| <<<Back 1 day (to 2012/11/07) | 2012/11/08 |
mvrhel_laptop | henrys you there? | 05:40.19 |
| oh never mind I figured it out | 05:41.18 |
Robin_Watts_ | kens: Scott is on a plane today. | 10:57.24 |
| Coming to London. | 10:57.56 |
kens | Robin_Watts_ : I didn't realise that | 10:58.17 |
| At least my reply tells them something is happening | 10:58.42 |
| chrisl Robin_Watts_ have either of you applied for an ESTA recently ? | 11:12.53 |
chrisl | kens: no, I think I did mine at the beginning of the year | 11:17.11 |
kens | Its OK, I found the problem. I had scripting blocked on the site. It was asking for my 'personal identification number' on my passport, and I don;t have one.... | 11:17.40 |
| WShy allowing scripting gets rid of that I have no idea | 11:17.59 |
chrisl | region checking? | 11:18.17 |
kens | Perhaps. | 11:18.28 |
| Of course I now get to fill out the form again, as it forgot all my entries when I turned on scripting.... | 11:18.52 |
correct_doc | for the pdfwriter option ConvertImagesToIndexed the doc says unsupported but when used the resulted PDF is different ? | 11:22.06 |
kens | And what is the question ? | 11:22.57 |
correct_doc | this is a bug in doc . | 11:23.26 |
kens | No | 11:23.31 |
| The option is not supported, the documentation is correct | 11:24.06 |
correct_doc | but it does something | 11:26.34 |
kens | What exactly ? | 11:26.43 |
chrisl_r61 | Cor, useful information, there..... | 11:46.47 |
kens | Detailed and insightful | 11:47.51 |
| I *know* pdfwrite doesn't convert images to indexed colour spaces, its on my todo list.... | 11:48.31 |
chrisl_r61 | I *could* be an indeterminism caused by the length of the command line - but with no information, who cares....? | 11:49.23 |
| s/I/It | 11:49.31 |
kens | Its probably the fact that the date and timew stamp are different. A simple compare or dif or MD5 checksum will show the files as 'different' | 11:50.11 |
| But since he offered no evidence, I don't really see a reaosn to pursue it | 11:50.49 |
chrisl_r61 | Yes, could be that. Actually, it made me wonder - how difficult would it be to dump the pdfwrite configuration as metadata into the output file? | 11:51.01 |
kens | Umm. | 11:51.19 |
| We could put it in the PostScript dict easily enough as a private key, but the XML is governed by the schema and thre's no provision in that for that kind of information | 11:51.50 |
| Also it would make the PDF file bigger (admittedly only a little bit) | 11:52.06 |
chrisl_r61 | Ah, I thought there was a metadata type that was just arbitrary data | 11:52.29 |
kens | In XML ? Yes, but in the PDF schema, I don't think so | 11:52.47 |
| (I could be wrong) | 11:53.00 |
chrisl_r61 | Well, in that case, we could just use a comment | 11:53.16 |
kens | Yes, we could do that, I don't think it would save us a whole lot | 11:53.42 |
chrisl_r61 | Well, I was thinking of all the times it takes to-ing and fro-ing with the reporter to finally get them to admit the exact command line they use | 11:54.25 |
kens | Yes, but that to-and-fro gives you an instant feel to whether you are dealing with a clueless muppet or a sensible person :-) | 11:54.57 |
chrisl_r61 | Comparing the alleged command line with the configuration read from a PDF would give you that idea pretty quickly, too! | 11:55.38 |
kens | True. | 11:56.13 |
chrisl_r61 | It was that guy with the reusablestreamdecode issue that made me think about it - "oh, but here's command I actually used.... oh, no, I mean this one......." | 11:57.37 |
kens | I guessed that was who you had in mind.... | 11:58.18 |
chrisl_r61 | He didn't fill me with confidence when, *after* I had confirmed I'd found the problem, he posted the Distiller result - gee, thanks..... | 11:59.13 |
