| <<<Back 1 day (to 2014/10/12) | 20141013 |
rayjj | found that ELF 2 is 64-bit -- I'll need to dig into the compiler/linker flags for pcl6. At least I have gs working so I can compare | 04:35.11 |
jogux | morning all | 10:11.31 |
kens | Morning | 10:11.36 |
paulgardiner | jogux: Your MuPDF commits on joseph/master look fine to me | 11:24.10 |
jogux | great, thanks Paul | 11:34.38 |
| morning tor8, thereâs an iOS 1.6 mupdf build on testflight if you fancy trying :) | 12:10.43 |
tor8 | jogux: seems to work fine. any new features I should look for? | 12:21.29 |
jogux | I donât think so. the only thing I fiddled with was to make it run at native res on iphone6/6+ | 12:21.55 |
tor8 | I don't have one of those so can't test that :) | 12:22.07 |
jogux | I figured :) | 12:22.31 |
kens | Miles is up early.... | 12:22.32 |
jogux | I guess Iâll submit it to Apple for review then | 12:22.37 |
tor8 | Miles asked for everybody's itinerary for december; when's everybody else going to be there? | 12:23.12 |
kens | I'm staing overnight 1 night, I'll travel up on the day of the first meeting (that's the 4th right ?) | 12:23.56 |
| But apparently Helen is trying to organise something at the Tower on the Monday so we may be there for that too. | 12:24.29 |
jogux | argh. I need some more screenshots before theyâll let me submit it. Gr. | 12:24.55 |
tor8 | kens: right, that's sort of what I'm wondering, if there's anything extra going on I want to arrive early/stay late for | 12:25.07 |
kens | So far, that's all I know. Apparently Robin mentioned it to me but I have no memory of that conversation | 12:25.37 |
tor8 | jogux: yeah, whenever they add new devices they want new screenshots at the new resolutions used | 12:25.39 |
jogux | Pete and I are heading down on Wednesday and back on Saturday (but Robin already communicated that to Miles a week or so back I believe) | 12:26.00 |
| tor8: just annoying the wait till the very last minute to flag it as an error ;-) | 12:26.20 |
Robin_Watts | tor8: My parents-in-law have an old-school-club outing thing going on to the tower of london on the monday after the staff meeting. | 12:26.31 |
| You get a tour of the tower after the public have left, fish and chip supper, open bar etc, and get to see the tower getting locked up. | 12:26.58 |
| Then you all get kicked out the side door at 9pmish. | 12:27.09 |
| I am not sure how many tickets they can get though. | 12:27.28 |
kens | thinks we should probably let the Jonny foreigners have first shout | 12:28.01 |
Robin_Watts | I *thought* I'd mentioned it to everyone, but possibly I may only have mentioned it to Americans. | 12:28.05 |
| kens: yeah, that was my logic. | 12:28.14 |
| If anyone is interested, please say, and i'll see how many tickets I can get. | 12:28.43 |
tor8 | Robin_Watts: yeah, that sounds fair enough :) We did a tower visit last time I was there. | 12:29.00 |
Robin_Watts | So far I have Scott+1, Henry+Sabrina, Marcos, and Ken+Stella as being interested. | 12:29.23 |
kens | I suspect Stella has already indicated interest to Helen, but like I said, tourists first | 12:29.24 |
tor8 | Robin_Watts: I'd be interested if you can get a ticket, otherwise I'll pass. I'll need to know soon to be able to book flights and let Miles know. | 12:32.22 |
Robin_Watts | yeah, I'll tell people asap, but it's reliant on people getting back to my parents-in-law, who are on holiday etc. | 12:33.08 |
tor8 | Robin_Watts: anything else planned for the saturday or sunday? | 12:37.53 |
Robin_Watts | tor8: Nothing specifically planned. | 12:38.10 |
| We'll be coming home on friday afternoon/evening. | 12:38.23 |
| Open invite for anyone who wants to come visit us over that weekend. | 12:38.37 |
