| <<<Back 1 day (to 2013/09/09) | 2013/09/10 |
paulgardiner | tor7: ping | 11:38.48 |
Robin_Watts | kens, tor: Helen has been looking into flights to/from Hawaii. | 12:32.24 |
kens | any news ? | 12:32.30 |
Robin_Watts | Virgin want about 1000 each (ecomony both ways, to honolulu, then got to fly to Maui and back). | 12:32.52 |
| So she called "flight centre" a company we've used before. | 12:33.04 |
kens | I'd pretty much discoutned Virgin, they fly to the wrong airport | 12:33.14 |
Robin_Watts | Supposedly they have a way to get out/back for 7xx each, with the return leg being PE. | 12:33.33 |
| outbound to honolulu, inbound from Maui. | 12:33.52 |
| cos we want to spend a couple of days on hololulu first. | 12:34.07 |
kens | I found I could fly Delta, PE all the wya, with the outbound leg to LAX being a Virgin codeshare | 12:34.07 |
Robin_Watts | kens: fair enough. | 12:34.57 |
kens | I don't remember the costs | 12:35.20 |
russcriptor | good morning | 12:35.23 |
Robin_Watts | russ: morning. | 12:35.32 |
russcriptor | is there a way to build a shared library of mupdf? | 12:35.40 |
Robin_Watts | make XCFLAGS="dfsdfsdfsdf" | 12:36.06 |
| where dfsdfsdfsdfs is whatever flags you want. | 12:36.30 |
kens | Robin_Watts : Looks like the delta cost (economy) is ~£1400 | 12:37.26 |
Robin_Watts | kens: per person? ouch. | 12:38.03 |
kens | per person, yes | 12:38.14 |
| The cheap options all run to many more hours travelling, at least ther ones I found | 12:38.39 |
| It *is* a long way.... | 12:38.52 |
Robin_Watts | kens: indeed. | 12:38.58 |
| Air Canada all seemed to be 2 changes as far as I could see. (Toronto and Vancouver) | 12:39.19 |
kens | On the return yes, on the way out there's a 1-stop in calgary | 12:39.37 |
Robin_Watts | wasn't showing me that. | 12:39.48 |
kens | but air canada don't do premium economy | 12:39.48 |
Robin_Watts | http://www.hotelclub.com/blog/the-worst-airlines-in-the-world/ | 12:40.33 |
| "I'd rather walk than fly air canada" | 12:40.42 |
kens | The Delta site doesn't want to tell me how much PE will cost | 12:40.43 |
russcriptor | Robin Watts: so I did export XCFLAGS=-fPIC and then make && make install | 12:41.15 |
| Robin Watts: does that sound right? I'm trying to update this project: https://github.com/rk700/python-fitz to mupdf 1.3 (basically the python bindings) | 12:41.56 |
Robin_Watts | russcriptor: sounds plausible. | 12:42.14 |
kens | Best price on Air Canada is £578 per person | 12:43.06 |
russcriptor | Robin_Watts: I have the following in /usr/local/lib/libmupdf: libmupdf.a and libmupdf-js-none.a, shouldn't the be .so files? | 12:43.24 |
Robin_Watts | russcriptor: The XCFLAGS won't have changed the link used, come to think of it. | 12:43.51 |
| You probably need to tweak the Makefile. | 12:43.57 |
kens | I still see AC851 to Calgary, then AC043 to Kaluhui, cost is £356 for that leg, flight time is 18 ours 5 minutes | 12:44.01 |
russcriptor | Robin_Watts: I see there was a patch, but doesn't look like it was ever implmeneted: http://bugs.ghostscript.com/show_bug.cgi?id=693009 | 12:44.28 |
kens | NB the aurcraft is a 'new interior' and has seat back video, laptop power and possibly other stuff. | 12:45.03 |
| THere are 4 return options at 22.5 hours, all have 2 stops (Calgary+TOronto or Vancouver+Toronto) | 12:47.05 |
| THat leg costs £225 | 12:47.17 |
| Total £581 | 12:47.30 |
| But.... no optionto upgrade to Premium Economy, just busines class at 5 times the price | 12:47.51 |
Robin_Watts | I shall leave this to Helen to sort. | 12:48.38 |
kens | She might be interested in what I found, but obviously its up to you guys | 12:48.56 |
zeevb | Hello, I have a question regarding Bug 693578 - http://bugs.ghostscript.com/show_bug.cgi?id=693578 | 12:49.01 |
Robin_Watts | I will forward it on. thanks. | 12:49.07 |
kens | I zeevb we have nothign to report on that bug | 12:49.28 |
zeevb | Is there a way to detect if a file has different embedded subsets of the same font with the SAME prefix? | 12:50.04 |
kens | zeevb, not unless you want to write an application, it requires parsing the PDF file | 12:50.29 |
zeevb | kens, I see. I tried pdffonts but it only lists the font names but nothing about the subset itself | 12:52.20 |
| kens, I used PyPDF2 to try and parse my input file but I didn't manage to detect subsets with the same prefix which contain different charsets | 12:58.57 |
kens | Robin_Watts : I was wrong on the Delat flights, its £752 with them, the web site was confusing | 12:58.59 |
| zeevb if there are two fonts with the same subset prefix the likelihoof is high that they contain different sets of glyphs | 12:59.30 |
zeevb | kens, I see. Is that information contained in the /CharSet ? | 13:02.45 |
