizing would be a start :-) | 14:10.36 |
Robin_Watts | paulgardiner: The 'new' warning detection sometimes gets confused and thinks all warnings are new etc. | 14:10.58 |
Dirrk | I know it should describe the margins on the page where the device can't print | 14:11.04 |
| but I don't know what ghostscript actually does with this information | 14:11.21 |
kens | See the GS documentation, in particular langauge.htm and devices.htm | 14:11.30 |
Dirrk | so a pointer would be nice. :-) | 14:11.32 |
paulgardiner | Robin_Watts: oh okay. Thanks | 14:11.37 |
Dirrk | It says "Size of non-imageable regions around the edges of the page, in points (units of 1/72in; see the notes on measurements in the documentation on devices). " | 14:14.02 |
kens | Yes, that's what it does. | 14:14.18 |
| The imageable area of the media is the media size minus the HWMargins | 14:14.36 |
Dirrk | so what is the effect? scale page down? leave a border? | 14:14.40 |
kens | Depends on teh device.# | 14:14.48 |
Dirrk | yes. it's the "bit" device. | 14:14.54 |
kens | Normally you would use this with a printer device | 14:14.59 |
Dirrk | the cups driver uses it with the "bit" device. | 14:15.09 |
kens | You can experiment easily enough | 14:15.13 |
| Draw a border round the page, send it to the device, see what it draws | 14:15.29 |
Dirrk | need some way to display the raw data, though :-) | 14:15.31 |
| experimented already with gv, no change | 14:15.39 |
kens | I know nothing of gv. | 14:15.59 |
Dirrk | ghostview | 14:16.03 |
kens | gv != ghostview | 14:16.13 |
| AFAIK gv is not the same as gsview | 14:16.38 |
| And gsview does all sorts of stuff, so I very much doubt it will have any effect there | 14:17.00 |
Dirrk | I don't have gsview | 14:17.04 |
kens | gsview is ghostview | 14:17.15 |
Dirrk | anyway, I'll try to display the raw data. thx. | 14:17.24 |
Robin_Watts | Dirrk: I would be suprised if setting margins caused the page to scale down. | 14:20.05 |
kens | Robin_Watts : I wouldn't be | 14:20.22 |
| If you send an A4 page to a printer with A4 media, but a 1/4" non-printing margin you have a choice | 14:20.47 |
Dirrk | I would be surprised, too, but better check | 14:20.53 |
kens | Don't print some of the job, or scale it very slightly to fit the imageable area | 14:21.01 |
Robin_Watts | really? Well, I bow to your greater experience, kens. | 14:21.02 |
kens | Normally I would expect no scaling, but I have seen both done. | 14:21.26 |
| IIRC PPD files can contain an ImageableArea key which tells you wht the printer is capable of. | 14:22.21 |
| God a PDF file whcih sets the CTM to a degenerate matrix, then sets the text matrix the same way *and* leaves garbage on the stack ? | 14:23.49 |
Dirrk | ok, for the record: The bit-driver uses HWMargin to blank out the specified margins, without reducing the size of the drawing area. | 15:06.20 |
| this might actually be related to my problem, I'll have a closer look. | 15:06.35 |
paulgardiner | Robin_Watts: small Android change on paul/master | 15:21.21 |
ray_laptop | morning all | 15:21.56 |
paulgardiner | Hi ray_laptop | 15:22.10 |
Robin_Watts | paulgardiner: Is there a reason we don't put BROWSABLE in now? | 15:22.10 |
| Morning ray_laptop | 15:22.13 |
ray_laptop | did anyone get my itinerary I sent on the 23rd ? I got a strange email from Miles that sounded like he didn't get it | 15:23.02 |
paulgardiner | Robin_Watts: leaving it out avoids the problem of MuPDF trying to open urls which cause it to crash. | 15:23.03 |
ray_laptop | (sorry for the off topic query) | 15:23.17 |
Robin_Watts | URLs cause us to crash? | 15:23.21 |
paulgardiner | Robin_Watts: when a link is passed directly from K9 without first downloading. | 15:23.55 |
Robin_Watts | ray_laptop: 24th for me (00:01 :) ) but yes I got it. | 15:24.52 |
henrys | ray_laptop: I got your itinerary | 15:24.55 |
Robin_Watts | Sent to Everybody + Miles | 15:25.01 |
ray_laptop | Robin_Watts and henrys: Thanks I guess Mles just missed it | 15:25.13 |
paulgardiner | Robin_Watts: I'm guessing it's passing a "content" url which is not based on a concreate file. | 15:28.35 |
Robin_Watts | paulgardiner: Right. | 15:28.37 |
| I thought I'd made changes so that worked now. I remember doing something to make gmail work. | 15:28.37 |
kens | ray_laptop : I got your itinerary yes | 15:28.38 |