kens | Yeah, its amazing how many people do that. And I've got one who keeps posting me the WIndows crash dialog, but won't send me a sample file.... | 12:00.05 |
chrisl_r61 | I saw that. Did you close that bug? At this stage, I think I would..... | 12:00.40 |
kens | Not yet, but I will soon if I don't hear more. Its not actually assigned to me ;-) | 12:01.08 |
chrisl_r61 | I lack sympathy since it looks like Ghostscript is being used by a commercial package (even if it wasn't installed with it) | 12:02.35 |
kens | Yeah I'm suspicous about that too, I think my question made him back off on that. | 12:03.07 |
| No way is it a 'Save As' in Powerbuilder, unless they have spcifically built that in and told people to use GS | 12:03.35 |
chrisl_r61 | Exactly | 12:03.57 |
| kens: http://tinyurl.com/cblrtva | 12:04.54 |
kens | Ha, well there we are then. | 12:05.37 |
| Nice up to date documentation too. | 12:05.54 |
chrisl_r61 | Well, I don't know what PowerBuilder version that's for, it's what google found | 12:06.29 |
kens | I can't believe customers let them get away with that | 12:06.35 |
chrisl_r61 | *Especially* as that is "PowerBuilder Enterprise" | 12:07.08 |
kens | Yeah, I waonder how much that costs ? ..... | 12:07.25 |
chrisl_r61 | According to InfoWorld in 2009: "PowerBuilder (12) costs $2,995 for a new license" | 12:10.31 |
paulgardiner | Robin_Watts_: I've fixed the wrong-page-render bug. | 12:11.06 |
kens | And its PDF support is limited to 'download this open source app form somewhere else, good luck with it' | 12:11.11 |
chrisl_r61 | Currently, a single seat, PowerBuilder 12.5 license is 2,439.80 Euros - and is only available for Windows x86(!!!) | 12:12.10 |
kens | Hmm, a product on its way out ? | 12:12.33 |
| damned expensive too | 12:12.45 |
chrisl_r61 | Oh, and that license price doesn't include support, that'll be another 1,097.80 Euros, please..... ffs! | 12:13.39 |
kens | Hmm gouging at its best | 12:14.03 |
chrisl_r61 | I'm unimpressed, to say the least | 12:14.49 |
kens | I wonder if this is actually a Sybase person | 12:15.36 |
| Its a hotmail account, and he went quite when I inquired about where he got it, then came back 24 hours later with 'it happens at the command line too' | 12:16.07 |
chrisl_r61 | Possibly, that would not explain the failure to provide a test file | 12:16.25 |
kens | Yes, that was another thought I had too | 12:16.53 |
| Well if he doesn't do so soon I will close the report | 12:17.08 |
chrisl_r61 | kens: I think this answers the question: http://tinyurl.com/cqbmx6w | 12:19.26 |
kens | Yeah that pretty much answers the question. I'm going to add a note to that effect in the thread and close it then. | 12:22.54 |
kens | lunches | 12:35.28 |
| chrisl just forwarded you a private mail I receioved from Mr Gogas | 13:22.38 |
Robin_Watts_ | kens: any chance you could copy support? </Nosey> | 13:24.41 |
kens | Hmm, I'll send it to you if you like RObin :-) | 13:26.21 |
| There you go. | 13:26.51 |
Robin_Watts_ | paulgardiner: Those 3 commits all look good to me. | 13:33.22 |
paulgardiner | Great thanks | 13:33.42 |
Robin_Watts_ | tor8: Do you want to move the tag up? | 13:34.15 |
chrisl_r61 | kens: well, at least he is off our backs...... | 13:45.17 |
kens | Yes, looks like it was a hardware problem, you would think he would have tried another computer earlier.... | 13:45.47 |
| And I just checked the tghread and I definitely sadi I couldn't reproduce the problem | 13:46.09 |
paulgardiner_ | Yes. He said we could accept the information in any form, and we can create bugs appropriately. | 13:48.35 |
ray_laptop | that's a strange one. A support inquiry at 2:14AM (PST) from a guy in South San Francisco, CA. Thanks from replying, kens. | 15:14.08 |