kens | zeevb If I remember correctly the CharSet is only valid for CIDFonts | 13:03.06 |
zeevb | kens, hmm... I see | 13:03.26 |
kens | But I could me mistaken. In any event the only place you cna be sure of checking the glpyhs available is the font program | 13:03.29 |
zeevb | kens, "the font program" - what is it? | 13:04.39 |
kens | THe FontFile object | 13:04.50 |
zeevb | kens, oh, ok | 13:05.06 |
kens | It contains a type 1/TrueType/CFF font program | 13:05.11 |
zeevb | kems, I wish I knew how to parse it | 13:06.06 |
kens | The specifications are availabe, but its not simple | 13:06.20 |
russcriptor | Robin_Watts: do you know what I need to tweak in the Makefile to build the shared .so libraries? | 13:06.23 |
zeevb | sigh | 13:06.50 |
kens | zeevb the simple asssumption is that if you have 2 fonts weith the same subset prefix and name, they won't live happily together | 13:07.27 |
| zeevb why do you need to know ? | 13:07.51 |
Robin_Watts | russcriptor: Not offhand, but you should be able to figure it out by looking at the patch. | 13:07.55 |
russcriptor | Robin_Watts: I'm trying to, but C isn't my strong suit. Is there a reason that this feature wasn't rolled in? | 13:08.55 |
Robin_Watts | yes. | 13:09.05 |
| Share library builds only make sense if you consider the API frozen. | 13:09.20 |
| and we haven't frozen all the API. | 13:09.38 |
zeevb | kens, my application receives a user generated PDF and runs gs on it (for prepress etc.) and I got an input file that yielded the same error reported in Bug 69357 | 13:09.56 |
kens | zeevb well that's not the same thing as you are asking about. | 13:10.25 |
| The error is from Adobe Acrobat and no other application complains about the fonts. We have no idea what Acrobat is complaining about but there is no reason to believe its multiple fotns with the same subset. | 13:11.15 |
| When pdfwrite creates the output PDF it is more or less guaranteed that there will *NOT* be any such occurences. | 13:11.51 |
russcriptor | Robin_Watts: so is libfitz now libmupdf? | 13:12.29 |
zeevb | kens, I see. The problem I have is that the output file is sent to our print partner and they complained that they can not print it since it was missing some fonts | 13:12.35 |
| then we checked with Adobe reader and saw this error | 13:13.00 |
kens | zeevb that still doesn't sound like the same problem. | 13:13.04 |
| Its not impossible to create a PDF file with fonts not embedded using pdfwrite, but its hard. | 13:13.27 |
| I would have to see the original input, and the command line used with Ghostscritp to say any more. | 13:13.57 |
zeevb | kens, how can I deliver that to you? | 13:14.33 |
kens | Well.... | 13:14.40 |
| You could open a bug report and attach the file, or if its a sensible size you could email it to me | 13:14.56 |
Robin_Watts | tor7: ping | 13:16.14 |
zeevb | kens, ok - I'll check if there are any restrictions sending this file (it is of one of our clients). Thanks | 13:18.10 |
kens | OK | 13:18.18 |
tor7 | robin_watts: paulgardiner: ping | 13:23.19 |
| or pong even | 13:23.22 |
kens | tor7 is still tired ;-) | 13:23.34 |
tor7 | kens: looking at the delta flights, I don't see any appealing return flights... all of the ones that come up OGG-LHR are 2 stops on the return | 13:23.56 |
| or 7-8 hrs in LAX | 13:24.05 |
kens | tor7 there are very few returns on any airlines that aren't multiple stops | 13:24.17 |
tor7 | CPH-OGG routes through frankfurt with delta code shares | 13:24.28 |
kens | The Air Canada one is multiple stops too | 13:24.29 |
| tor7 it may be easier for you to go straight back to Copenhagen, rHTER THAN COMING BACK THROUGH lONDON ? | 13:24.57 |
tor7 | multiple stops does mean you can get out and stretch your legs properly :) | 13:24.59 |
kens | tor7 yes, I did consider that a bonus actually | 13:25.16 |
| the one I'm looking at is 22.5 hours return, stopping at LAX and JFK | 13:25.31 |
paulgardiner | tor7: what do you suggest as the best way to move PDF files onto an iPad or iPhone? | 13:25.46 |
kens | So 3 stints in the air, 6,7 and 5 hours | 13:25.53 |
tor7 | kens: yeah, could be, but then I'd have to come in via LHR first | 13:26.30 |
| paulgardiner: download a pdf in safari, then click the "Open in MuPDF..." button that should appear | 13:26.51 |
kens | I could meet you there instead of yo ucoming here | 13:26.51 |
paulgardiner | Ah yes, of course. Ta | 13:27.22 |
tor7 | kens: yes, I see two options -- meet up on the final or semi-final leg so we have company when being zombiefied at arrival | 13:28.19 |
| or I come visit for a day before and after :) | 13:28.27 |
kens | tor7 you are en welcome to come and visit us, Stella keeps telling me to make sure and offer. You are of course welcome to stop over on both directions | 13:29.07 |