paulgardiner | Robin_Watts: yeah we handle some cases of content urls - ones where there is an underlying file | 15:28.38 |
Robin_Watts | yeah, in onCreate we call either openFile or openBuffer | 15:28.39 |
henrys | 2 minutes | 15:28.39 |
paulgardiner | Robin_Watts: hmmm, You'd think openBuffer would handle most cases. | 15:29.20 |
| Robin_Watts: perhaps I should check it out further | 15:29.40 |
henrys | meeting time⦠paulgardiner are we really comfortable mvrhel_laptop's demo is not patent encumbered. There was one sentence in your patent followup that concerned me, about briefly using the intermediate format. | 15:31.08 |
| ? | 15:31.09 |
mvrhel_laptop | monring | 15:32.17 |
Robin_Watts | henrys: The current demo loads into Edr, then outputs from that. So we do use Edr, currently, yes. | 15:32.46 |
paulgardiner | henrys: I don't know enough about patents really. One way of interpreting that patent could mean it applies to the use of PDF as an intermediatory | 15:33.11 |
mvrhel_laptop | was there an email that told me where to grab the apk by chance? | 15:33.16 |
Robin_Watts | Whether the use of Edr conflicts with the Edr patent is the big question. | 15:33.21 |
| mvrhel_laptop: It's in pauls home dir on casper. | 15:33.32 |
mvrhel_laptop | ok. I will grab it thanks | 15:33.43 |
chrisl | Robin_Watts: is there anything you can do with the code that doesn't go through Edr? | 15:34.18 |
paulgardiner | chrisl: only in the very very long term | 15:34.52 |
Robin_Watts | That demo contains all the wasp rendering code too - though the way we invoke it currently doesn't exercise it. Stripping that out is one possible job I might be able to do indepdendent of Pauls work. | 15:34.56 |
henrys | I had another legal question for the meeting ;-^ can we release this demo? mupdf is GPL so we have it using something without source but we are the holders of the mupdf license. | 15:35.21 |
marcosw | morning, sorry to be late. | 15:35.31 |
Robin_Watts | henrys: No, we shouldn't release the demo as is. | 15:35.48 |
chrisl | paulgardiner: I mean as the code we bought stands, is there any functionality that doesn't route through Edr? | 15:35.54 |
mvrhel_laptop | henrys: we can pay a license to ourselves... | 15:35.55 |
paulgardiner | henrys: we must be able to grant ourselves a licence for that. | 15:36.08 |
| Too slow. Bah | 15:36.20 |
mvrhel_laptop | :) | 15:36.28 |
Robin_Watts | For the benefit of those that don't know... the Picsel code potentially relies on a load of Patents that Picsel took out years ago. | 15:36.49 |
| Those patents were sold to Samsung and Picsel got a license back from Samsung to use those patents for fileviewer applications. | 15:37.23 |
paulgardiner | chrisl: the main thing we wanted was the agents (which interpret the office file variants). The agents output EDR | 15:37.34 |
mvrhel_laptop | and what is EDR? | 15:37.46 |
Robin_Watts | It came to light over the weekend that those patent licenses may not have survived the liquidation. | 15:38.04 |
chrisl | paulgardiner: so without the right to use those patents, the code is essentially useless, as it stands..... | 15:38.07 |
henrys | paulgardiner, mvrhel_laptop I don't understand what you mean - is Stallman going to march up to San Rafael and tell miles to release the Office source or face the wrath of GPL lawyers. | 15:38.17 |
chrisl | henrys: as Artifex holds the copyright to mupdf, we can release it however we want | 15:39.05 |
henrys | chrisl: but can we still call mupdf GPL? | 15:39.32 |
| seems like we'd need a new license. | 15:39.41 |
Robin_Watts | chrisl: There is some question as to which of the patents still apply. | 15:39.50 |
| Picsel took the view that they had rejigged the code to avoid the need for patent licenses. | 15:40.11 |
| We need to convince ourselves of this. | 15:40.22 |
chrisl | henrys: there are already customers shipping mupdf in closed source apps, that hasn't precluded us from shipping AGLP mupdf | 15:40.27 |
henrys | Robin_Watts: I am assuming the patents aren't an issue. | 15:40.27 |
paulgardiner | chrisl: Certainly not useless. Worse case it it's a good basis for developing our own office format readers | 15:40.36 |
| I mean parser not reader | 15:40.51 |
Robin_Watts | Certainly, I think that by the time we get it integrated within MuPDF properly, we'll be free of the patents. | 15:41.02 |
