| <<<Back 1 day (to 2014/11/23) | 20141124 |
hjklmn | Hello, is this MUpdf channel? | 02:58.24 |
tor8 | Robin_Watts: a handful of commits on mupdf tor/master for review | 14:58.26 |
Robin_Watts | tor8: Did you see the MuPDF thing from the customer last week. | 14:58.46 |
| I sent an email a skype/remote debugging thing I did with them. | 14:59.01 |
tor8 | Robin_Watts: probably not, since I can't remember | 14:59.06 |
Robin_Watts | They have a file that has been incrementally added to 100s of times. | 14:59.29 |
tor8 | oh, that one! | 14:59.34 |
| yes, now I remember | 14:59.42 |
Robin_Watts | each one adds a new xref, alters object 1 and adds some new objects at the end. | 14:59.44 |
tor8 | paul did the incremental xref work, so I think we should consult him | 14:59.58 |
Robin_Watts | We could solve it by having the xrefs be a linked list. | 15:00.15 |
tor8 | I thought each new incremental section was limited to just the individual xref chunks not the whole object set | 15:00.23 |
Robin_Watts | Each section that read_old_xref reads is a full array from 1... xref_len-1 | 15:00.54 |
paulgardiner | robin: I remember you suggested a different memory allocation scheme... not for this reason, but to avoid pointers to the xref becoming stale | 15:01.07 |
tor8 | yeah. one linked list of xrefs for each 'level' of incremental update (or just one global, with the 'level' in each node) | 15:01.09 |
paulgardiner | I wonder if the same solution might give us sparseness | 15:01.20 |
| I think rather than a link list, before you suggested a an array of smallish arrays | 15:02.08 |
| Array of pointers to small arrays | 15:02.34 |
tor8 | another option (maybe stupid) would be to have a linked list of entries sitting off the main array (so each object has a linked list of its old versions) | 15:02.55 |
| and keep the head inline in the array, to avoid needless mallocs for the common file types | 15:03.30 |
| but that might wreak havoc with the existing code for incremental writes, etc? | 15:03.39 |
Robin_Watts | They have just skyped me again, asking for an ETA for a fix. | 15:03.40 |
tor8 | Robin_Watts: easiest fix -- make a flag whether to keep or flatten incremental xrefs | 15:04.14 |
| going back to the pre-paul behaviour when you don't care about incremental updates | 15:04.29 |
Robin_Watts | tor8: True. | 15:04.43 |
tor8 | which is only relevant for digital signatures | 15:05.03 |
| at least, as far as I can remember | 15:05.13 |
Robin_Watts | The only problem I can maybe forsee here... if we load a sparse xref from the file, then interpret the file, that will cause various objects to have xref pointers in them, right? | 15:05.37 |
| Then we do something that updates an object. This might pull an object into the latest xref. | 15:06.02 |
| If that's not in the range covered by the current xref, what do we do? | 15:06.18 |
| Do we 'flatten' the sparseness into a single block? Won't that leave stale pointers behind? | 15:06.46 |
tor8 | the objects keep pointers to the pdf_document rather than the xref don't they? | 15:06.53 |
Robin_Watts | tor8: Honestly, I can't remember. | 15:07.06 |
tor8 | I remember having an obj->xref pointer, but that was before we renamed pdf_xref to pdf_document | 15:07.19 |
Robin_Watts | Can we have 2 threads running at once? Probably not. | 15:07.21 |
paulgardiner | The stale pointers aren't fatal at the moment. I think we just considered them a potential trap | 15:07.24 |
Robin_Watts | stale pointers sound pretty fatal to me! | 15:07.39 |
paulgardiner | We don't ever hold onto them | 15:08.09 |
Robin_Watts | Or are you saying "we don't have the use case where stale pointers are an issue at the moment". | 15:08.11 |
| Right. | 15:08.14 |
paulgardiner | potential stale pointers not actual ones | 15:08.22 |
