| | <<<Back 1 day (to 2012/11/04) | 2012/11/05 |
| mvrhel_laptop | Robin_Watts_: For the logs. No. They have not fixed that. But it is basically one key to get from the tile screen to the desktop and there are a pile of keyboard shortcuts to use. I do still find myself occasionally looking for the start button but with looking at a large graphical list on the screen vs. a written list from the start button to find something is not too bad | 04:50.34 |
| Robin_Watts_ | tkamppeter_: Hey. Are you around ? | 11:32.06 |
| emi | hello anybody! | 13:35.53 |
| Robin_Watts_ | hello emi | 13:36.03 |
| emi | I hope someone can help me... | 13:36.53 |
| Chumper | HI guys, i am currently deeping into the android sourcecode of mupdf. I want to be able to edit/add annotations to a pdf, therefore i looked into the c source and on stackoverflow.com. http://stackoverflow.com/questions/11396951/editing-a-pdf-with-mupdf I found the pdf_annot.c there are some functions to handle annotaions but are there more functions elsewhere that i am missing? Do i still need to create the objects with the spe | 13:36.56 |
| | dictionary? | 13:37.29 |
| Robin_Watts_ | Chumper: Are you looking at git source or the last release ? | 13:38.20 |
| Chumper | I am looking at the git source | 13:38.36 |
| Robin_Watts_ | ok. | 13:38.47 |
| emi | I'm using gs to convert .PS to PDF and it works very well.... but i need to convert at the same time the VECTORIAL PDF to RASTER PDF. Any idea?? | 13:38.54 |
| Robin_Watts_ | So at the moment, mupdfs public API for handling annotations is pretty much just enough to load them and display them. | 13:39.49 |
| emi | .PS is obtained by a parametric CAD. I'have some tecnicals drawings | 13:39.52 |
| Robin_Watts_ | With some extras in there for form handing. | 13:39.56 |
| | If you're wanting to manipulate the annotations at the PDF object level (create new ones etc), then you need to do that yourself, by calling the PDF object functions within mudfp. | 13:40.45 |
| Chumper | Ok, sounds good, so to be able to add annotations it is all on me | 13:40.46 |
| | ok, just want to make sure i don't miss anything | 13:41.02 |
| | thanks | 13:41.04 |
| emi | No Robin, i don't need to visualize pdf because i use acrobat reader | 13:41.05 |
| Robin_Watts_ | Chumper: Let me know how you get on. | 13:41.20 |
| Chumper | all right :) | 13:41.32 |
| Robin_Watts_ | emi: Sorry, none of my answers fo far were to you :) | 13:41.35 |
| emi | in a rasterized pdf when you zoom +++ you cannot see very well all the lines and i like this... | 13:41.52 |
| Robin_Watts_ | emi: Essentially, if I understand you correctly you want to go from a PDF full of 'move here, draw line to here' etc, to a PDF that's really just a wrapped around a bitmap image ? | 13:42.35 |
| kens | We don't make that kind of PDF | 13:43.22 |
| emi | robin: yes, you've understand what I mean!! | 13:43.36 |
| Robin_Watts_ | kens: We could, if he pretended to have some transparency in there :) | 13:43.42 |
| kens | Would still have to set compatibility tto 1.3 | 13:43.56 |
| | at default level we just preserve transparency | 13:44.12 |
| Robin_Watts_ | emi: We go to some lengths to AVOID doing that sort of step, specifically because it loses information. | 13:44.22 |
| | The one case where we DO have to drop back to making PDFs with just an image in there is when we think that a) there is transparency in use in the file, and b) we are outputing to a version of PDF that doesn't support transparency. | 13:45.13 |
| emi | robin: I need to protect their information design!... | 13:45.43 |
| | and drawings cannot be measured... | 13:46.04 |
| kens | I can measure a bitmap easil | 13:46.18 |
| | Acrobat even has tools for that | 13:46.30 |
| emi | kens: but is not very simply, consider to measure 0.05mm or 0.5 degree... | 13:47.06 |
| Robin_Watts_ | emi: So essentially you're looking at doing this so as to deliberately produce a "degraded" output and somehow preserve the value of the original. | 13:47.08 |
| emi | robin: YES!!! | 13:47.24 |
| | this problem is felt by meny designers and engineers... | 13:47.59 |
| kens | I don't think so, first time I've heard of it | 13:48.12 |