| s/en/entirely/ | 13:29.21 |
| THe only downside is your trip would be very long. | 13:29.38 |
tor7 | kens: considering how pointless and painful the trip is going to be, visiting would be a bright spot in it all :) | 13:29.43 |
kens | You only say that because you haven't met Melanie yet :-D | 13:30.03 |
tor7 | kens: haha :D | 13:30.09 |
| kens: my giant table was delivered today! my god it's HUGE. | 13:30.32 |
| just perfect :) | 13:30.42 |
kens | But can you put all the expansion aon it at the same time ? | 13:30.51 |
tor7 | 240x120cm ... I believe it should be just about possible | 13:31.07 |
kens | Wow, that is indeed big | 13:31.15 |
| Just as well you have a long reach | 13:31.22 |
tor7 | 4'x8' | 13:31.30 |
| now I just need more chairs to go around the table... | 13:31.49 |
kens | Hah, who needs chairs anyway :-) | 13:32.04 |
tor7 | kens: huhm... the prices that aircanada are quoting now are a *lot* higher than when I was looking in the US | 13:33.53 |
kens | Hmm, they are the same for me, I wonder if its because you are searching from a Swedish IP address.... | 13:34.15 |
tor7 | oh wait, wrong airport. | 13:34.31 |
kens | Oh boy, bad start :-) | 13:34.42 |
tor7 | AC wants GBP 2000 to take me from CPH to LHR via Toronto :) | 13:34.46 |
kens | Wow | 13:34.59 |
| For me, going via Calgary out and Vancouver/Toronto home its £580 | 13:35.23 |
tor7 | 800 executive class, 400 flex | 13:35.28 |
kens | really ? I'm seeing executive at about 2000 | 13:35.40 |
tor7 | 870 for executive, calgary out | 13:36.22 |
kens | strange | 13:36.36 |
tor7 | and the same coming back | 13:36.43 |
| probably because of the extra leg :) | 13:36.52 |
kens | Seems ot be a pretty reasonable price to em | 13:37.11 |
tor7 | and I can take a vancouver/toronto/london route on the way back | 13:37.29 |
kens | THa's the routes I see, but they are way more expensive for me using executive class | 13:37.50 |
| I should get you to book me the flight | 13:38.13 |
tor7 | if they didn't cancel your whole trip if you miss out the first leg, you should book coming from copenhagen as well :) | 13:38.39 |
kens | THat's what I meant yes :-) | 13:38.49 |
| Or maybe I shoudl come visit you and start from Lunde :-) | 13:39.02 |
tor7 | that would be even more painful, and we're short on spare beds, but you'd be more than welcome if you want! | 13:39.37 |
kens | I'm seeing 18:05 out, 22:30 back outbound via Calgary, return via Vancouver+toronto | 13:39.48 |
kens | was kidding | 13:39.59 |
tor7 | AC85/AC854/AC9967 coming back via vancouver/heathrow | 13:40.33 |
| outbound, I don't see any options for going via heathrow though | 13:40.55 |
kens | Hmm different flights altogether | 13:41.04 |
| I see 851/043 out, 058/034/856 return | 13:41.25 |
kens | will use the AC web site instead of the price comparison site, just a moment | 13:41.55 |
tor7 | http://www.flightstats.com/go/FlightStatus/flightStatusByFlight.do?&airlineCode=AC&flightNumber=854 | 13:42.09 |
kens | Hmm, which day are you looking for return dates TOr ? | 13:43.46 |
tor7 | outbound: the 8th | 13:43.51 |
kens | Very strange | 13:44.04 |
| I see 4 options on return, 2 each via Vancouver/Toronto and Calgary/Toronto | 13:44.28 |
tor7 | outbound the 6th I have two appealing options, via calgary or vancouver. both at 23-24hrs | 13:44.41 |
kens | Slight longer than me, but I was looking at flyong out on the 6th Via Calgary | 13:45.03 |
tor7 | yeah, there's an extra 3hrs minimum for me due to the extra leg to frankfurt | 13:45.24 |
kens | FGor me that's ac851 to Calgary and then AC043 to Kaluhui | 13:45.34 |
tor7 | return on the 8th I have multiple 30hr options | 13:45.40 |
kens | FOr me AC043 departs Calgary for Kaluhui at 16:10 (aircraft 763) | 13:46.08 |
tor7 | AC845 to calgary, then AC43 to Kahului | 13:46.14 |
kens | OK so we'd be on the same flight from Calgary, that would work | 13:46.31 |
tor7 | yeah 16:10, aircraft 763 | 13:46.32 |
| coming in for me that leaves a 3h30m layover | 13:47.20 |
kens | On the 8th there are 2 flight s from Kaluhui that I see, AC044 and AC058 | 13:47.21 |
tor7 | in calgary, which should be plenty | 13:47.25 |
kens | I'm not sure how much I would get, need to continue further,just a moment | 13:47.42 |
tor7 | on the 8th, AC44 wants me routed via frankfurt, ac58 routes through heathrow | 13:48.25 |
kens | Looks like I'd get a little over 2 hours | 13:48.26 |
| tor7 that is odd, given they both go to Vancouver or Calgary :-) | 13:48.49 |
tor7 | yeah... :) | 13:49.03 |
kens | But then, its not shopwing me your options when i COME BACK TO lONDON | 13:49.07 |
tor7 | maybe they just pick the "best" route and skip the leftovers | 13:49.15 |