kens | His auto-reply said he was travelling (out of office) so I guess he's probably in Europe at the moment | 15:23.47 |
Robin_Watts_ | tor8: ping | 15:32.38 |
ray_laptop | I had sent Scott an email I got that said they could provide a list of IT managers at Fortune 1000 companies. I wonder if he followed up. Enough support-only contracts could be a good revenue stream. | 15:46.05 |
Robin_Watts_ | ray_laptop: ooh, I really hope we don't follow up on that. | 15:46.38 |
ray_laptop | have to restart to install an AV engine update. bbiab | 15:46.56 |
Robin_Watts_ | Giving any business to firms that spam you with offers to sell you other peoples addresses would make me feel very dirty. | 15:47.16 |
paulgardiner | Anyone know why a PDF document might have text widget annotations marked as being hidden? | 15:51.14 |
Robin_Watts_ | And more specifically, why acrobat would show them anyway. | 15:51.44 |
kens | Weel there's no accounting for what Acrobat does | 15:52.00 |
Robin_Watts_ | paulgardiner: RadioCheckBox_AcroForm.pdf has a clear button that doesn't work. | 15:53.17 |
kens | One suggestion would be tghat annotations coujld form a history, when the task is done (editing or whatever) you cna hide tehm. If you ever need to see the history you can (presumabluy) unhide them | 15:53.25 |
Robin_Watts_ | Is it possible that we should make fields 'unhidden' when we generate appearance streams ? | 15:54.05 |
kens | Robin_Watts_ : I don't really feel I know enough about annotations to comment | 15:54.25 |
Robin_Watts_ | kens: This is a file where we can click on an (initially empty) field, and enter a string. | 15:54.31 |
paulgardiner | Robin_Watts_: the docs say I shouldn't even be allowing the user to interact with them. | 15:54.38 |
Robin_Watts_ | Oh. | 15:54.52 |
| What does chrome do? | 15:55.02 |
| Chrome does not allow us to access those fields. | 15:56.16 |
| So I'm tempted to say that chrome is right, and acrobat is just broken. | 15:56.39 |
paulgardiner | Ah ok. Sometime I should change it so that those fields cannot be access probably. | 15:57.29 |
kens | "The Guide is definitive. Reality is frequently inaccurate." | 15:58.05 |
| I expect people will regard Acrobat as correct, and Chrome as broken | 16:00.08 |
| Oh great, zero-day vulnerability in reader X | 16:04.11 |
mvrhel_laptop | uhoh. hope you feel better soon henrys | 16:50.37 |
Robin_Watts_ | ew. | 16:51.03 |
henrys | thanks guys | 16:57.21 |
paulgardiner | Robin_Watts_: Clear button in RadioCheckBox_AcroForm.pdf works now, but I'm waiting on a cluster run to be sure I haven't broken anything. | 17:03.33 |
ray_laptop | I think I may be coming down with what my daughter had :-( Not horrible yet, tho'. | 17:07.11 |
Robin_Watts_ | TTOTD: Avoid using variables containing words you regularly misspell. | 17:08.26 |
kens | goodnight all | 17:11.43 |
mvrhel_laptop | bye kens | 17:11.48 |
paulgardiner | Night kens | 17:12.14 |
Robin_Watts_ | paulgardiner: Gotta be fast to catch kens. | 17:12.48 |
Robin_Watts_ | finishes hacking huge amounts of out ETS, and looks to see if it still works :) | 17:13.16 |
ray_laptop | Robin_Watts_: I wonder if the 16 white lines thing is intentional so that the error term will have reduced to (close enough to) 0. 16 lines may be a bit conservative. | 17:15.07 |
Robin_Watts_ | ray_laptop: Right, but the distance figures won't have decreased to zero. | 17:16.05 |
| They *increase* each time. | 17:16.13 |
ray_laptop | increase to the max (as close as we get to infinite) | 17:16.35 |
Robin_Watts_ | 2*x+1 and 2*y+1 | 17:16.56 |
| so they need only be 33. | 17:17.14 |
mvrhel_laptop | brb | 17:20.50 |
mvrhel | Robin_Watts_ yes that does not sound correct | 17:23.02 |