tor8 | Robin_Watts: can you fit everyone? :) | 12:38.53 |
Robin_Watts | The brewery should be open for tours on saturday, and I can take people clay pigeon shooting. | 12:38.57 |
| If people want to stay, we'll probably resort to rooms at a local pub (nice place, henry and sabrina have stayed there before) | 12:39.25 |
tor8 | Robin_Watts: count me in, I could do with a bit of away-from-home time during the dark days of december | 12:40.09 |
Robin_Watts | cool. | 12:40.46 |
tor8 | and then I'll book a flight home on monday or tuesday, depending on the availability of tickets for the tower | 12:41.11 |
jogux | oh ffs. in the course of producing the screenshot Iâve found a bug. (luckily for me, Iâm betting itâs not iOS specific). | 12:48.38 |
tor8 | jogux: oh :( | 12:51.48 |
jogux | I made this happen once before and we couldnât reproduce :-S just writing it up | 12:55.25 |
| tor8: http://bugs.ghostscript.com/show_bug.cgi?id=695596 - Iâve not had a chance to investigate anymore (I shouldnât really be doing mupdf at allâ¦). Interestingly the annotated document there displays okay in os x preview, so it seems perhaps mupdf has created a file it canât read :-( | 13:06.46 |
tor8 | jogux: sounds like a bug for paul, but he's not supposed to be doing mupdf either :( | 13:07.32 |
kens | GS says it has an incorrect stream length and (possibly) a spurious ET operator | 13:08.54 |
| The stream seems to be borked | 13:11.42 |
jogux | as often seems to happen when I break things, itâs an artifex created document that I started with :( | 13:12.28 |
kens | The content stream ends: | 13:12.58 |
| 117.02 702.51 l | 13:12.58 |
| 117.126 702.49 l | 13:12.58 |
| 117.231 702.464 l | 13:12.58 |
| 117.337 70endstream | 13:12.58 |
| endobj | 13:12.59 |
| Whic looks wrong to me :-) | 13:13.07 |
jogux | :( | 13:13.11 |
| right, thatâs the 1.6 release submitted to Apple for review, anyway. (I presume this isnât a new bug and probably isnât iOS specific, so doesnât seem to be a pressing reason to not submit) | 13:15.33 |
kens | I think it may be caused by the fact that the page has several content streams | 13:15.35 |
| And they don't break in a logical place, it is busy constructing a path when it suddently stops the content stream and starts a new one, which then continues to appaend to the path. | 13:16.11 |
jogux | Producer: Acrobat Distiller 6.0 (Windows) â¦. | 13:16.56 |
kens | The rewritten PDF file only has one content stream for the first page, looks like it just didn't cater for the additional streams at all | 13:17.02 |
| jogux : yeah Distiller does this a lot for no obvious reason. | 13:17.17 |
jogux | hehe | 13:17.26 |
kens | Anyway, that's as far as I cna go with it. | 13:18.08 |
jogux | thanks Ken, Iâve dumped that onto the bug for future reference :) | 13:20.02 |
henrys | kens: I forgot chris is on vacation - put some changes for him to review in my repo, you'd be next in line to review them ... not important. | 13:46.27 |
kens | Hmm, I don't think I have your repository as a remote. I think maybe it would be better to wait for Chris unless you're in a hurry | 13:47.07 |
henrys | okay | 13:48.58 |
tor8 | henrys: you should run "touch /home/henrys/repos/ghostpdl.git/git-daemon-export-ok" on casper | 13:50.36 |
| so you can fetch using the git:// protocol | 13:50.46 |
| kens: you can run "git remote add henrys git://git.ghostscript.com/user/henrys/ghostpdl.git" to add henry's repo as a remote, once he's set that up | 13:52.21 |
kens | tor8 I can do that (I already have Chris's) I'm just dubious that I can usefully review a commit from Henry :-) | 13:53.09 |
tor8 | kens: that's a better objection -- i.e. one I can't do anything about ;) | 13:53.38 |