kens | stupid caps lock ley | 13:49.16 |
| tor7 I guess that could be,m what time do the flights leave Maui that you see ? Mine all depart at 21:55 I think, or possibly 21:35 | 13:49.47 |
| oh yes, 21:35 | 13:49.52 |
tor7 | stupidly long layouver though, arrival in vancouver at 05:05, departure at 18:15 | 13:50.12 |
kens | oh :-( | 13:50.26 |
tor7 | I see 21:35 for both AC44 and AC58 | 13:50.29 |
| which explains the 32hr trip as opposed to 23 coming in | 13:50.44 |
kens | Yes, I can get either and they leave at the same time, one to Vancouver and one to Calgary | 13:50.56 |
tor7 | going home via frankfurt, the layover is still stupidly long, but that's 13:50 rather than 18:15 | 13:51.22 |
kens | Looks like IO'd have 1:40 on the way in to change, which is kind of tight, but I woudl hope they know what they are doing | 13:51.24 |
tor7 | hopefully calgary has less of an issue with US immigration :) | 13:51.58 |
kens | Well I would hope it would not be too bad | 13:52.10 |
tor7 | in european airports, slightly more than 1hr is the smallest they allow for connecting flights | 13:52.47 |
kens | Yeah, an hour is definitely tight, but I'd have nearly 2, assuming no delays | 13:53.05 |
| I'm inclined to go for it. But the executive price is huge | 13:53.33 |
tor7 | 3x the regular for you? | 13:53.41 |
kens | More than that | 13:53.48 |
| > 5x | 13:53.55 |
tor7 | Tango is dirty cheap, but it's only twice the Flex rate for me | 13:54.01 |
kens | FOr me Tango return is £222 and executive lowest is 1283 | 13:54.28 |
tor7 | and I'm sure they haven't added all the "airport taxes" and other crap charges on the AC website | 13:54.30 |
kens | I think they have, they match the price comparison sites, and are only a bit lower than the US airline options | 13:55.10 |
| AC do allow me to select a bulkhead/exit seat though for an extra small charge | 13:55.43 |
| So as long as I book real soon I would do that | 13:55.55 |
tor7 | kens: I'd be happy to do Delta or BA or anything. the lack of inflight entertainment doesn't bother me, I always bring my own. | 13:56.05 |
kens | tor7 the last stretch would not be a problem, its a short hop anyway | 13:56.30 |
| Its more about convenience for you | 13:56.50 |
tor7 | kens: ah, that's nice. it's the insecurity of not being able to guarantee an exit/bulkhead seat that makes me refuse economy | 13:56.51 |
kens | Form the comparison sites everyone is coming in at about £580 for the cheapest flights | 13:58.07 |
| WHich is what the AC site offers me | 13:58.15 |
| THe main reason for looking at Delta was because they would do Premium Economy, but their web site won't tell me what it costs until I book it. | 13:59.02 |
| All it says is 'from $9' | 13:59.17 |
tor7 | grand total comes to $3200 for executive | 13:59.44 |
kens | Ah that's more like it, that's more or less what it woudl cost me | 13:59.57 |
tor7 | kens: and then they'll say "sorry, not available for this booking" | 14:00.07 |
kens | tor7 yeah :-( | 14:00.17 |
| TO be honest, I think we want to book early for this trip, no matter who its with | 14:00.42 |
tor7 | kens: well, we're only staying two nights and not bringing anybody else, so miles may be willing to splurge for it if he can get over his "but *I'm* only flying coah" deal | 14:00.53 |
kens | Hmm, somehow I doubt it..... | 14:01.11 |
Robin_Watts | russcriptor: Yes, libmupdf now I think. | 14:01.14 |
russcriptor | Robin_Watts: cool, thanks, I think I was able to build the shared lib now | 14:01.52 |
| Robin_Watts: Now to figure out how to call it from python | 14:02.07 |
tor7 | kens: it's roughly half of what flying me to SFO used to cost in the early days | 14:02.18 |
kens | Wow! | 14:02.29 |
tor7 | yeah. the first few years it was dirt cheap, then the prices jumped a lot for a few years up to the $1500 range | 14:03.28 |
Robin_Watts | tor7, paulgardiner: My fix for reflow needs review when you get a chance - no hurry. | 14:03.47 |
| I've just pushed all but one of zenikos fixes too. | 14:03.58 |
kens | Hmm, Delta is offering me 1 flight with a 35 minute layover, thats... ambitious.... | 14:04.01 |
tor7 | eheh, yeah, that's ambitious to say the least! | 14:04.14 |
| unless it's a layover that lets you stay onboard | 14:04.26 |
Robin_Watts | The one I didn't push was... http://git.ghostscript.com/?p=user/zeniko/mupdf.git;a=commitdiff;h=db07085ca369b05979da26cfa5fd4476c1b925ee | 14:04.31 |
kens | Hmm that;s a possibility | 14:04.37 |
Robin_Watts | I can't see how that doesn't just run us into problems later. | 14:04.44 |
tor7 | robin_watts: an error like that should drop us into repair mode | 14:05.11 |
| but there are plenty of apple quartz generated pdfs with broken xref tables like that | 14:05.27 |