Robin_Watts_ | balls. I've broken the random noise. | 17:28.19 |
mvrhel | it was already broken :) | 17:29.21 |
| brb again. having to restart each of these machines with the install of windows phone sdk and Hyper-V | 17:30.09 |
Robin_Watts_ | mvrhel_laptop: Yeah, but it's really broken now. I'm getting black pixels in white areas. | 17:40.04 |
ray_laptop | Robin_Watts_: I thought there was code that forced white areas to be white: /* White pixels stay white */ | 18:03.12 |
Robin_Watts_ | ray_laptop: There is, BUT... turns out that white is 0x10000 not 0. | 18:03.30 |
| That should actually be 'black pixels stay black'. | 18:03.41 |
| which means that the code is even further wrong, cos that should be resetting the distance. | 18:04.12 |
ray_laptop | oh | 18:04.14 |
Robin_Watts_ | It surprised me. | 18:04.22 |
ray_laptop | well, with black the expected distance to the next pixel s/b 0, shouldn't it ? | 18:06.00 |
| and it sets dst_ptr[i] = 0 | 18:06.47 |
Robin_Watts_ | Right, but we need to state that we set the pixel. | 18:06.51 |
| Turns out we are working in 'amount of light' not 'amount of ink' in this code. | 18:07.13 |
| oh, boy I hate this code. | 18:27.23 |
| mvrhel_laptop: Are you in a position to spare me a few minutes? | 18:27.40 |
paulgardiner | Robin_Watts_: bmpcmp looks ok | 18:33.07 |
Robin_Watts_ | paulgardiner: Cool. | 18:33.17 |
| So do we think this is the last change required for the release? (famous last words) | 18:33.41 |
paulgardiner | Seems to have fixed another problem that was apparently quite common | 18:34.15 |
| Lefthand side is the new ones, right? | 18:34.40 |
Robin_Watts_ | Yes. | 18:35.04 |
paulgardiner | Yeah, that's fixed a lot of cases where we were getting both of a choice selected. Have to admit I'm not sure why. | 18:36.05 |
mvrhel_laptop | Robin_Watts_: I am here | 18:39.23 |
Robin_Watts_ | OK, pushed. | 18:39.24 |
| mvrhel_laptop: I've just stumbled upon something that has really confused me. | 18:39.40 |
mvrhel_laptop | crap phone hold on | 18:39.50 |
Robin_Watts_ | To the point where I don't understand how this code is working at all. holding. | 18:39.57 |
mvrhel_laptop | ok I am back | 18:44.05 |
| Let me open the project | 18:44.32 |
| Robin_Watts_: ok | 18:45.00 |
Robin_Watts_ | My previous understanding of the code was that we were working in 'amount of ink' in this central loop. | 18:46.17 |
mvrhel_laptop | That was my assumption to as we are looking at the distance to a dot | 18:46.51 |
Robin_Watts_ | I thought that imo == im = 0 meant white | 18:47.03 |
| BUT, I now think that im = levels-1 means white. | 18:47.19 |
| Which throws me into utter confusion. | 18:47.27 |
| sorry, imo == levels-1, im == 65536 means white. | 18:48.16 |
| could I trouble you to sanity check me please? | 18:48.37 |
| (run a known file through, look at the values etc) | 18:48.55 |
mvrhel_laptop | I can try | 18:49.15 |
| what file and what parameters | 18:49.59 |
| should I update first? | 18:50.13 |
Robin_Watts_ | Any file you like, and -m0 -e0 -r0 should be enough to show it. | 18:50.26 |
mvrhel_laptop | ok so I need to make a pam cmyk file I suppose | 18:51.03 |
Robin_Watts_ | examples/tiger.pam ? | 18:51.15 |
mvrhel | oh I did not realize we had the pam in there | 18:51.29 |
Robin_Watts_ | Or your .pgm file should be fine. | 18:51.46 |
mvrhel | ok let me use the .pgm | 18:52.17 |
Robin_Watts_ | Aargh. | 18:53.01 |
| pgms represent white as FF. | 18:53.09 |
| pams as 00 | 18:53.13 |
mvrhel | oh | 18:53.22 |
Robin_Watts_ | So that explains something. | 18:53.23 |
mvrhel | yes | 18:53.26 |
Robin_Watts_ | God knows how stuff has been working then :( | 18:53.55 |
mvrhel | That has been the big question since we dug into this | 18:54.12 |