Robin_Watts | OK. So someone needs to spend some time on this. | 15:08.39 |
tor8 | touches his nose. | 15:08.47 |
Robin_Watts | It could be any of us, I think. | 15:08.53 |
tor8 | I'm neck deep in EPUB and HTML, and would like to be able to finish up enough to have a demo going by the meeting | 15:09.25 |
Robin_Watts | I'm happy to take some time off SOT to have a go, unless paulgardiner wants to make a bid for it. | 15:09.31 |
| tor8: I understand. | 15:09.36 |
henrys | wastes a lot of time campaigning for open source again, almost, but no luck. | 15:11.06 |
paulgardiner | I don't mind. I'm maybe getting somewhere with Symantec, so just at the moment less desperate to get away from SOT than usual | 15:11.15 |
kens | Henrys, maybe a bit more grief for Miles and Scott, and you'll get your way :-) | 15:11.37 |
Robin_Watts | We cannot release open source. | 15:12.20 |
| Our Patent License prohibits that | 15:12.34 |
kens | True. Free then. | 15:12.43 |
henrys | We'd go back to the patent holder and ask for a GPL exception for the patents. | 15:14.11 |
tor8 | henrys: like the DjVu patents? | 15:14.55 |
henrys | or raph's patents remember? | 15:15.09 |
tor8 | vaguely | 15:15.38 |
paulgardiner | open source sounds good: spread the pain! | 15:15.40 |
tor8 | but yes | 15:15.41 |
Robin_Watts | I would think that at least one of the companies attempting to compete with us are entirely unscrupulous enough to cherry pick fixes out and ignore the GPL. | 15:16.04 |
henrys | Robin_Watts: yes that's miles and the obvious argument against. | 15:16.36 |
jogux | it's possible that could actually make sorting things out easier. | 15:16.59 |
henrys | Robin_Watts: you were very much in favor of open source before and we knew that was a risk all along. Your opinion has changed? | 15:21.49 |
Robin_Watts | henrys: When I thought we were pulling the core into MuPDF, I was very in favour of Open Source. | 15:22.21 |
henrys | anyway miles isn't ready for it... but someday perhaps. | 15:22.37 |
Robin_Watts | When it became clear that we were working on the source 'as is', (and it became clear *quite* how unscrupulous our competition is), I think that idea falls by the wayside. | 15:23.23 |
chrisl | I doubt open sourcing would work with the build system as it is now..... | 15:23.27 |
henrys | I believe it would establish us the owners of the code, stealing would happen and we'd see it and make clear our competitors are violating the GPL. | 15:29.26 |
| jogux: what exactly did you mean? | 15:29.49 |
Robin_Watts | chrisl: Yes, because no open source projects ever have wacky build systems. | 15:30.05 |
| henrys: The people stealing the code would be just as free to claim "but we have a license through another route" then as they are now. | 15:30.51 |
chrisl | Robin_Watts: I just mean we'd end up spending *masses* of time talking people through using it - and we're not exactly flush with spare time as it is | 15:31.37 |
jogux | henrys: I think very similar to what you are thinking. | 15:31.48 |
| robin_watts: not for code that was created by Artifex. | 15:31.59 |
Robin_Watts | jogux: When we get to the stage where even 1% of the code has been created by Artifex, we can revisit the idea :) | 15:33.05 |
henrys | I think the owners will probably fall soon so you think btc is the problem? | 15:34.15 |
Robin_Watts | has been listening to predictions of the founders imminent demise for a very long time now. | 15:35.32 |
henrys | a few fixes cherry picked for gpl is not going to change a thing. | 15:35.54 |
Robin_Watts | Their lifespan appears to be about the same of that as that of a "short term hack so we can get it shipped". | 15:35.58 |
henrys | s/for/from | 15:36.01 |