| might be he wants to open them without running repair mode | 14:05.35 |
kens | Ooh, there's an option to fly Delta between LAX and OGG while flying Virgin between LHR and LAX, in both directions | 14:06.04 |
| And I *know* Virgin do proper Premium Economy | 14:06.26 |
tor7 | well, the 12hr flight is the only important one | 14:06.40 |
kens | Actually the return is OGG->LAX->JFK->LHR and its only the JFK section that's Virgin | 14:07.04 |
tor7 | if you can get us on that, I'll just do the gatwick trip and not have to fight miles about flying executive | 14:07.13 |
| well, that's a leg stretcher at least so I wouldn't mind too much | 14:07.37 |
kens | At the moment that's looking like my best option. 20:19 out, 21:15 back | 14:08.18 |
| 3 hour layover in LAX | 14:08.29 |
| But after customs and immigration that won't be long | 14:08.47 |
| That's on the way out of course | 14:08.57 |
tor7 | yeah. better than the 11 hour layover in vancouver on the way back... | 14:09.05 |
kens | the return has 2 short layovers, 2:18 and 1:09 | 14:09.28 |
| which is probably OK, the 1:09 at JFK is a little worrying | 14:09.49 |
tor7 | I prefer that to waiting 7+ hours in an airport, to be honest | 14:10.04 |
kens | I'm inclined to agree | 14:10.11 |
| I'll be carry on bags only anyway, I assume you'll be the same | 14:10.23 |
tor7 | JFK-LHR probably has more flights per day so missing that one won't be too bad | 14:10.31 |
| yeah, I never check bags | 14:10.47 |
kens | ROFL, going to have to abandon that one, its > £3000 apiece | 14:11.57 |
| I'm not sure why its so expensive, I'll try a different search | 14:12.27 |
Robin_Watts | kens: Call Flight Centre, tell them what you want, and let them quote. | 14:13.33 |
| They can search flights WAY better than you can. | 14:13.42 |
kens | Probably true | 14:13.56 |
Robin_Watts | http://www.flightcentre.co.uk/ | 14:14.04 |
kens | THe problem is that I'm being picky | 14:14.15 |
Robin_Watts | LHR-> LAX: Economy 350, PE 1792, Upperclass 1136. ??!?! | 14:19.04 |
kens | Bonkers | 14:19.12 |
Robin_Watts | I know you get assholes in upper class, but really... | 14:19.24 |
kens | Americna Airlines Premum Economy is silly prices | 14:26.50 |
russcriptor | Robin_Watts: hmm not sure if I built the shared library properly. I'm getting /usr/local/lib/libmupdf.so.1.0: undefined symbol: jpeg_resync_to_restart | 14:29.17 |
kens | tor7 : I found a way to get the Delta flights at a sensible price, £998 including all taxes and fees | 14:35.47 |
| As far as I can see its a choice of that or the Air Canada flights | 14:38.49 |
| GOing to fetch some more coffee, back shortly | 14:39.03 |
Robin_Watts | russcriptor: Sounds like it's missing the jpeg stuff? but honestly, shared objects and stuff - all unsupported as far as I'm concerned. tor7 may be prepared to help you more. | 14:53.39 |
| kens, tor7, chrisl, paulgardiner: In case you're remotely interested... | 14:59.41 |
| it looks like the best option for us is to go: LHR -> SFO (with virgin), 4 hour layover, SFO -> HON (delta) | 15:00.14 |
| We then stay in honolulu for 3 nights, and fly to Maui. | 15:00.27 |
| Then on the evening after the meeting, we'd fly back to hon, stay overnight, fly home via LAX. | 15:01.03 |
kens | Mhm, Virgin do a codeshare flight to Honolulu with Hawaiin airlines | 15:01.06 |
Robin_Watts | kens: That's the one we'd be taking back. | 15:01.23 |
| but to do that on the way out, you have 2 hours at LAX to clear immigration, pick up and recheck baggage. | 15:01.48 |
kens | But I don't really want to fly to Honolulu..... I doubt my ability to complete the journey to Maui after a flight of that duration | 15:01.48 |
Robin_Watts | which seems tight to me. | 15:01.54 |
kens | Robin_Watts : yes tht does seem tight. Normally I woudl say its the airlines problem, but.... | 15:02.14 |
Robin_Watts | kens: If we miss the connection it'd be their problem to fly us out the next day, but it'd mean an overnight stop in LA, and would screw up plans for honolulu, so I'd really rather not. | 15:03.09 |
kens | Yeah, that was whgat I meant | 15:03.22 |
sebras | Robin_Watts: hi robing! | 15:03.36 |
tor7 | kens: okay. either is fine with me, if you can det PE on delta at a reasonable pricing (or was that only for the virgin leg?) then that'd probably be better than AC. | 15:04.13 |
kens | tor7 I'm certain of PE being available on the Virgin legs. They *say* they do PE on Delta, but I've learned to distrust US airlines (or indeed any airline, I prefer to stick with what I know). I'm sending you a PDF with what I found for Delta so you can consider it, you already ahve the options for Air Canada well sorted. | 15:05.32 |