Robin_Watts_ | ok, so in the light of that, I need to regenerate all those pngs again. | 18:56.07 |
| Thanks. | 18:56.10 |
| I'll do that before I waste any more of your time. | 18:56.21 |
mvrhel | ok. need to call wife back. she got a ticket and apparently I was not sympathetic enough..... | 19:00.56 |
Robin_Watts_ | ah. | 19:08.18 |
| Right, images regenerated. What this shows is that the random noise is REALLY bad :( | 19:27.47 |
| I can really see ETS working in the teeth now. | 19:28.41 |
paulgardiner | Robin_Watts_: not managed to reproduce that bad redraw yet | 19:29.27 |
Robin_Watts_ | on android? | 19:29.47 |
| I had the phone in portrait mode. | 19:29.53 |
| With the page scaled to be about twice the width of the screen. | 19:30.18 |
paulgardiner | Yes, that's about what I have. | 19:30.41 |
Robin_Watts_ | Select "Option 2" and then scroll so the clear button is in the middle of the page. | 19:31.13 |
| then click it. | 19:31.17 |
paulgardiner | Ta. Got it | 19:31.36 |
Robin_Watts_ | Phew | 19:32.00 |
paulgardiner | Ah! Pretty sure I know what it is and how to fix it... I'll leave the disappointment of finding I'm wrong for tomorrow. | 19:37.12 |
Robin_Watts_ | paulgardiner: Ah! Excellent. | 19:37.23 |
| Good plan :) | 19:37.26 |
Ch3rrryC0ke | Hey there-- I'm trying to support printing in a PDF viewer that uses mupdf. The problem I'm having is that even when rendering text documents at a high resolution, when printed, they look aliased/fuzzy. | 19:49.12 |
| My thinking was that perhaps the anti-aliasing mupdf applies looks great on screens but not when printing. Is there an easy way to turn it off? | 19:49.31 |
| Or are there any other tips for printing with mupdf? | 19:49.39 |
Robin_Watts_ | Ch3rrryC0ke: hi. | 19:54.05 |
| The current plotters in mupdf only know how to go to 8bpc rgb or grayscale. | 19:54.27 |
| mudraw can go to 1bpp output, but it does that by rendering to 8bpp contone then dithering down internally. | 19:55.06 |
| Presumably you are using mudraw here ? | 19:55.31 |
Ch3rrryC0ke | I'm not using mudraw-- I'm using the libmupdf library. This is a Win8 app, so I'm basically rendering a bitmap, wrapping it in a Windows control, and then passing that control to the Win8 print framework. | 19:56.27 |
Robin_Watts_ | If you use -b 0 you will disable antialiasing, in theory at least. | 19:56.28 |
| Right. | 19:56.39 |
Ch3rrryC0ke | How can I do that in code? (E.g. not command line) | 19:57.01 |
Robin_Watts_ | fz_set_aa_level(ctx, 0); | 19:57.22 |
Ch3rrryC0ke | Ah awesome, let me try that real quick. Sorry if I'm using the wrong terminology here, I'm still getting up to speed on mupdf. | 19:57.48 |
Robin_Watts_ | no, you've been bang on the money so far. | 19:58.05 |
Ch3rrryC0ke | When you said mudraw, I thought that meant strictly the command line tool | 19:58.12 |
| Does the fz_set_aa_level apply globally? | 19:58.34 |
Robin_Watts_ | It applies to all renders done on that context. | 19:58.49 |
Ch3rrryC0ke | Ah gotcha. So right now how I have things structured is similar to the Win32 mupdf Viewer. I have a single "mupdf" class , (technically a "WinRT component"), and I instantiate one object of this per document. | 20:00.29 |
| It has methods to render a page, or to render a "print page" etc, and only has one context | 20:00.47 |
Robin_Watts_ | ok, that's fine, unless you ever want to start doing multithreaded magic | 20:01.28 |
Ch3rrryC0ke | Yeah-- I want to eventually do multithreaded magic. How can I learn about all that / are there some good examples? | 20:02.10 |
| Especially because the Win8 stuff is ALL about async functions. | 20:02.41 |
Robin_Watts_ | doc/overview.txt | 20:02.50 |