chrisl | paulgardiner: useless "as it stands" - i.e. nothing we can use "out of the box" | 15:41.08 |
Robin_Watts | There is only 1 patent that *might* be an issue with Pauls current demo. | 15:41.17 |
henrys | paulgardiner: I ask because doing a demo as we have it now on iPhone and Android could attract customers. | 15:41.18 |
| paulgardiner: but we obviously don't want to GPL the office code. | 15:41.49 |
Robin_Watts | henrys: At the moment, we'd have to put the demo out under the Artifex commercial license. | 15:42.03 |
paulgardiner | chrisl: it was always expected to be useless as it stands. It was only when henrys thought of the "convert to PDF" idea that it became possible to do anything quickly with it | 15:42.07 |
Robin_Watts | paulgardiner: That's not quite true - we had talked about rebranding the existing SmartOffice app and continuing to sell that. | 15:42.36 |
paulgardiner | Robin_Watts: oh yes, sorry | 15:42.49 |
Robin_Watts | but that was an afterthought, not the original driving idea of taking the license in the first place. | 15:43.03 |
vtorri_ | Robin_Watts: ok, thanks for the list of changes | 15:43.05 |
chrisl | paulgardiner: I was just wondering if the patent rights could be considered an implicit part of the sale - I guess not | 15:43.22 |
mvrhel_laptop | paulgardiner: so I can't grab the MuPDFOffice*.apk file you have the permissions locked down | 15:43.24 |
Robin_Watts | chrisl: The problem is that the liquidator doesn't have the patent rights to sell. | 15:43.44 |
henrys | paulgardiner: okay well something to mull over. | 15:43.48 |
vtorri_ | no idea when the 1.4 wil be out ? | 15:43.52 |
paulgardiner | henrys: IANAL, but it isn't EDR that is patented, it is the use of an intermediary format, so I don't understand why it doesn't cover display lists as used in many applications. | 15:44.10 |
Robin_Watts | Miles is talking to various people about this now. | 15:44.18 |
ray_laptop | the liquidator didn't have "sub license" rights ? | 15:44.25 |
paulgardiner | mvrhel_laptop: oh sorry. Hang on | 15:44.27 |
Robin_Watts | paulgardiner: I think it may be something specific about the structure of the intermediate format. | 15:44.50 |
paulgardiner | mvrhel_laptop: should be okay now | 15:45.11 |
Robin_Watts | I'd need to read the patent carefully, and I have real problems reading patents. My brain turns to mush. | 15:45.13 |
mvrhel_laptop | paulgardiner: yes. thank you. | 15:45.23 |
Robin_Watts | s/to mush/even more mushy/ | 15:45.30 |
chrisl | Robin_Watts: not an uncommon reaction to reading patents, IME :-( | 15:45.52 |
Robin_Watts | Some of the picsel patents seem impossibly broad. | 15:45.56 |
henrys | mvrhel_laptop: do you have a ballpark idea when gsview will be out - you said the meeting. Does that include distilling? I do think postscript should be part of this beast. | 15:46.09 |
Robin_Watts | If people are interested, I can post a list of the patents to tech. | 15:46.14 |
paulgardiner | Robin_Watts: I think we could fairly easily make variants for each office subformat and then we wouldn't be using a common format. In fact I think we'd need to. | 15:46.16 |
Robin_Watts | vtorri: within a couple of months. | 15:46.25 |
mvrhel_laptop | henrys: yes, I will plan to have gs hooked up too by then | 15:46.37 |
| henrys: but I can't guarantee a polished ready to release app at that point. but I will try | 15:47.00 |
henrys | mvrhel_laptop: sounds good. Are you just going to exec a process or use the API? | 15:47.33 |
mvrhel_laptop | a bit over a month, so it is possible. henrys: my plan was to use the API | 15:47.51 |
henrys | kens:gsprint? It would be nice for gsview and gsprint to go out together. | 15:49.18 |
| chrisl and all - do you think we should add the urw commercial fonts to the private customer area? so customers can see we have these fonts. We don't do anything to advertise them. | 15:50.42 |
ray_laptop | henrys: the problem with having them in the same area is that some customers may not realize that they are licensed separately | 15:51.36 |
chrisl | henrys: I understood these to be a cost option for customers. Anyway, we should probably wait until they actually work | 15:51.42 |
paulgardiner | henrys: not being a patent expert, I don't know how relevant this is, but when we integrate GhostDocs with MuPDF we wont be using EDR as a common intermediary for all fomats. Well be using different mechanisms for PDF XPS JPG etc. | 15:51.56 |