| Robin_Watts_ | So, short answer: The existing code will not do this for you, but we could probably make it do so without too much work. But that's the kind of work we wouldn't undertake ourselves, unless it was for a paying customer. | 13:48.19 |
| | kens: I think that's a fair statement, right? | 13:48.37 |
| kens | And we still don't make that kind of PDF unless we have absoluytely no choice | 13:48.40 |
| | Robin_Watts_: q it would be quite a lot of work | 13:48.55 |
| | IOI would not start form pdfwrite | 13:49.01 |
| | Use a raster driver and wrapu up the output in P{DF'ism | 13:49.30 |
| Robin_Watts_ | Couldn't we have a -dForceBitmap=1 thing that forced the transparency detection to 1 ? | 13:49.35 |
| kens | HGorrible idea, I wqon'rt hear of it | 13:49.46 |
| Robin_Watts_ | emi: OK. make that a *highly* paying customer then :) | 13:50.32 |
| kens | If you want this, start with a raster driver, and go from there, do *NOT* attempt to do this with pdfwrite | 13:50.38 |
| | We did dicsuss this for PCL ROPS if you recall, and that was my answer then, I stan dby it | 13:51.08 |
| | I reckon that's 3 days to 2 weeks for someone who knows what they are doing | 13:52.19 |
| emi | i could convert .PS to TIFF and so .TIFF to PDF but I don't like this... | 13:52.21 |
| kens | Its exactly what you are describing, why do you not like it ? | 13:52.38 |
| emi | because I've to menage many files and i've to create a batch procedure.. | 13:53.39 |
| kens | Seems to me that this is easily done as a btach prcess | 13:53.54 |
| emi | but with gs can i convert ps to TIFF? | 13:56.03 |
| kens | Yes | 13:56.08 |
| | many ways | 13:56.11 |
| | and many, many options | 13:56.20 |
| emi | sorry but i'm a semi-beginner... | 13:58.01 |
| | there is not a command .bat in GS lib... | 13:58.36 |
| kens | No, correct. | 13:58.45 |
| | You will have to learn to use the GS command line | 13:58.52 |
| emi | or better.. i don't find it.. | 13:58.57 |
| | ah! | 13:59.07 |
| kens | there are numersou tiff devices producing monochrome, grey, RGB or CMYK tiffs, you need to specify the resolution and for some of the tiffs you can specify whether you want nicely anti-aliased output and so on | 13:59.33 |
| Robin_Watts_ | gswin32c.exe -sDEVICE=tiff32nc -o out.%d.tif -r300 -o input.pdf | 13:59.34 |
| | or maybe: | 13:59.47 |
| kens | 2 -o in there Robin_Watts_ | 14:00.00 |
| Robin_Watts_ | gswin32c.exe -sDEVICE=tiffscaled24 -dDownScaleFactor=3 -o out.%d.tif -r900 input.pdf | 14:00.20 |
| | kens: oops, yes. emi: Ignore the second -o in that first line. | 14:00.32 |
| emi | a moment... | 14:04.47 |
| | thanks, I try your solution | 14:05.35 |
| Robin_Watts_ | The second one will give nicer results. | 14:05.50 |
| emi | ok | 14:06.09 |
| | sorry but imput must be .PS and not .PDF | 14:09.16 |
| Robin_Watts_ | emi, that's fine, makes no odds. | 14:10.04 |
| emi | a question | 14:12.42 |
| | out.%d.tif maybe for example test.tif or out.testd.tif? | 14:13.09 |
| Robin_Watts_ | out.%d.tif means put the first page as "out.1.tif", the second as "out.2.tif" etc. | 14:14.09 |
| | Obviously you can change that string as you'd prefer. | 14:14.19 |
| emi | i've only-always 1 page | 14:14.40 |
| | ah ok | 14:14.53 |
| Robin_Watts_ | Then -o whateveryoulike.tif | 14:14.53 |
| emi | mmm... i've load this command line in DOS and GS starts but not go next... | 14:21.59 |
| | OK, it works very welllllll!!!! | 14:24.31 |
| | and so.... to reconvert this tiff in pdf? | 14:25.13 |
| kens | We don't have any way to do that, use ImageMagick | 14:25.52 |
| emi | uff!!...... | 14:26.55 |
| chrisl | I suspect ImageMagick uses the libtiff tiff2ps tool, and then Ghostscript to go PS to PDF..... | 14:27.08 |
| emi | but for money, is there anyone who can give me a command .PS (vectorial) to PDF (raster)? | 14:27.46 |
| kens | POssibly, nobody I know | 14:28.14 |
| emi | i don't wanto to have to do .ps > .pdf > .tif > .ps > .pdf !!!!!!! | 14:28.49 |
| chrisl | Almost everybody will do everything possible to *avoid* that kind of result | 14:28.52 |
| | ("raster" PDF, that is) | 14:29.02 |
| kens | we've had that discussion already.... | 14:29.11 |
| chrisl | emi: you don't have to do ps->pdf->tif, you can go ps->tif | 14:29.37 |
| kens | Yes, he wants a PDF wiht a bitmap in it | 14:29.48 |
| | Starting form PS | 14:30.02 |
| chrisl | Yes, and in the flow above the first step emi lists is ps->pdf | 14:30.26 |