Ch3rrryC0ke | Does fz_aa_level need to be set before fz_run_page, or can it be set after fz_run_page and before fz_run_display_list | 20:03.47 |
Robin_Watts_ | I suspect it needs to be set before you create the draw device. | 20:04.45 |
Ch3rrryC0ke | Robin_Watts: thanks! I just tried printing with AA set to 0, and the text looks MUCH more crisp. | 20:10.30 |
Robin_Watts_ | fab. | 20:10.36 |
Ch3rrryC0ke | Any other optimizations for printing? | 20:10.45 |
Robin_Watts_ | Ch3rrryC0ke: Not that I can think of offhand. Will this be available at some point? | 20:11.07 |
| We'd love to see a metro app for mupdf. | 20:11.29 |
Ch3rrryC0ke | Robin_Watts_: Yeah, the app is in the Win 8 app store right now, it's called PDF Touch. It's actually a PDF annotator. | 20:11.38 |
| here's a link: | 20:11.49 |
| http://apps.microsoft.com/webpdp/app/pdf-touch/c9d9c388-68ca-499b-ae0f-d829f19ad613 | 20:12.40 |
| And I finally got it compiling for ARM, so if you have a "Surface", it will work on that too | 20:13.07 |
Robin_Watts_ | Ch3rrryC0ke: So, your app is released under the GNU GPL, right? | 20:13.57 |
| I can't see a declaration of that anywhere. | 20:14.07 |
Ch3rrryC0ke | Robin_Watts_: no, I'm paying for a mupdf license. | 20:14.13 |
Robin_Watts_ | oh, right, sorry! | 20:14.20 |
Ch3rrryC0ke | Robin_Watts_: No worries, I understand ;) | 20:14.44 |
| I would love any feedback if you're able to check it out. | 20:15.17 |
Robin_Watts_ | I don't have windows 8, but mvrhel does... | 20:15.29 |
mvrhel | hello? | 20:16.19 |
Robin_Watts_ | mvrhel: Ch3rrryC0ke has written a Metro app using mupdf. It's on the app store. | 20:16.43 |
mvrhel | I am playing around with writing a win8 app right this second | 20:16.46 |
| cool | 20:16.49 |
Robin_Watts_ | See the link above. | 20:16.49 |
mvrhel | I will check it out | 20:16.54 |
Ch3rrryC0ke | One more Q if you have a sec-- my app is a PDF annotator, and I'm using iText to write annotations back to the PDFs. With certain PDF files, after saving my annotations, they "move" once re-rendered by mupdf, as if the units of the PDF are somehow different than what I expect. This only happens with certain files, and I can't figure out why-- any idea how I can debug this / figure out what's going on? | 20:17.06 |
| I saw that pdf 1.6 introduces custom units, but one of the files that this repro's on is pdf version 1.4. | 20:17.31 |
Robin_Watts_ | I am unfamiliar with iText. | 20:18.00 |
Ch3rrryC0ke | mvrhel: thanks. Would love if you can e-mail me feedback at support@pdftouchapp.com. | 20:18.08 |
Robin_Watts_ | How much do they move ? | 20:18.17 |
mvrhel | ok grabbing your app right now | 20:18.19 |
Ch3rrryC0ke | They move about half an inc down and a half an inch left | 20:18.32 |
Robin_Watts_ | Consistently, no matter where they are placed on the page? | 20:18.50 |
| I'd be suspicious that the document was rotated or something, but if you're getting the same offsets whereever you are on the page, that seems unlikely. | 20:19.48 |
Ch3rrryC0ke | Yeah, it seems so (down and left), and on some pages, they dissapear entirely | 20:20.26 |
Robin_Watts_ | Could be to do with odd margins or something. | 20:20.56 |
| Every page has various bounding boxes, (MediaBox, CropBox etc). | 20:21.14 |
| Typically they are [0 0 width height] | 20:21.24 |
mvrhel | neat app | 20:21.32 |
Robin_Watts_ | but sometimes they might be [xoff yoff width+xoff height_yoff] | 20:21.45 |
Ch3rrryC0ke | I can show you an example / send the file. But basically I use mupdf's text extraction to highlight the text, then I convert those units to "pdf units" (1/72 inch, origin at bottom left", and then write the annotation back to the file. | 20:21.55 |