henrys | ray_laptop: there is no such warning for PCL and it is licensed separately. | 15:52.45 |
chrisl | henrys: did we actually establish the license conditions for the LL3 font set? | 15:53.22 |
henrys | chrisl: commercial | 15:53.34 |
| chrisl: I don't know exact dollars and terms | 15:53.56 |
chrisl | So, we probably don't just want to give them to all customers | 15:54.17 |
henrys | paulgardiner: okay so as long as we are going forward with that I think we are safe with the demo. | 15:54.52 |
Robin_Watts | http://www.google.co.uk/patents/US6925597 <- That's the 'Edr' patent I think. | 15:55.20 |
| mvrhel_laptop: 'Edr' is picsels internal representation for documents. | 15:55.39 |
mvrhel_laptop | Robin_Watts: ok | 15:55.47 |
chrisl | henrys: I would say that making (current and future) customers aware of the availability of the LL3 font set is more Miles and Scott's territory than ours | 15:55.54 |
Robin_Watts | Electronic Document Representation, or Epage Document Representation or something. | 15:56.00 |
paulgardiner | henrys: I looked at the iOS app again and we do handle interaction with other apps (or at least when provoked by the other app). From DropBox, and a file browser I have, I can bring up MuPDF via "Share" and then "Open in". | 15:56.34 |
henrys | chrisl: oh you mean something for the Artifex page. | 15:56.41 |
chrisl | henrys: that and in their direct contacts with customers, too | 15:57.06 |
mvrhel_laptop | hmm claim 1 may cover all display lists..... | 15:57.14 |
paulgardiner | henrys: I think good reader can actually enumerate Dropbox files, but that's using an API very specific to DropBox | 15:57.21 |
chrisl | henrys: although, tbh, I seriously doubt there will be many takers | 15:57.44 |
Robin_Watts | There is dropbox reading code in the picsel source if we wanted to look at that. | 15:57.56 |
paulgardiner | mvrhel_laptop: that's what I think, and so there is loads of prior art | 15:58.11 |
mvrhel_laptop | yes | 15:58.16 |
| it would be *very* difficult for one to defend this | 15:58.31 |
henrys | paulgardiner: looking at mupdf on iOS there is just no draw at all, xps is dead pdf isn't needed - if we had office maybe we get some folks interested in licensing, that's why I asked about using the shortcut like the demo. But now I think it best to wait. | 15:58.31 |
Robin_Watts | mvrhel_laptop: The problem with all such patents is that they claim the world. | 15:58.49 |
| You can throw away much of the fluff and get down to just a few core claims that are the meat of the patent. | 15:59.20 |
paulgardiner | henrys: oh GhostDocs on iOS? | 15:59.23 |
Robin_Watts | but it's always a slight lottery where the fluff ends and the meat begins. | 15:59.43 |
henrys | chrisl: true, unless they really do a good job with cyrillic, greek etc. in the commercial - then I could see that being a distinguisher ... | 15:59.44 |
mvrhel_laptop | yes | 15:59.54 |
| brb. need to help get kids out door | 16:00.15 |
paulgardiner | henrys: we have the forms and annotations in the iOS app now. I may have misunderstood. I thought that was what I was supposed to be doing. | 16:00.16 |
| henrys: and reflow | 16:00.51 |
| ... but maybe that's all handled by the default viewer. | 16:01.15 |
henrys | paulgardiner: it is yes, but I brought up releasing something like mvrhel_laptop's demo because I thought the new formats would be attractive. | 16:01.26 |
chrisl | henrys: well, that's to be established (the fonts I looked at did not have the alleged extra glyphs), and given that most people use their own fonts on their own workstations, and embed them, most companies I've encountered don't consider the extra fonts worth extra cost | 16:02.04 |
ray_laptop | chrisl: the only folks we've had slight interest from are printer companies that need to be able to state 136 PS fonts. But most printer companies already are sunk into MT | 16:03.24 |
henrys | ray_laptop beat me to it. | 16:03.36 |
chrisl | ray_laptop: yep, and they also don't relish paying twice - once for the extra fonts, and once for the extra storage to hold them..... | 16:04.07 |