tor7 | kens: fab. we'll need to ack with miles, but if we can get PE on the virgin legs I'm okay with that option, and I'll come over and visit you | 15:06.19 |
kens | Sounds good to me :-) I don't know what it will cost to do PE, but at least I know for sure it will exist ;-) | 15:06.48 |
chrisl | Delta seem to do "economy comfort" rather than "premium economy" | 15:08.25 |
kens | chrisl yes, but its +4 inches leg room | 15:08.38 |
chrisl | Yeh, I was trying to see if there was laptop power, too | 15:09.19 |
kens | I can't tell :-( | 15:09.31 |
chrisl | As usual, it seems on Delta that *some* economy seats on *some* aircraft have power..... not really helpful :-( | 15:12.01 |
kens | well the Virgin sections probably do | 15:13.57 |
| Do you want to see the PDF I sent to Tor ? | 15:14.10 |
chrisl | Sure, thanks | 15:14.38 |
kens | OK | 15:14.41 |
chrisl | So, I'm bored with comparing the shortest flight times, so I think we should also do the longest (single bookings only) - I've a found a 56 hour outbound one | 15:16.22 |
kens | O.O | 15:16.52 |
chrisl | I think the return was about 43 hours, too | 15:17.27 |
kens | for people who like flying obviously | 15:18.15 |
chrisl | I guess so. Opodo listed some fairly decent flight times, but I think they weren't really available as they disappeared when you try to book them | 15:19.13 |
kens | is bored with the whole flight thing | 15:19.19 |
Robin_Watts | circumnavigate, with a stop over in maui :) | 15:20.35 |
chrisl | Or maybe a cruise - lots seem to go to Maui.... | 15:21.14 |
Robin_Watts | chrisl: It goes without saying that if you wanted to come on the same dates as us etc, you'd be welcome to join us around honolulu etc. | 15:22.45 |
chrisl | Robin_Watts: thanks, I may take you up on that - I still need to checkup what family stuff I might be in for around then....... | 15:23.50 |
Robin_Watts | tor7: So, JNI bindings etc... I have been thinking about it. | 15:27.32 |
| In order for it to feel like a 'natural' java interface, we need the java objects to automatically reference count the underlying native ones. | 15:28.38 |
| This means we have to free the native object on the 'finalize' method of the java object. | 15:29.02 |
| Which means we need a context to be available within the finalize. | 15:29.19 |
| Which means either we need every object to be bound to a context (which is bad as the finalize can occur in any thread at any time, and hence could happen while the context is in use in another thread)... | 15:30.18 |
| or, we use ThreadLocal storage to keep a context per thread. | 15:30.48 |
| Every jni call would try to retrieve a context from the thread local storage. If it's there, great, just use it. | 15:31.20 |
tor7 | robin_watts: oh... trouble's brewing there I see... | 15:31.26 |
Robin_Watts | If it's not, make one. | 15:31.35 |
tor7 | I'd vote for TLS context with java | 15:31.44 |
Robin_Watts | store that one in the ThreadLocal storage, and use that one. | 15:31.51 |
| That way the context vanishes from the Java interface, which feels nice. | 15:32.08 |
tor7 | robin_watts: agreed. we might still want an "init mupdf context" function to set the store size though? | 15:32.38 |
Robin_Watts | tor7: I think that'd be a top level method (possibly a static one) on a MuPDF class. | 15:33.24 |
| I think the context will need to wrapped in a java object internally so it can get garbage collected (presumably the ThreadLocal stuff is smart enough to throw away all the thread local variables when a thread dies) | 15:34.31 |
| OK, so if you're unoffended with that, I'll try pursuing it for a bit. | 15:35.32 |
tor7 | go ahead. | 15:38.37 |
paulgardiner | Robin_Watts, tor7: maybe the store size could be held in a static member of the document class (or one of the other classes - whichever seems most appropriate) | 15:43.20 |
tor7 | kens: so, delta won't give me a direct CPH-LHR flight | 15:43.31 |
kens | Oh, but if you stay over with us that's OK ? | 15:43.51 |
tor7 | which means if we choose that option, I'll come to gatwick and visit | 15:43.52 |
kens | Sure, we'd be happy to see you | 15:44.00 |
| Air Canada and Delta are the best options I've found | 15:44.29 |
tor7 | kens: so if you go ahead and book that option, would it be better if you include me on the same reservation (in case there are changes etc)? | 15:44.49 |
| also might be trouble for me to book a trip starting from the UK if I'm located in sweden (not sure, haven't tried to do that before) | 15:45.42 |
kens | tor7 makes sense | 15:46.17 |
tor7 | kens: I'm off for dinner. we reckon we should ask miles and see what he says about PE/business before booking, though. | 15:49.26 |
| s/we/I/1 | 15:49.38 |
kens | makes snese too | 15:50.02 |