| | emi: btw, libtiff comes with a tiff2pdf tool, too | 14:30.52 |
| | So, ps->tiff->PDF | 14:31.03 |
| emi | ok chrisl, you are right | 14:31.48 |
| | but i dont'like to use another program command line to convert tiff to pdf... | 14:32.17 |
| | other idea? | 14:32.23 |
| kens | GS won't do it so you will have to use 2 progtams | 14:32.38 |
| chrisl | ImageMagick will probably "hide" all the steps from you (but still using GS and libtiff), but you'll get less control, and we won't be able to help if/when it goes wrong | 14:33.34 |
| emi | thanks kens, robin and ghrisl! | 14:34.06 |
| kens | You're welcome | 14:34.14 |
| emi | you were very clear! | 14:34.21 |
| | now I will work to create the batch procedures necessary | 14:34.26 |
| | good job! | 14:34.42 |
| Robin_Watts_ | tkamppeter_: ping | 15:19.46 |
| chrisl | Robin_Watts_: You probably want to wait for the "_" postfix to disappear.... | 15:21.46 |
| henrys | now our clocks have changed we are sync'd up again. | 16:07.09 |
| paulgardiner | Good. So back to 4pm tomorrow for the forms meeting. | 16:19.48 |
| henrys | yes it is 8:20 now Pacific. | 16:21.03 |
| chrisl | henrys: oh, on the subject of meetings, I'll very likely have to miss the 9am/5pm IRC meeting tomorrow - I will check the logs, but if anything important comes up for me, if someone could ping me an e-mail. | 16:25.10 |
| henrys | that's fine - I was going to ask you about freetype, have a real date yet? | 16:25.52 |
| | a change like that I'd like to have in as far before the release as possible. | 16:26.17 |
| chrisl | Well, the thing that's holding me up just now is the artificial boldening, I'm pretty happy with the rest just now | 16:26.42 |
| henrys | XL, XPS or both? | 16:27.44 |
| chrisl | XL, Tor did a higher level solution for XPS, so I'm not too worried about that right now | 16:28.11 |
| henrys | yeah I wonder if Tor's code would work with XL and we could toss the smearing bitmap stuff. | 16:29.30 |
| | ? | 16:29.33 |
| Robin_Watts_ | We do artificial boldening by rendering the glyph and then smearing the image? | 16:30.18 |
| chrisl | It's certainly a *much* nicer approach than the bitmap smearing, yes. In both cases, though, we lose cacheing | 16:30.21 |
| henrys | plchar.c:bits_smear_horizontally Robin_Watts_ | 16:31.49 |
| Robin_Watts_ | implemented 'glow' for Picsel by smearing and post processing. | 16:31.58 |
| chrisl | Robin_Watts_: in XL, yes. In XPS, Tor calculates a stroke width for added "boldness" and fills then strokes the glyph | 16:32.36 |
| mvrhel_laptop | good morning | 16:32.50 |
| henrys | hi mvrhel_laptop | 16:32.59 |
| Robin_Watts_ | Interesting. | 16:33.00 |
| | mvrhel_laptop: Morning | 16:33.04 |
| | mvrhel_laptop: Can we set some time aside to run through this ETS stuff? (assuming you're free of customer panics etc) | 16:33.29 |
| mvrhel_laptop | Robin_Watts_: yes that would be fine | 16:33.49 |
| Robin_Watts_ | When would be good for you? | 16:33.57 |
| henrys | chrisl:so we don't have a good way to get a bitmap back from freetype? | 16:34.06 |
| mvrhel_laptop | Robin_Watts_: Do you want to plan on right after the Tuesday meeting? | 16:34.27 |
| | That will give me a little time to go over it better | 16:34.34 |
| Robin_Watts_ | Sounds OK to me. | 16:34.48 |
| chrisl | henrys: not like the current image smearing code relies on, no | 16:35.04 |
| henrys | I'd vote for getting the code in with it broken then I'll take it upon myself to play with the line width hack in XL. | 16:36.32 |
| chrisl | How often is the boldening "feature" used with Truetype fonts? | 16:37.06 |
| henrys | I have never seen a wild xl file with boldening so it isn't much of a risk. | 16:37.09 |
| | I will need to fix it before the release though. | 16:37.50 |
| chrisl | Do you know, off the top of your head, a simple test file with boldening in it? I'd like to see what the Freetype boldening is like | 16:39.08 |
| henrys | yeah I'll find it one minute | 16:39.27 |
| | pxlfts3.0/T311.BIN panel 01 | 16:42.20 |
| chrisl | Cool, thanks | 16:42.30 |
| | henrys: Freetype on left, AFS on right (obviously some fettling required, like not bolding char paths): http://www.ghostscript.com/~chrisl/ft_left-afs_right.png | 17:10.46 |