| anyway I don't think miles is ready for this change. I'm sorry to hear you aren't for it Robin_Watts... I just assumed you would be on board with it mistakenly. It's not going to be a meeting topic this time around. | 15:38.30 |
Robin_Watts | henrys: I'd be on board if it weren't for the founders. | 15:39.03 |
| But they are (one of) the biggest challenges we face with making this stuff pay for itself. | 15:39.43 |
| Anything that helps them has to be bad. | 15:40.04 |
paulgardiner | Making it more obvious that we own it could be bad for them | 15:40.39 |
Robin_Watts | In getting people to jump from them to us, "they aren't legal" is the stick and "look at what we've fixed" is the carrot. | 15:40.42 |
| open sourcing would remove the carrot. | 15:40.54 |
henrys | Robin_Watts: no I disagree, open source is another carrot. | 15:41.40 |
paulgardiner | "They aren't legal" may be a huge stick for legitimate OEMs | 15:41.40 |
henrys | I do believe it will establish us as the owners. | 15:42.40 |
Robin_Watts | It would help us appear to be the owners. | 15:44.42 |
| But would have no legal basis for improving stuff that I can se. | 15:44.59 |
henrys | paulgardiner: indeed it is, and I think going open source gives us the grounds to make theat sort of claim. The claim is weak now. We are just a player. | 15:48.00 |
jogux | henrys: wouldn't "one of the other things Miles is doing just now" fairly clearly set us apart as the owners too? | 15:49.43 |
kens | reboot brb | 15:50.25 |
henrys | jogux: yes it would, we did agree that a successful outcome there would be prerequisite to open source. | 15:50.37 |
jogux | nods | 15:50.46 |
Robin_Watts | henrys: If we get a successful outcome to the thing-that-miles-is-doing-now (and I'm thinking of the same thing-that-miles-is-doing that you two are), then I'd be in favour of open source. | 15:51.55 |
henrys | Robin_Watts: good to know thanks. | 15:52.45 |
jogux | :) | 15:53.03 |
henrys | jogux: thanks for bringing that up I should have said something. | 15:53.13 |
henrys | canceled the meeting tomorrow but that isn't the Tuesday before the meeting. so should we miss 2 meetings? I looked at my calendar and mistook my flight day for the meeting day. | 16:03.24 |
Robin_Watts | henrys: Let's have a meeting, g'wan.... | 16:03.48 |
kens | doesn't mind either way, technically we should meet | 16:04.07 |
henrys | I'll send out email. | 16:04.30 |
chrisl | thinks Robin_Watts is clearly demented..... or replaced..... Stepford Robin_Watts..... | 16:04.45 |
mvrhel_laptop | so there is a meeting tomorrow? | 16:47.23 |
kens | there is now | 16:47.29 |
| Henry sent an email | 16:47.34 |
Robin_Watts | mvrhel_laptop: There is. When are you flying? | 16:47.35 |
mvrhel_laptop | Thursday | 16:47.50 |
henrys | mvrhel_laptop: if you made other plans no big deal. | 16:48.06 |
mvrhel_laptop | managed to catch a cold. hoping to get rid of it before then | 16:48.17 |
| no I will be here (at meeting) | 16:48.24 |
| Robin_Watts: so we have two packages that came for helen that we will be bringing | 16:48.43 |
Robin_Watts | mvrhel_laptop: Thanks. | 16:49.07 |
mvrhel_laptop | brb. need to reboot this machine. it is misbehaving | 16:49.46 |
henrys | on tuesday the 9th I have to get to London St. Pancras train station by 9:15am from the hotel. Any transportation suggestion? | 16:51.55 |
kens | Which hotel ? | 16:52.09 |
| cab is probably easiest | 16:52.16 |
henrys | miles' hotel in hyde park. | 16:52.29 |
| hilton | 16:52.34 |
kens | I owuldn't fancy the Underground at rush hour | 16:52.35 |
Robin_Watts | henrys: cab. That'll be rush hour. | 16:52.38 |
henrys | Robin_Watts: that car company you recommended wouldn't do that sort of trip? | 16:53.19 |
Robin_Watts | henrys: The car company I found was specifically an airport car company. | 16:53.36 |