ray_laptop | I think most of the staff that were really against TLS are gone now. I tried to push TLS for our 'instance' needs to get rid of globals. IMHO, it's a LOT cleaner and quicker way to adapt non-threadsafe code than trying to bury the context in some structure that is usually available | 15:54.17 |
Robin_Watts | ray_laptop: I'm anti TLS in C. | 15:55.01 |
| It's the wrong choice for anything C based, due to the lack of garbage collection etc. | 15:55.20 |
ray_laptop | henrys was one of the staff that was really against it, but the rest of them went away. Note that I have no objection to using context if it's designed in, however | 15:55.22 |
Robin_Watts | and the fact that it just can't be done portably. | 15:55.40 |
| but given that it's inbuilt within Java/Android, it seems most sensible to use it. | 15:56.08 |
ray_laptop | robin_watts: I agree that it's not totally portable, but all systems I've seen have it | 15:56.23 |
Robin_Watts | it's probably got a large overhead though :( | 15:56.24 |
chrisl | Performance varies greatly with TLS across systems...... | 15:56.45 |
ray_laptop | well, with gs, the point is moot now, but the work to eliminate globals was rather painful | 15:57.04 |
| on the kernels I've written, TLS was always quite fast. It sort of falls out of the thread context switch | 15:58.09 |
| so I guess no meeting today ? | 15:59.28 |
mvrhel_laptop | I am guessing that too | 16:00.29 |
ray_laptop | mvrhel_laptop: I've been working on bug 694297. It's really strange. Somehow pdfwrite is deciding to not put 'q' 'Q' around a block that includes a 'gs' that sets some ExtGState stuff and that is making a subsequent dropshadow disappear | 16:04.32 |
mvrhel_laptop | sounds like kens needs to be dragged in... | 16:10.25 |
kens | :-( | 16:10.30 |
ray_laptop | hmm... looks like it's gs when displaying the PDF. Opening the 'bad' PDF with Reader is OK | 16:11.07 |
| mvrhel_laptop: sorry, but it's you (and I), and not kens I guess :-( | 16:11.27 |
| nm. I need to make sure I'm testing the right versions... | 16:13.20 |
| mvrhel_laptop: kens: OK. Confirmed. pdfwrite is OK. The problem is that when we render the PDF that pdfwrite makes, we get it wrong. | 16:15.12 |
kens | is happy to hear it.... | 16:16.20 |
ray_laptop | now to turn on some debug (-Zv -dPDFDEBUG). At least I've got two very simple files to deal with | 16:16.24 |
Robin_Watts | henrys is standing in a show at the moment. Given he's not here, I'd guess he's not got WiFi. | 16:17.12 |
| hence no meetings, probably. | 16:17.20 |
| Wasn't there a large chunk of agenda we never got to ? | 16:17.34 |
kens | I was assuming no meeting | 16:17.37 |
ray_laptop | robin_watts: If he really wanted to, he could tether like he did in the meeting | 16:17.41 |
| but I suspect that he has had enough meeting for a week | 16:18.03 |
Robin_Watts | actually, it was just "Some day projects", "Noted Accomplishments", "Added at/after meeting" and "High priority per engineer" that we didn't do, I think. | 16:18.59 |
kens | I think that's correct | 16:21.03 |
ray_laptop | I should suggest that he move Noted Accomplishments to the top (I think it used to be there), so that Miles and Scott and Juichi hear that part | 16:21.04 |
| just in case I ever make the list ;-) | 16:21.27 |
kens | plays sad violin for Ray | 16:23.36 |
ray_laptop | Oh, I did (along with kens, chrisl and the mupdf team). Now I'm really sorry that we didn't cover it. It means I have to find something notable to do before the next staff meeting | 16:23.37 |
chrisl | thinks: "Notable achievements: Ray - persuading henrys to move "Notable Achievements" to the top of the agenda again"....... | 16:25.49 |
ray_laptop | chrisl: Hey, I like that. A recursive achievement | 16:26.13 |
| or is it a meta-achievement ? | 16:26.29 |
mvrhel_laptop | ray_laptop: ok pushed my stuff so I can look at what you have now | 16:58.34 |
| please send me the file when you get a chance | 16:58.45 |
ray_laptop | mvrhel_laptop: OK, so the issue is OPM 1. With that, an RGB device doesn't show the SMask, but the tiff32nc device does. With OPM 0 and OP true, then BOTH devices fail to show the SMask. Acrobat shows the SMask whether or not I have "Overprint Preview" set to Never or Always | 17:07.15 |
| mvrhel_laptop: sending the file now.... | 17:07.22 |
| mvrhel_laptop: OK the file is on casper:/home/ray/public/bad.pdf It is sort of large since it is decompressed and it has a large image in it for the SMask | 17:09.03 |
kens | I'm puzzled as to why this exhibited a differnce with a change to clipping | 17:09.04 |
mvrhel_laptop | me too | 17:09.20 |
chrisl | The clipping change might have changed pdfwrite's emitting of q/Q pairs? | 17:11.38 |
kens | wwwwwwI guess that's possible | 17:11.58 |