| henrys | if there are real files in the wild I imagine using freetype would go unnoticed. Panel 01 won't get passed most of our customers though. That's quite different. | 17:15.45 |
| Robin_Watts_ | Looks to me like AFS doesn't increase the bbox for the glyph when it boldens it. | 17:15.58 |
| henrys | right there should be a clip. | 17:16.33 |
| Robin_Watts_ | whereas freetype does. | 17:16.34 |
| henrys | the HP output is very similar to ours. | 17:16.48 |
| Robin_Watts_ | chrisl: Sorry, you said earlier that we couldn't use our existing boldening code, because what we got back from freetype was in the wrong format or something? | 17:18.03 |
| chrisl | Robin_Watts_: the current code doesn't really get a bitmap from the font world, it gets an outline from the font world, and renders that to a mem device, then retrieves the bitmap from the mem device. | 17:19.21 |
| Robin_Watts_ | ok, but the boldening code itself is actually just a process that is performed on a 1bpp bitmap, right? | 17:19.52 |
| chrisl | Yes, my fallback plan is to pull that code into FAPI, and do the smearing en-route from Freetype back into GS | 17:20.43 |
| Robin_Watts_ | (hmm, badly expressed). Our boldening code can be seen as a process that gets the outline, renders it to a mem device, and then calls a 'boldening core' that takes a bitmap -> a bitmap. | 17:21.00 |
| | Right. | 17:21.00 |
| | That fallback plan sounds to me like it should give us a much closer match to the old code. | 17:21.30 |
| kens | Off to sort out dinner, goodnight asll | 17:21.50 |
| Robin_Watts_ | night kens. | 17:21.59 |
| chrisl | Bye kens | 17:22.01 |
| | Robin_Watts_: yeh, I had hoped that boldening the outline, rather than the bitmap smearing would get a nicer result, but there's nothing in it,really | 17:22.39 |
| Robin_Watts_ | What is the downside to that fallback plan? (i.e. why is that a fallback plan, rather than the primary plan) | 17:22.42 |
| | chrisl: Right. | 17:22.49 |
| chrisl | see above ;-) | 17:22.51 |
| Robin_Watts_ | :) | 17:22.55 |
| chrisl | The advantage of the bitmap smearing in FAPI approach is that it is scaler agnostic - we're not relying on a feature of the font library | 17:23.50 |
| spacebarbarian | converting a pdf to jpeg is there a built in way to multithread gs ? | 17:56.23 |
| Robin_Watts_ | spacebarbarian: Yes. | 18:00.00 |
| | The interpretation can't be multithreaded, but the rendering can be. | 18:00.17 |
| | It also depends on the resolution being high enough to not run in page mode, AIUI. | 18:00.39 |
| | gs -sDEVICE=jpeg -o out.jpg -dNumRenderingThreads=4 in.pdf | 18:01.03 |
| spacebarbarian | oh, i meant having different page conversions be assigned different threads, would multi threaded rendering perform better ? | 18:03.09 |
| | im not rendering very high res, final image is just 1280x900 | 18:03.54 |
| Robin_Watts_ | spacebarbarian: No, we don't currently support different page conversions having different threads. | 18:15.22 |
| | and indeed such a thing would present problems for our device interface. | 18:15.49 |
| spacebarbarian | ah ok would you recommend doing the page level multithreading over the rendering flag ? | 18:16.25 |
| Robin_Watts_ | spacebarbarian: Well, your choices are "page level multithreading which isn't implemented, and so won't work" or "multi threaded rendering which should work, but may not give you as much speedup as you'd like". | 18:18.46 |
| | spacebarbarian: If it was me, and I had many files to process, I'd probably just process them several at a time. | 18:19.11 |
| spacebarbarian | Robin_Watts_, i meant i would just scrip[t it to do page level | 18:19.15 |
| | thanks for the help ill just time both methods see which does better :) | 18:21.03 |
| mvrhel_laptop | bbiab | 18:28.35 |
| chrisl | henrys: at least a first pass (again, FAPI on the left): http://www.ghostscript.com/~chrisl/fapi_with_boldening.png | 19:44.30 |
| Gigs- | What was the name of the visual diff program I used for regression testing a while back? | 21:38.14 |
| Robin_Watts_ | bmpcmp ? | 21:38.40 |
| Gigs- | that was probably it thanks | 21:39.11 |
| Robin_Watts_ | or cmpi? or compare ? | 21:39.12 |
| mvrhel_laptop | bbiaw | 21:40.12 |
| Gigs- | yeah it was bmpcmp thanks | 21:40.21 |
| | Forward 1 day (to 2012/11/06)>>> | |
IRC Logs
