| <<<Back 1 day (to 2014/05/25) | 2014/05/26 |
kens | Morning tor8 | 09:43.13 |
tor8 | hi kens | 10:09.01 |
kens | :-) slow reaction, no breakfast yet ? | 10:09.14 |
| tor8 could you take a real quick look at this one please: | 10:09.45 |
| http://stackoverflow.com/questions/23808529/how-to-apply-androidtextisselectable-option-for-mupdf-viewer | 10:09.45 |
| I suspect its a dumb question but I'm not certain | 10:09.46 |
tor8 | kens: yeah... had a bout of food poisoning over the weeked :( | 10:10.18 |
| so I'm a bit lethargic in general... | 10:10.27 |
Robin_Watts | tor8: ouch. in sympathy with Scott? | 10:10.32 |
kens | :-( | 10:10.34 |
tor8 | most likely! | 10:10.42 |
Robin_Watts | I think my little dog may be having (or have had) a stroke, so am waiting for the vet to call back. | 10:11.07 |
| Why does this stuff always happen on a bank holiday? | 10:11.21 |
tor8 | kens: the question is about android so I'm not really sure about the answer. | 10:11.34 |
kens | otr8 THat's fair comment, you probably know more than me though :-) | 10:11.53 |
tor8 | robin or paul would know better, I've made a point of keeping my fingers out of the android viewer :) | 10:12.13 |
Robin_Watts | kens: AIUI, the answer is "code a load of android UI stuff" | 10:12.28 |
kens | I had a suspicion that they wanted to use some kind of built-in Android text-selection thingy, an that's not going to work I woudl assume | 10:12.32 |
Robin_Watts | certainly it's not something that is going to be a couple of lines. | 10:12.54 |
| Let me try and scribble an answer. | 10:13.23 |
kens | Thanks Robin_Watts | 10:13.28 |
Robin_Watts | posted. | 10:40.33 |
| tor8: Good news from the slowdown bug. Apparently the patch made it faster. And we're now twice as fast as 1.2 was :) | 10:49.00 |
tor8 | Robin_Watts: oooh! | 10:49.08 |
| does that include the newest bug that compares it to 1.2? | 10:49.24 |
Robin_Watts | That was from the bloke who opened that bug, yes. | 10:49.42 |
tor8 | ah, yes, I see that comment now. good :) | 10:50.27 |
| Robin_Watts: have you looked at 695260? it looks like the try/always/catch blocks in run_xobject aren't fine grained enough | 10:57.46 |
| there's a number of different states that might be unbalanced depending on where an exception is thrown | 10:58.10 |
Robin_Watts | looking now. | 10:58.49 |
| the way I'd do it, I think, would be to have a volatile int state = 0; | 11:01.25 |
| then state = 1 after the pdf_gsave; | 11:01.38 |
| then in the always: if (state >= 1) pdf_grestore(); | 11:01.54 |
| etc. | 11:01.55 |
| that way we only have the one try/catch overhead, but we can correctly clean up. | 11:02.10 |
| The overhead on a try/catch is a lot more than simply updating a state var. | 11:02.51 |
tor8 | yeah. similar to the chained goto cleanup1, cleanup2, etc labels we used before exceptions. | 11:03.02 |
| indeed. and not just the overhead of excessive source indentation :) | 11:03.22 |
Robin_Watts | yeah. | 11:03.28 |
| I can have a look at that, or you can. | 11:03.42 |
tor8 | Robin_Watts: I'll let you take this one :) (are you headed back to the SOT salt mine otherwise?) | 11:04.15 |
Robin_Watts | yeah. | 11:04.26 |
| I got the font stuff into SOT at the end of last week, so I feel I've earned a distraction :) | 11:04.52 |
| Update from vet - it probably was a stroke, he seems fine now, there is nothing they can do, so just keep him at home and watch him. | 11:05.28 |
tor8 | speaking of font stuff, I should poke henrys about the URW progress | 11:05.41 |
Robin_Watts | I had a email from an SOT client this morning, so will have to look into that. I think the basic answer is "yes, you've put the font into the system, but unless you update the font map to tell it that that font provides chinese, it won't get used" | 11:06.32 |
kens | Hmm, is aspirin poisonous ot dogs ? | 11:06.57 |
Robin_Watts | kens: Possibly. I know that ibruprofen kills them. | 11:07.21 |
| and paracetamol is bad too IIRC. | 11:07.38 |
| basically, I'm not risking it :) | 11:07.50 |
kens | Yeah that's true (also chocolate), but not sure about spirin | 11:07.50 |
| I guess vet would suggest it idf its sensible, so presumably its a bad idea | 11:08.25 |
tor8 | paracetamol is very dangerous to overdose, and the normal dose is not very far from a lethal dose | 11:08.41 |
| so a human dose of paracetamol to a dog would be quite the bad idea, I reckon | 11:09.14 |
Robin_Watts | yeah. paracetamol would never have been given the "over the counter" status that it enjoys had it been brought to market in recent years. | 11:09.16 |