ray_laptop | kens: with the clipping change, pdfwrite no longer puts the 'q' ... 'Q' around the execution of the part that executes the 'gs' with OPM 1 | 17:12.26 |
kens | Hmm, ok | 17:13.08 |
ray_laptop | kens: I think because it no longer sets the clippath before drawing that part | 17:13.13 |
kens | well at least that makes some sense | 17:13.26 |
ray_laptop | since it didn't reduce the clip path, it didn't need the gsave/gresotre | 17:13.35 |
kens | its just exposing a pre-existing problem | 17:13.47 |
ray_laptop | mvrhel_laptop: were you able to get the file ? | 17:14.01 |
mvrhel_laptop | ray_laptop: hold on a sec | 17:14.16 |
ray_laptop | line 81 of object 8 is the /OPM setting | 17:14.42 |
mvrhel_laptop | so is OP true in the file? | 17:15.10 |
ray_laptop | mvrhel_laptop: oops. I may have sent you the one with OPM set to 0. Yes, /OP is set true in that same object | 17:15.27 |
mvrhel_laptop | ok | 17:15.33 |
ray_laptop | since it's decompressed, you can easily edit it. | 17:16.05 |
mvrhel_laptop | ok I have the file. let me look it over | 17:16.30 |
| blending mode is multiply at some point | 17:17.12 |
| with an alpha of 0.75 | 17:17.32 |
| according to acrobat overprint is false | 17:18.12 |
| but I will dig into the file to see | 17:18.17 |
ray_laptop | mvrhel_laptop: I think that's after the problem, but not sure. From the RAW_DUMP files, the first differences are: 07)aTrans_Group_Pop_748x769x6.raw, 07)bImageTOS_748x767x6.raw and of course, 07)eComposed_748x767x5.raw | 17:18.26 |
mvrhel_laptop | ray_laptop: ok give me a sec. to run the file and look | 17:18.51 |
ray_laptop | mvrhel_laptop: sorry if I sent you a changed file. It is supposed to have /BM /Normal /OPM 1 /TK true /OP true /SA true | 17:19.56 |
mvrhel_laptop | ray_laptop: ok can you put the file that has the problem up there please | 17:20.25 |
ray_laptop | that way it's fine with tiff32nc, but not with tiff24nc (or any RGB device) | 17:20.27 |
| mvrhel_laptop: bad.pdf has the problem with gs. It generates a blank page with: gswin32c -r72 bad.pdff | 17:21.04 |
mvrhel_laptop | ok. so you are telling me that the file you gave me has issues with an rgb device? | 17:21.45 |
ray_laptop | but I'll put it up there again making sure I have the right ExtGState object 8 | 17:21.46 |
mvrhel_laptop | ray_laptop ^^ | 17:21.55 |
ray_laptop | mvrhel_laptop: correct. It is blank with RGB devices, but not with Acrobat even with "Overprint Preview" set to "Never" | 17:22.38 |
| mvrhel_laptop: Now it may just be that it is a valid Overprint difference with the RGB vs. CMYK devices, and that Acrobat is always doing Overprint | 17:23.32 |
mvrhel_laptop | file renders fine for me | 17:25.20 |
| ray_laptop ^^ | 17:25.24 |
| like I said, this file has no overprint | 17:25.42 |
| did you put the other file up there? | 17:25.52 |
ray_laptop | mvrhel_laptop: yes. What does line 81 have in the file that "has no Overprint" ? | 17:26.48 |
mvrhel_laptop | I dont know | 17:26.57 |
| I just ran it with gs out to a rgb device | 17:27.08 |
| and it was fine | 17:27.12 |
| getting other file now | 17:27.17 |
ray_laptop | mvrhel_laptop: so you saw the SMask ? | 17:27.27 |
mvrhel_laptop | ray_laptop: I see the same output that Acrobat has | 17:27.44 |
ray_laptop | mvrhel_laptop: sorry. Maybe that line and the /OP true line blanked out. | 17:27.50 |
mvrhel_laptop | ray_laptop: what is the name of the new file? | 17:28.42 |
| or is it the same name? | 17:28.47 |
ray_laptop | mvrhel_laptop: same name. bad.pdf | 17:28.54 |
mvrhel_laptop | ok | 17:28.57 |
ray_laptop | Ignore (or rename) the one that works if you want | 17:29.21 |
| if you look at "8 0 obj" you will see /OPM 1 as well as /OP true | 17:30.17 |
mvrhel_laptop | ok this file does have OP and OPM true | 17:30.35 |
| with multiply blending mode | 17:30.49 |
| and I get the blank page | 17:31.43 |
ray_laptop | mvrhel_laptop: I haven't touched the /BM /Multiply line. | 17:31.44 |
| mvrhel_laptop: I have to run something over to the school soon. | 17:33.44 |
kens | is off for dinner | 17:34.29 |
ray_laptop | mvrhel_laptop: I'll check back in a few minutes (going home to pick up what I need to take to the school)... | 17:35.04 |
mvrhel_laptop | ray_laptop : back to fixing my cie issue for a bit | 18:03.29 |
| hmm so there is a problem with my profiles generated from CIEA | 19:01.42 |
| lunch for a bit | 19:02.28 |
| ok I *think* I have this CIE stuff working now | 19:56.12 |
| with the white point fill-in off littleCMS turned off, a bug in the CIEA color space conversion was revealed | 19:57.03 |
| ok. now back to ray's bug | 20:01.20 |
| bbiaw | 20:41.33 |
Robin_Watts | marcosw: ping | 23:50.51 |
| Forward 1 day (to 2013/09/11)>>> | |