henrys | ah | 16:53.46 |
Robin_Watts | But they might do transfers like that too. | 16:53.50 |
kens | You could get a minicab, but I'd suggest a black taxi myself | 16:53.55 |
mvrhel_laptop | Robin_Watts: so I booked a car through them for when we arrive | 16:53.57 |
| rather reasonable | 16:54.02 |
| thanks for finding that | 16:54.06 |
Robin_Watts | (and I have no personal experience of that company, just that they seemed remarkably reasonable!) | 16:54.16 |
| mvrhel_laptop: No problem. | 16:54.24 |
mvrhel_laptop | This will make life easier with two jet lagged kids + luggage | 16:54.48 |
kens | I just checked the Tube map, take a cab :-) | 16:55.43 |
henrys | kens: will do. | 16:56.17 |
kens | You owuld need 5 stops on the Central line, then 3 on teh Picadilly, both lines will be massivley busy at that time | 16:56.51 |
henrys | although Robin_Watts has it stuck in my brain I should visit every tube station during my visit, but sabrina is removing that brain section. | 16:56.56 |
kens | *every* tube stop would be quite a challenge..... | 16:57.29 |
| Especially the limited opening hours ones (and indeed, teh closed ones ;-) | 16:58.11 |
henrys | I got this crazy idea I wanted to do a slow "Tube Challenge" | 16:58.31 |
kens | You would probably need the week just to do it, some of the liones go out a long way | 16:58.49 |
| lines* | 16:58.53 |
henrys | we should set up the optimization problem and use the clusters ;-) | 16:59.25 |
kens | Hmm, osmone managed it in 18 hours, I am impressed | 16:59.36 |
henrys | certainly NP | 17:00.01 |
kens | But that was in 1960, there may be more stations now | 17:00.06 |
| wow, 14 hours..... | 17:00.33 |
henrys | http://www.thetubechallenge.com/ | 17:00.39 |
kens | Well that's more recent than wikipedia's entry | 17:01.05 |
Robin_Watts | 14? 16 hrs 23 mins or something? | 17:01.27 |
henrys | kens: cab it is thanks guys | 17:01.39 |
kens | wikipedia has a 1960's time at 14 hours something | 17:01.51 |
| 4th July 1964; 14 hours 17 minutes | 17:03.17 |
| But of course, the network does change. | 17:03.36 |
henrys | kens: I think I read something about the rules changing and the 14 is no longer considered good for guiness I forget what it was. | 17:03.43 |
kens | Its possible, also do they include the Docklands Light Railway for example. | 17:04.09 |
Robin_Watts | kens, right. The site you linked figures out the "probable best possible time" as being 15hrs 18mins. | 17:04.44 |
kens | that's henrys link :-) | 17:04.56 |
| ANd I notice it specifically excludes the DLR in the rules section | 17:05.07 |
| jogux : pedro_mac just had an email from Miles asking you both to join us for dinner on THursday of the meeting, and would you liek to bring along partners/children. I'll forward his email | 17:07.46 |
henrys | anyway sabrina has just about nixed my tube dreams ... it would be warmer than parading around outside but she wasn't convinced that was enough ;-) | 17:08.02 |
pedro_mac | kens: thatâs very good of him; I think weâll be there (and probably joguxâs wife) but Iâll be on my own | 17:09.10 |
kens | NP I've forwarded the email to you both, if you could reply to Miles that would be best | 17:09.31 |
pedro_mac | nods - thanks | 17:09.39 |
kens | Does anyone know *which* Hilton at Hyde Park we are going to ? | 17:10.18 |
| I seem to recall there were 2 | 17:10.24 |
jogux | Bayswater I believe | 17:10.29 |
Robin_Watts | kens: The hilton, not the doubletree. | 17:10.32 |
kens | Right, that's fine, thanks | 17:10.39 |
Robin_Watts | I believe mvrhel_laptop is at the doubletree though. | 17:10.45 |
| they are only 200 yards apart. | 17:10.56 |
kens | Indeed, but it's nice to go to the correct one :-) | 17:11.10 |
jogux | kens: thanks | 17:11.18 |