| apparently aspirin itself is OK (under vetinary supervision), but some of the stuff they often have in with the aspirin really isn't. | 11:09.50 |
| (according to Professor Google) | 11:09.59 |
kens | How old is the dog ? | 11:10.26 |
tor8 | aspirin has a similar effect to ibuprofen of inhibiting inflammation, but it doesn't work as well as a pain relief | 11:10.27 |
kens | No, but it is good for stroke patients, thins the blood | 11:10.43 |
| at least, it does in humans | 11:10.50 |
| maybe not dogs | 11:10.54 |
Robin_Watts | kens: rescue dog, so no idea. We've had him 11 years, and he was at least 3 when we had him. | 11:10.54 |
kens | Well that's a fair age | 11:11.21 |
tor8 | kens: yeah. you use aspirin for the medical effects, not the pain relief, is what I was trying to get across :) | 11:11.27 |
kens | True, same for ibuprofen in my experience | 11:11.41 |
Robin_Watts | yeah, he's like the duracell bunny, pretty much. | 11:11.49 |
kens | Probably too risky to prescribe much witha dog that age | 11:12.42 |
tkamppeter | ow talking on #control-center | 14:06.05 |
| <tkamppeter> I have a problem with color spaces. I can convert the file launch_leaflet.pdf correctly to CUPS/PWG Raster (gdevcups.c) with CMYK color space up to 8 bits per color, with 16 bits I get wrong colors only for text/vector parts, images are correct. Can someone help me debugging this? | 14:06.06 |
kens | Probably Ray | 14:06.40 |
tkamppeter | Poppler does this correctly. | 14:06.49 |
| kens, thanks, will try again at night. | 14:09.22 |
kens | probably in 2 or 3 hors till | 14:09.42 |
| hours* | 14:09.52 |
tkamppeter | ok | 14:16.15 |
Robin_Watts | today is is US holiday, I believe. | 14:42.53 |
| s/is is/is a/ | 14:43.02 |
kens | didn't realise that | 14:43.08 |
Robin_Watts | Memorial day, maybe? | 14:43.19 |
| tkamppeter: Are the image colors byte reversed? | 14:44.00 |
| i.e. are you getting 0x3412 when you expect 0x1234 ? | 14:44.13 |
| And on what platform is this? | 14:44.24 |
| sorry, my question should have been "Are the incorrect colors byte reversed?" | 14:46.18 |
kens | Appears you are right Robin, May 26th Memorial Day | 14:56.46 |
tkamppeter | Robin_Watts, I will look into it ... | 15:00.39 |
tor8 | so since it's a holiday in the US I expect marcos hasn't turned on the cluster machines he turned off over the weekend...? | 15:05.31 |
| Robin_Watts: a handful of commits on tor/master for review | 15:26.04 |
| the last one (the font extraction one) has lots of cluster differences, but they're either progressions or rounding differences | 15:27.01 |
Robin_Watts | 5 mins, and I'll look. | 15:28.18 |
| tor8: The tm/ctm one looks fine. | 15:38.38 |
| I've looked at the other two before, apparently. | 15:38.57 |
| and they are fine. | 15:39.25 |
tkamppeter | Robin_Watts, did you see my answers? | 15:53.22 |
Robin_Watts | I did not. | 15:53.36 |
| nor did the logs, it seems. | 15:54.20 |
kens | I didn't see it either fwiw | 15:54.33 |
tkamppeter | Robin_Watts, if I switch byte order of the CUPS Raster output images stay correct (they seem to be CMYK 8-bit and conversion to 16-bit by GS seems to be independent of Endianess=. | 15:56.52 |
| The text and vector stuff is wrong in both cases, but less wrong )near correct) with byte order changed. I am on Intel PC (probably little endian). | 15:56.59 |
| Robin_Watts, I do not know whether I get 0x3412 when I expect 0x1234, I see only the visible results of the rendered files on my screen. | 15:57.11 |
| Robin_Watts, here they are again ... | 15:57.38 |
Robin_Watts | tkamppeter: OK, if you're on an LE platform, then I would be very surprised if it was an endianness issue. | 15:57.38 |
| Intel PC is indeed LE | 15:58.02 |
| tkamppeter: So, I'd suspect that something is wrong with the encode_color and decode_color routines. | 15:58.44 |
| Do the encode_color and decode_color routines correctly pack 16 bits per color? | 15:59.29 |
tkamppeter | Robin_Watts, according to a quick view it seems so, but one has to look deeper into it. | 16:00.16 |
| Robin_Watts, I will look into it and tell you later or tomorrow. I have to leave now. | 16:03.35 |
Robin_Watts | ok. | 16:03.41 |
tkamppeter | Robin_Watts, thanks so far. | 16:03.42 |
Robin_Watts | no worries, I haven't done much :) | 16:03.49 |
kens | Goodnight all | 16:22.47 |
el_mendi | Hi! | 16:57.17 |
| does anyone know how to set an image as an annotation? | 16:57.58 |