henrys | paulgardiner: I do hope the iOS features you are working on get us some looks but "office" support is a big thing, certainly if we had the office stuff before you started the interactive iOS and now I wonder if everything should be dropped and we go full steam ahead with Office integration. What do you think? | 16:06.29 |
| s/interactive IOS/interactive iOS, we'd have made it higher priority. | 16:07.22 |
| Sorry that is terribly confusing. | 16:07.29 |
| meeting over 7 minutes ago, sorry also about that. | 16:07.53 |
marcosw | kens: can I ask a question about adding support for mediaposition in ps2write? | 16:08.08 |
kens | We no longer support pswrite marcosw, its deprecated | 16:08.21 |
paulgardiner | henrys: yeah that makes perfect sense. I had been imagining the office stuff was the priority, but just looking to see if a quick further fiddle with the iOS app would give us something releasable. | 16:08.31 |
ray_laptop | ps2write | 16:08.35 |
chrisl | ray_laptop: the customer is referencing pswrite | 16:09.20 |
kens | ray_laptop : the customer asked about pswrite | 16:09.23 |
marcosw | and I'm asking about ps2write... | 16:09.38 |
kens | :-) You said pswrite..... | 16:10.01 |
paulgardiner | henrys: so I'd looked at opening docs from dropbox etc and found it was alread working, so maybe we have enough for a release even if it isn't a big draw | 16:10.03 |
marcosw | no, I didn't, check the logs :-) | 16:10.21 |
henrys | paulgardiner: how did you do that? | 16:10.37 |
paulgardiner | henrys: just tried every conceivable gesture within DropBox. | 16:11.18 |
kens | OK I read it as pswrite | 16:11.41 |
paulgardiner | henrys: hang on, I'll just get my iPad | 16:11.46 |
kens | Anyway, we don't support it | 16:11.46 |
marcosw | I know, that's why I said "adding support". | 16:12.05 |
kens | We deliberately don't support most of the page requests. | 16:12.39 |
marcosw | yes, I know we don't support it, I'm asking about adding support (I'm the one who just got up yet appear to be the one who is awake). | 16:13.09 |
paulgardiner | henrys: in drop box tap a doc, then the icon top right consisting of an arrow coming upward out of a box. then select "open in" | 16:13.18 |
kens | The outpout is intended to be device independent. And I don't know any way t map PCL trays into PostScript ones | 16:13.18 |
| marcosw end of a long day | 16:13.46 |
henrys | paulgardiner: ship it ;-) | 16:14.49 |
ray_laptop | we have over 350G of testdocs so far (still dribbling in) | 16:15.12 |
marcosw | presumably one could specify the output tray on the command line, using the -dMediaPosition option, the same was as is done for other output devices. | 16:15.38 |
paulgardiner | henrys: :-) | 16:16.49 |
kens | marcos how would that work with PCL ? | 16:16.55 |
marcosw | It wouldn't, I didn't say anything about PCL. I'm asking about ps2write. | 16:17.20 |
kens | but the input is PCL ? | 16:18.22 |
Robin_Watts | marcosw: Presumably kens is talking about using the pcl exe set to use ps2write out? | 16:18.24 |
marcosw | nope. | 16:18.34 |
kens | What is the input then ? | 16:18.41 |
henrys | the input is avadhut what more needs to be said? | 16:19.07 |
marcosw | kens: I believe GhostScript can only read PostScript and PDF, so presumably it's one of those. | 16:19.08 |
kens | marcosw if the input is PostScript what's the point ? If its PDF then the PDF interpreter does a setpagedevice which I htink may well knowck out the media position | 16:19.41 |
marcosw | kens: the customer is doing PS->PS to embed fonts that aren't present on the printer. | 16:24.44 |
kens | marcosw then he should make his PostScript better in teh first place | 16:25.32 |
paulgardiner | henrys: one point I didn't quite get: should I go ahead with the office integration work proper, or do we want to look at applying your trick to the iOS app? My guess is we just need a demo and the Android app gives us that, so no need for another platform, but maybe there are other issues. | 16:25.45 |
marcosw | kens: I don't think it's his postscript, I think he's being sent the postscript from someone else. | 16:26.04 |
| kens: never mind, I'll just open a bug for the feature and assign it to you. you can close it and I'll have the customer contact miles who'll discuss it with henrys, they'll decide if it's worth doing and either the customer will be told to go away or the bug will be reopened (it's beginning to feel like Artifex is a big company). | 16:26.09 |