| henrys: aw. still, plenty of other fun things to do in London :-) | 17:11.30 |
kens | wonders if there's an endurance record on the London Eye | 17:11.53 |
jogux | kens: there are emergency toilets and supplies in each capsule ;-) | 17:12.38 |
kens | So I've heard :-) | 17:12.48 |
pedro_mac | jogux: when you say âemergency toiletâ do you mean âhole in the floorâ ? ;) | 17:13.28 |
jogux | :) | 17:13.39 |
kens | We could have leased one for the meeting | 17:13.45 |
chrisl | What, an emergency toilet?? | 17:13.59 |
kens | Hmm they have an ice rink there this year too | 17:14.44 |
Robin_Watts | kens: in every pod? :) | 17:14.53 |
kens | THat wuold be chilly..... | 17:15.07 |
henrys | a toilet on the eye? | 17:15.22 |
kens | THough apparently the pods are decked out in a 'frosty' theme | 17:15.32 |
henrys | according to the forecast the rains start the day I arrive, just like last trip. | 17:20.56 |
kens | I wouldn't worry, it will be raining long before you get here | 17:21.15 |
jogux | henrys: on the plus side, the long range forecasts are often fantastically incorrect | 17:21.33 |
kens | suspects predicting rain in London during December is a safe bet | 17:22.01 |
jogux | kens: snow and thunderstorms remain a distincy possibility ;-) | 17:22.36 |
henrys | temperatures seem to be holding above freezing though .... even the lows | 17:22.46 |
kens | Snow is possible, but really pretty rare | 17:22.49 |
| henrys it usually is above freezing in December, most of the time | 17:23.05 |
| Especially in the city which is a couple of degrees (Centigrade) warmer than the country | 17:23.24 |
| It may not feel much warmer by the time you add the wind and rain though | 17:23.53 |
jogux | looks like it'll (just) be warmer than Glasgow | 17:24.25 |
kens | THe UK residents did point out the likely weather when Miles gave us the destination :-) | 17:24.53 |
| Admittedly, we didn't protest very hard, but then weighing a night in London against a long-haul flight might have been influencing us slightly | 17:25.27 |
jogux | :) | 17:25.40 |
chrisl | I think the word I used for the likely weather was "grim"..... | 17:26.08 |
henrys | I do think there is something special about London around christmas time... possibly it is an american thing - | 17:27.07 |
kens | At least Henry won't have to shovel snow (though getting to his front door when he goes home might be tricky) | 17:27.32 |
| (of course, now that I've said that it'll dump a couple of feet, just to prove me wrong) | 17:28.34 |
| OK time to go, goodnight all | 17:29.39 |
Robin_Watts | Who will be around for food on friday night? | 17:30.36 |
| We'd planned to be there all day so we could go out and get an early meal on friday with you all before heading home. | 17:31.21 |
| I suspect if there is a group of us heading out for food we really ought to book. | 17:31.46 |
chrisl | It sort of depends on what time we finish on Friday - if we finish early, I was thinking of heading off home..... | 17:32.37 |
Robin_Watts | chrisl: I was vaguely hoping you might head out with me to a music shop... | 17:33.00 |
chrisl | Robin_Watts: well, that would make it worth staying until Saturday :-) | 17:33.28 |
chrisl | thinks: must leave credit card at hotel....... | 17:34.14 |
mvrhel_laptop | Robin_Watts: we (the vrhel's) will be around for food friday night | 17:37.18 |
jogux | robin_watts: Yukko & I will probably be around then | 17:42.33 |
| I think I'd probably say "will be very lucky if we have an option other than very expensive fixed-price menu turkey" rather than "probably should book" (given it's the 3rd Friday before christmas we're well into corporate christmas party territory...) | 17:44.08 |