| Or if there is a better way to save some image in an already existing PDF file? | 16:58.49 |
Robin_Watts | tor8: Bug fix on robin/master then. | 17:05.14 |
tor8 | Robin_Watts: did you push? | 17:22.12 |
| I can't see anything on robin/master | 17:22.18 |
Robin_Watts | Sorry, try now. | 17:31.25 |
el_mendi | Anyone has a clue of how to put a multicolor ink annotation? | 17:41.37 |
Robin_Watts | el_medi: A single ink annotation with more than one color in it? Can't be done, AIUI. | 17:52.03 |
| macossers may be interested in this: https://github.com/fredericgermain/valgrind/commits/homebrew | 17:57.38 |
| A version of valgrind that works on OSX Mavericks. | 17:57.50 |
el_mendi | And is there an option to introduce an image as an annotation? | 17:59.14 |
| May be an annotation with 2 parameters, the area of the image and an Array of colors of each pixel. | 18:00.49 |
Robin_Watts | el_mendi: Check table 8.20 in the PDF 1.7 spec and you'll see the list of annotations possible. | 18:01.54 |
| Image is not one. | 18:02.00 |
el_mendi | Ok, so there is no way to introduce an image inside an existing PDF? | 18:03.26 |
| Overwriting it, like an annotation | 18:03.39 |
Robin_Watts | el_mendi: Sure there is, just not using an annotation. | 18:03.46 |
el_mendi | May be like a Watermark? | 18:04.00 |
Robin_Watts | yeah, we've had requests for watermarks in the past. | 18:04.13 |
| I've done some work towards this in mupdf. | 18:04.20 |
el_mendi | I need to implement this function in the Android app of mupdf | 18:05.05 |
Robin_Watts | I have to go now, but will be back tomorrow (earlier than this) and I can talk about it then. | 18:05.31 |
el_mendi | Ok | 18:05.48 |
| thank you | 18:05.51 |
Robin_Watts | You'd basically need to run the page contents through a pdf-op-filter into a pdf-op-buffer | 18:05.51 |
| That would allow to recreate a buffer with the page contents in. | 18:06.06 |
| Then you can append to that using fz_device calls to draw your extra content. | 18:06.30 |
| gotta go! | 18:06.33 |
el_mendi | ok! thank you! | 18:06.43 |
tor8 | Robin_Watts: shouldn't the state setting be after the begin_softmask/etc calls? | 18:30.36 |
Robin_Watts | tor8: No. | 18:31.57 |
| Some functions (like begin_softmask) handle the error internally, and increment error_depth. | 18:32.26 |
| and in those cases, you must always call the closing function. | 18:32.49 |
| hence what the state records is "that I tried to call begin_softmask" | 18:33.04 |
| and therefore that you need to call "end_softmask" | 18:33.12 |
| gotta run, sorry! | 18:33.30 |
tor8 | Robin_Watts: okay. LGTM then. | 19:04.48 |
henrys | marcosw: I just upgraded Mac Pro to maverick | 23:02.55 |
Robin_Watts | henrys: https://github.com/fredericgermain/valgrind/commits/homebrew | 23:10.01 |
| check that out to get a version of valgrind that should work. | 23:10.14 |
henrys | yeah I use the valgrind from svn and build it myself on mac, but I find I'm valgrinding on linux these days. | 23:13.31 |
| Robin_Watts: ^^^ | 23:13.35 |
Robin_Watts | right, but mainline valgrind doesn't work on mavericks. | 23:13.48 |
henrys | oh this is a fork? | 23:14.07 |
| thanks for that. | 23:14.21 |
Robin_Watts | This is valgrind + 6 commits to make it work, yes. | 23:14.30 |
marcosw | henrys: it shouldn't affect the cluster, my macpro has been running mavericks since it came out. | 23:14.36 |
henrys | marcosw: I'm rather annoyed I have to pay for the server again. | 23:15.15 |
Robin_Watts | I have been using my mac to watch amazon instant prime video, and I find it judders a bit. | 23:15.33 |
henrys | marcosw: I guess I could screw around with it and get it all working myself but the server is convenient | 23:15.47 |
Robin_Watts | I wonder if that's because of their funky timer stuff to save power. | 23:15.50 |
marcosw | henrys: isn't it cheap? | 23:16.02 |
henrys | marcosw: yeah 20 bucks - just whinin' | 23:16.38 |
| Robin_Watts: is amazon prime big over there? | 23:18.34 |
henrys | just finished Boardwalk Empire 4. Any scene from the series could be a painting. Visually awesome. | 23:20.47 |
Robin_Watts | BWE looks brilliant. | 23:26.37 |
| I feel the writing lurches a bit in the middle of series, but the period detail is great. | 23:27.38 |
| Amazon prime (the delivery) is big here, yes. | 23:28.01 |
| The fact that you get video included with it now, has raised hackles though, cos they bumped the price by 35% to cover it. | 23:28.37 |
| and frankly I'd rather have my 35% back and pay netflix :) | 23:28.48 |
| Forward 1 day (to 2014/05/27)>>> | |