kens | Its a dumb request, andf the device is designed to discard that kind of information | 16:26.32 |
henrys | paulgardiner: office integration proper | 16:26.32 |
kens | People like it that way marcosw | 16:26.39 |
paulgardiner | henrys: okay thanks. | 16:26.46 |
chrisl | kens: I've rejigged the history on the xpswrite branch a little - when you have a moment: http://tinyurl.com/o5heoxo | 16:29.22 |
kens | ok | 16:29.42 |
chrisl | So, your commit adds the code, but the code isn't built. Mine adds the build changes etc. That way we keep all the commits building properly | 16:30.31 |
henrys | marcosw: it would be easier for us all to look at if their were a bug. | 16:30.50 |
kens | makes sense little busy right now, will look soon | 16:30.53 |
chrisl | kens: no rush | 16:31.04 |
Robin_Watts | henrys: I wondered whether it was worth me doing a bit more work on the office stuff. | 16:32.48 |
| The current "convert office to PDF" exe has capabilities we aren't going to use, like the ability to render to bmp etc. | 16:33.31 |
| Is it worth me spending some time paring that down? It would be orthogonal to Pauls work. | 16:34.01 |
henrys | Robin_Watts: I think that the high priority thing mupdf-wise. | 16:34.13 |
Robin_Watts | It all depends if we think that there is a separate market for an "office to pdf" converter | 16:34.37 |
| A side effect of the work would be to find out exactly what wasp entry points will be needed to be reimplemented for the mupdf integration work. | 16:35.11 |
| (i.e. what path construction routines are used). | 16:35.23 |
henrys | Robin_Watts: I never saw it more than a stopgap until proper integration could be done. | 16:35.56 |
Robin_Watts | So there is no market for a converter than could be sold to printer companies that have their own PDF solution? | 16:36.41 |
henrys | Robin_Watts: Yes but the "proper" integration, as we're calling it, will do that. | 16:41.48 |
Robin_Watts | will it? | 16:42.14 |
henrys | Robin_Watts: wouldn't it make more sense for you to just work with paulgardiner on the "proper" integration? | 16:42.20 |
Robin_Watts | Only if we get pdfwrite finished. | 16:42.27 |
henrys | Robin_Watts: okay I missed something I thought the proper integration would include "SO's" pdf generation | 16:44.19 |
Robin_Watts | SO does pdf generation by converting from Edr to PDF. | 16:44.58 |
henrys | Robin_Watts: right I see that now. | 16:45.10 |
Robin_Watts | (Well, more technically from DisplayList to PDF, where DisplayList is a structure derived from doing layout on Edr) | 16:45.40 |
henrys | Robin_Watts: do we know how the 3rd party print module works - the one miles talks about from China? | 16:46.30 |
Robin_Watts | henrys: I assume it's bitmap in -> printer data. | 16:46.50 |
| i.e. converts rendered RGB data to data to send to a range of printers. | 16:47.11 |
henrys | Robin_Watts: so you're thinking we can bolt this demo on the side and use it whenever pdf generation is needed? | 16:48.53 |
Robin_Watts | I'm saying we could take the exe that the demo uses, and pare it down a bit and we'd get a standalone 'office to PDF' converter. | 16:49.43 |
| And we could potentially license that to people that had an existing PDF solution, but wanted to add office format support. | 16:50.12 |
| If it was me doing the integration with MuPDF, I'd do this as a first step, but I know that paulgardiner disagrees with me on this approach. | 16:50.53 |
| My approach would be to pare it down as much as possible staying within the epage system where we have something that we know builds and can be tested. | 16:51.22 |
| then make the jump to mupdf. | 16:51.30 |
henrys | Robin_Watts: that does make sense but we really do want pdf export in the final app, right? That seems almost necessary. | 16:51.49 |
Robin_Watts | paulgardiner favours pulling stuff into the mupdf world and getting it building there as he goes. | 16:52.13 |
| pdf export would be nice to have, yes. | 16:52.22 |
henrys | Robin_Watts: in the meantime though what you are suggesting makes senst. | 16:53.50 |
| s/senst/sense | 16:53.57 |
| are you doing JNI bindings? | 16:54.11 |
Robin_Watts | At the moment, no. I'm rejiging the internals of MuPDF to allow us to 'filter' operator streams in a cunning manner. | 16:54.39 |
| It should open up the way to being able to do bates numbering etc. | 16:54.50 |
| as well as lots of other nice things. | 16:54.58 |