Robin_Watts | jogux: Indeed. Turning up with a party of 12 will just get us laughed at unless we have a reservation. | 17:44.57 |
pedro_mac | Robin_Watts: Iâll be around Friday too | 17:45.01 |
Robin_Watts | R/H, J/Y, P/E, M/S+kids is 10. | 17:45.43 |
| I think Paul and Linda will be around then too. | 17:45.56 |
pedro_mac | Iâll be on my own. Its better that way ;) | 17:46.09 |
Robin_Watts | Scott + 2 maybe. | 17:46.12 |
chrisl | Robin_Watts: if we're hitting Denmark Street, you better count me in, too | 17:46.28 |
Robin_Watts | chrisl: Depends how long you think Denmark Street will take, but cool. | 17:46.51 |
pedro_mac | reckons I havenât been to Denmakr street before | 17:47.07 |
henrys | hmm I thought helen and sabrina were talking about something for friday night that was already set up. Did I miss somethingj? | 17:47.13 |
jogux | ditto | 17:47.14 |
| hm. I bet I never gave Yukko Helen's email address. oops. | 17:47.37 |
chrisl | Robin_Watts: well, I don't want to be getting the train after about 4pm on a Friday | 17:47.47 |
Robin_Watts | henrys: Helen + Sabrina had vaguely discussed the possibility of going to a 'nice' restaurant on Friday. | 17:48.22 |
| where nice is "nicer than we could expect Miles to pay for". | 17:48.38 |
| Like say Simpsons etc. | 17:48.47 |
| But I think we'd decided that it would be bad form to just head off and leave everyone else. | 17:49.15 |
jogux | robin_watts: well a "nice" restaurant would be okay with me (not entirely sure how much this might end up costing me :) ) | 17:51.40 |
henrys | Whatever you guys think is best sabrina sent me a mail from Helen about going to Simpsons or a few other choices. | 17:51.53 |
Robin_Watts | I suspect it's really too late to get a table at Simpsons (certainly for a large party) | 17:52.43 |
jogux | robin_watts: partly depends on time | 17:53.36 |
| (they have tables for 6, the most you can check online) at 18:00 or 20:30) | 17:53.50 |
Paul-the-Norther | Hello! I'd like to express my appreciation to MuPDF developers. I'm grateful to you! Your app is the best existing. Society needs you. | 17:54.11 |
henrys | we were looking at doing a cruise on the river Friday (with dinner) it's a bit expensive but it might be fun. | 17:54.12 |
Paul-the-Norther | From Russia with love. | 17:54.37 |
| Sorry for interrupting | 17:54.53 |
henrys | we are going out thursday together then I'll be locked in a tower with all of you, that might be enough ;-) | 17:55.08 |
| I'm doing this non Artifex sanctioned activity http://www.alternativeldn.co.uk/pub_tours.php saturday and 3:30pm if anyone would like to join. Probably a little touristy for the UK'ers. | 17:58.49 |
Robin_Watts | henrys: The river cruise sounds like a nice thing to do. You shouldn't feel tied to stay with the rest of us - you see us all the time, but are only in London rarely. | 17:59.44 |
henrys | s/and/at | 17:59.53 |
| Paul-the-Norther: thank you | 18:00.33 |
jogux | henrys: I'd be up for that if I was staying long enough. darn :) | 18:01.01 |
henrys | I see trying to get rid of me... | 18:01.17 |
| I think we will do the cruise. | 18:01.42 |
Robin_Watts | henrys: Oh, right. That tour will probably be bethnal green way. | 18:11.30 |
| Near the London hospital, where Jack the Ripper was operating. Plus the crays, and the pub that had a famous siege. | 18:12.04 |
sebras | __tim: if you are around, do you have any opinions on this patchset? https://bugzilla.gnome.org/show_bug.cgi?id=727826 (the first ought to be quite noncontroversial, the fourth is the interesting one) | 19:53.55 |
| __tim: slomo has already looked at it a bit, but I think wanted you to take a quick glance at it too. | 19:54.47 |
| ehm... | 19:55.53 |
| gstreamer, ghostscript... same, same right..? d'oh... | 19:56.22 |
| Forward 1 day (to 2014/11/25)>>> | |