Robin_Watts_ | If you were calculating the offsets into the page by pixels, you'd consistently be off by (xoff, yoff). | 20:22.30 |
| Sure, send me a file. robin.watts at artifex.com | 20:22.47 |
Ch3rrryC0ke | Robin_watts_: ok, I'll look at that, as well as rotations. Are "mediaBox" and "CropBox" PDF concepts or mupdf concepts? | 20:22.51 |
Robin_Watts_ | PDF concepts | 20:22.59 |
Ch3rrryC0ke | Thanks, sending right now.. this was sent to me by a user, so no judgements on the content ;) | 20:23.23 |
mvrhel | so how do you make money on this though Ch3rrryC0ke? | 20:24.48 |
Ch3rrryC0ke | mvrhel: Right now, I don't- I'm working on building a pro version that I hope to charge for, and my bigger goal is to build a company around making productivity applications for tablets. | 20:25.53 |
| PDF Annotator is hopefully the first of a suite of apps. | 20:26.26 |
| Robin_Watts_: File sent | 20:28.13 |
Robin_Watts_ | mvrhel: I've just added a -l flag to the ets test thing, so we can set the number of levels. | 20:29.00 |
| results look nice with -l4 | 20:29.13 |
| and I've pushed my sanitised code. | 20:29.34 |
mvrhel | Robin_Watts_: great | 20:29.39 |
| off to lunch. I will update and take a look | 20:29.51 |
Robin_Watts_ | The white spots in the tongue do not appear to be overflows though, so I'm confused. | 20:30.04 |
| s/tongue/mouth/ | 20:30.31 |
mvrhel | oh that is surprising | 20:30.35 |
Robin_Watts_ | actually, maybe we shouldn't be surprised to see the odd dot. | 20:38.25 |
| Maybe the shade just happens to be very close to the exact ratio that would cause it to be completely covered in dots, so just a few are showing? And the ets biasing tips it over. | 20:39.16 |
| Arranging to only bias in the lighter part of the range would solve that maybe. | 20:39.50 |
| Ch3rrryC0ke: /CropBox[17.0079002 17.0079002 331.600006 466.992004]/MediaBox[0 0 331.600006 484] | 20:42.44 |
| so if the disparity between CropBox and MediaBox IS what's going on here, we'd expect to see a 5% horizontal displacement. | 20:43.55 |
| I reckon we're seeing a 1.3cm horizontal displacement for a 25cm rendering | 20:45.52 |
| which is about right. | 20:46.18 |
| We'd expect a 3.6% displacement up/down... | 20:47.00 |
| I can't tell where you were expecting it to be. | 20:49.12 |
| "sage is still positive, ...ways." would make sense I guess. | 20:49.46 |
Ch3rrryC0ke | Robin_Watts: thanks, that's already really helpful. I'll try adjusting for the crop box and media box. | 20:56.09 |
| I'll read about those in the standard first and then go from there. | 20:56.39 |
| So for example the "When he tells" was meant to be highlighted in the first paragraph on page 6 | 20:57.19 |
| it shifted down and left | 20:57.27 |
Robin_Watts_ | Really? | 21:02.08 |
| I reckon "proved over time .... counsel." was highlighted. | 21:02.36 |
| and "find such friends.... Bible." | 21:03.16 |
| The offset to "When he tells" seems wrong to me. | 21:04.29 |
Ch3rrryC0ke | Here's a before and after, I used a different color this time to make it more clear ;) | 21:06.20 |
| http://static.inky.ws/image/3390/image.jpg | 21:06.22 |
Robin_Watts_ | Right, so that looks like the margins are plausibly to blame. | 21:09.42 |
Ch3rrryC0ke | Ok cool, are margins different than cropbox/ media box? | 21:10.28 |
Robin_Watts_ | No, I meant cropbox/mediabox, sorry. | 21:11.48 |
Ch3rrryC0ke | No prob, thanks so much, this has been very helpful! | 21:12.29 |
| Are "custom units" in PDFs fairly common? | 21:18.31 |
Robin_Watts_ | No, not at all. | 21:19.47 |
| That's an Acrobat hack. | 21:20.05 |
ray_laptop | I definitely got a virus (probably the same as my daughter had) | 22:07.20 |
Ch3rrryC0ke | Is there a way to get the cropbox of a page from mupdf? | 22:54.35 |
| Forward 1 day (to 2012/11/09)>>> | |