henrys | Robin_Watts: raed | 16:55.26 |
Robin_Watts | yeah. | 16:55.32 |
henrys | Robin_Watts: so split time between that and paring down the down is fine by me. | 16:56.50 |
| s/the down/down | 16:57.05 |
Robin_Watts | ok. | 16:57.05 |
henrys | or something like that ;-) | 16:57.13 |
| Robin_Watts: I am sort of wondering if we aren't going to end up bring ghostscript into this in some manner, you get pretty good printer coverage with just ps2write and pxlmono, and pxlcolor. | 17:22.46 |
Robin_Watts | henrys: So you want the office code to be able to feed into gs as well as MuPDF? | 17:23.44 |
| Actually, mooscript might be the answer there. | 17:24.30 |
henrys | Robin_Watts: I was thinking we'd just output pdf and gs would take it from there. | 17:26.35 |
Robin_Watts | That still requires working pdf export. | 17:26.59 |
henrys | Robin_Watts: yup | 17:27.07 |
| Robin_Watts: for example we have a customer now that has an extensive printing setup with ghostscript on the cloud. They want to see the office stuff - here it make sense to use the demo we have now to just send pdf to the cloud. | 17:28.13 |
Robin_Watts | henrys: I would have thought it made sense to sell them the office -> pdf converter to run on the cloud. | 17:28.49 |
henrys | Robin_Watts: yes that works too. | 17:29.10 |
| presumably you want to view the pdf on your phone first so you've done the conversion already assuming we just have the demo | 17:30.02 |
kens | is heading off, had enough for today | 17:30.34 |
| Goodnight all | 17:30.39 |
henrys | bbiab | 17:32.51 |
| paulgardiner: did mvrhel_laptop get the demo working | 18:01.06 |
| ? | 18:01.07 |
mvrhel_laptop | henrys: I have to charge up the tablet. I will try in a bit | 18:02.12 |
henrys | mvrhel_laptop: okay we have a meeting sooner and I have to get miles to install it and I wanted to make sure it worked for you first. | 18:03.08 |
mvrhel_laptop | oh ok. I will let you know in a bit. just plugged it in... | 18:03.38 |
| and it was very dead | 18:03.50 |
henrys | marcosw: are you going to be near miles in the near future? | 18:07.12 |
ray_laptop | moving the space_params to the gx_device struct is a real pain :-( | 18:28.49 |
henrys | paulgardiner: is the apk on casper? | 18:32.31 |
Robin_Watts | henrys: yes, it's in his home dir. | 18:33.23 |
henrys | now how do I get it to him. If I put it in a password protected html link will android do the right thing when the browser accesses it? | 18:39.10 |
Robin_Watts | probably | 18:40.02 |
henrys | mvrhel_laptop: I sent you an email about the apk. | 19:34.13 |
mvrhel_laptop | henrys; ok | 19:34.22 |
| let me give it a try | 19:34.41 |
henrys | htaccess is tits on a bull without ssl so I'll tear it down as soon as you're done | 19:35.21 |
mvrhel_laptop | ok is working but you would want to give him casper.ghostscript.com/~henrys/.apk as the address | 19:36.07 |
| let me put the apk on the device now | 19:36.16 |
henrys | mvrhel_laptop: definitely wonder how that happened I thought I copy pasted. | 19:38.03 |
mvrhel_laptop | henrys: the charger I had plugged was crap. switched chargers give me a few more minutes | 19:39.52 |
| although the charger says 1.8A should be plenty | 19:40.30 |
Robin_Watts | henrys: want to give me a URL? I'll try it on the transformer prime. | 19:41.11 |
henrys | no problem, I told miles to delay the trip to next week. I'd like to bash on this a bit anyway | 19:41.17 |
mvrhel_laptop | there it is starting up | 19:41.30 |
henrys | Robin_Watts: I meant that htaccess uses a plaintext password for logins. | 19:42.07 |
Robin_Watts | henrys: right, I understood that. | 19:42.21 |
henrys | oh I completely misunderstood you I thought transformer prime was some system to give me a secure password - wow ⦠I'll send you the email | 19:43.26 |
mvrhel_laptop | now let me grab a word doc | 19:44.28 |
| it is working. | 19:46.15 |
| some font issues with the simple doc I grabbed though | 19:46.25 |
| oh it did not handle exponents and sub indexes | 19:47.18 |
| hey the cat is making pages turn with her paws | 19:47.35 |
henrys | mvrhel_laptop: is there something strange miles won't understand installing this? How did you get a file to the app? | 19:48.31 |
mvrhel_laptop | henrys: so we will need to pick our documents carefully at this point | 19:48.35 |
henrys | I guess I should dig out my android too. | 19:48.44 |
mvrhel_laptop | henrys: I just dragged a file from my laptop into the DCIM folder | 19:49.10 |
| If he plugs in the tablet into his computer it should be pretty simple | 19:49.29 |
ray_laptop | mvrhel_laptop: DCIM as where pictures go ? | 19:49.59 |
mvrhel_laptop | yes | 19:50.04 |
| ray_laptop: well, no. there is a Pictures folder and a DCIM folder on the nexus | 19:50.24 |
henrys | we can add test files to the http directory, which would make it much easier | 19:50.29 |
mvrhel_laptop | henrys: yes | 19:50.38 |
| so when do we start posting rendering bugs for this stuff? | 19:51.21 |
Robin_Watts | works for me too. Neat. | 19:51.26 |
mvrhel_laptop | and who gets assigned them... | 19:51.30 |
Robin_Watts | mvrhel_laptop: It would be conversion bugs, rather than rendering bugs. | 19:51.40 |
mvrhel_laptop | Robin_Watts ?? | 19:51.41 |
Robin_Watts | and I have my finger on my nose. | 19:51.57 |
mvrhel_laptop | I don't think that works in this case for you Robin_Watts | 19:52.17 |
Robin_Watts | It does, cos paulgardiner worked for Picsel more recently than me. And I don't do fonts. etc. | 19:52.46 |
mvrhel_laptop | lucky paulgardiner | 19:53.18 |
Robin_Watts | but if you find bad files, let me know. We can see if they load properly into SmartOffice or not. | 19:53.30 |
| That way we can see if it's just the pdf export that's wrong. | 19:53.38 |
mvrhel_laptop | Robin_Watts: ok. I have one file that is just text | 19:53.44 |
| here | 19:53.46 |
| shall I send it to you? | 19:53.52 |
| or do we want to open bug tracking? | 19:54.06 |
henrys | mvrhel_laptop: assign them to miles | 19:54.15 |
| seriously I don't think we need to do that right now, presumably marcosw is going to go through a bunch of stuff - hopefully sorted by functionality from Quality Logic. | 19:55.17 |
| so we don't have to dive into fixing wild files first | 19:55.30 |
| need to "bug track" that is | 19:56.09 |
| so now to find interesting examples that work. | 19:56.58 |
| have you guys used the android emulator? Is it worth installing? | 20:02.09 |
Robin_Watts | henrys: I have used it. Don't bother installing it. | 20:03.02 |
| It claims to support a particular type of FP, and then crashes when it's used. | 20:03.21 |
| so it dies with mupdf. | 20:03.28 |
| cos of v8. | 20:03.34 |
henrys | be interesting to try it again with torscript | 20:08.02 |
jo0nas | what is licensing of toolbin/halftone/ETS/* ? | 20:13.29 |
| that dir includes a COPYING.txt file but none of the other files seem to refer to that, and README.txt explicitly says "This version of ETS is for commercial licensees" | 20:14.34 |
henrys | mvrhel_laptop: we did have it on the tech agenda we wanted a GPL release of ETS, I can't remember if something fouled that up, see jo0nas queston | 20:30.46 |
jo0nas | is preparing a future upgrade of Ghostscript officially in Debian, and if not (better) clarified will need to strip ETS :-( | 20:32.13 |
| (I say "future upgrade" because that upgrade waits for upgrade of LCMS2 to be done first) | 20:33.10 |
henrys | jo0nas: I agree it needs to be fixed. | 20:34.18 |
mvrhel_laptop | henrys: so I don't see why ETS would not be GPL | 21:15.14 |
| bbiab | 21:18.32 |
| henrys: are you there? | 23:15.46 |
henrys | hehe | 23:15.54 |
| they wanted it virtual I didn't have anything to demo with. | 23:16.24 |
| mvrhel_laptop: but yes I'm here | 23:18.45 |
mvrhel_laptop | henrys: yes. so I don't think we are doing the video recording like Miles suggested are we? | 23:22.24 |
| I guess I need to ask him | 23:22.31 |
| henrys: what versions of windows office docs can we handle? | 23:23.26 |
henrys | mvrhel_laptop: I don't know of limitations | 23:24.03 |
mvrhel_laptop | Oh. ok. I may give it a try to see. | 23:24.17 |
henrys | mvrhel_laptop: we really do need to start beating on this but marcosw was going to take the lead on that. | 23:24.43 |
mvrhel_laptop | right. I just don't want to find myself being asked a bunch of questions that I don't know the answer to so I will probably start asking them myself and playing around with it | 23:25.25 |
henrys | mvrhel_laptop: makes sense we've been clear we've just acquired it and it's a work in progress | 23:26.40 |
| mvrhel_laptop: be back in a bit. | 23:26.54 |
mvrhel_laptop | ok | 23:26.56 |
| Forward 1 day (to 2014/01/29)>>> | |