kens | Yeah sorry, I sort of put 2 problems in 1, and the second one is the more important..... | 13:54.13 |
| So I have henrys repo as a remote now | 13:54.51 |
| Might as well.... | 13:54.55 |
| Hmm so what should I be doing to get a copy of henrys repo ? I thought it was git remote but that fails | 13:56.00 |
| sorry git fetch | 13:56.05 |
| fingers running independently of brain.... | 13:56.16 |
henrys | I just review remotes on the web. | 13:56.58 |
kens | has no idea how to do that | 13:57.18 |
henrys | perfectly good diff tool | 13:57.20 |
| go to http://git.ghostscript.com/?p=user/henrys/ghostpdl.git;a=summary and click on commitdiff for the changes (example) | 13:59.05 |
| I wanted the NOCACHE ones reviewed | 13:59.38 |
kens | Hmm, I really do think it would be better for Chris to look at that | 13:59.48 |
henrys | sure np | 14:00.01 |
kens | Off the top of my head I don't see a problem, but I'm not really well up on the font cache these days | 14:00.20 |
henrys | tor8: I thought I was using the git protocol. | 14:01.04 |
| tor8: my remotes have git: prepended | 14:01.30 |
tor8 | henrys: yeah, but you need to create a blank file named git-daemon-export-ok in order for your repo to be available using the git: protocol | 14:03.26 |
| rigel ~/src/ghostpdl $ git fetch henrys | 14:03.48 |
| fatal: remote error: access denied or repository not exported: /user/henrys/ghostpdl.git | 14:03.48 |
| or you get messages like that | 14:03.57 |
kens | That's what I got | 14:04.03 |
tor8 | repos are not exported by default | 14:04.21 |
| (and as such, only available via ssh) | 14:04.37 |
| (which is dog slow in comparison) | 14:04.48 |
henrys | okay done, but that should be on by default for everyone | 14:05.16 |
kens | Aha, now I can fetch from henrys repo | 14:05.38 |
tor8 | henrys: yeah, it's an annoying default | 14:06.00 |
| no idea if you can change it in the git-daemon config somehow | 14:06.08 |
| should be possibly by passing the --export-all flag when starting dgit-daemon | 14:07.19 |
kens | Robin_Watts : ping | 14:07.38 |
Robin_Watts | pong | 14:09.08 |
kens | If you get bored sometime, could you have a quick look at the pdfwrite Bicubic downsampling filter ? I think it doesn't work | 14:09.39 |
| It processes the data, but the output is horrible | 14:09.50 |
Robin_Watts | I would expect the output to be horrible if it's downsampling by more than 1/2 | 14:10.22 |
kens | Its downsampling by 4.5 | 14:10.33 |
Robin_Watts | eq. | 14:10.38 |
| ew. | 14:10.41 |
kens | OK so I'm guessing that Acrobat is *actually* using the Average donsampling for very large downsampliung, even when Bicubic is requested then. | 14:11.15 |
| What a mess. | 14:11.30 |
Robin_Watts | hmm. maybe I'm being unfair. | 14:11.40 |
| can you make me a bug with a file and a command line and I'll have a look at some point? | 14:11.53 |
kens | OK thanks, I'll do that. | 14:12.14 |
tor8 | henrys: okay, I think I managed to change the default so git-daemon-export-ok won't be needed anymore | 14:12.26 |
henrys | tor8: great | 14:12.43 |
rayjj | morning, all | 14:18.04 |
kens | Morning | 14:18.17 |
rayjj | I assume that LHR is the best to come into ? | 14:18.19 |
kens | Do you have a choice ? | 14:18.31 |
| Don't go to City airport | 14:18.45 |
| If you come to LGW that's my local airport | 14:19.04 |
| I'd be surprised if you can go direct to Luton or Stanstead from LA | 14:19.25 |
rayjj | kens: I guess not if I want non-stop | 14:20.14 |
nsz | tor8: i ran the official tests from https://github.com/tc39/test262 | 14:20.22 |
| and they found many issues | 14:20.27 |
kens | rayjj That's pretty much whta I thought | 14:20.39 |
nsz | (these are pedantic conformance tests so probably not all failures are serious) | 14:20.54 |
| (and i found a bug in the ecma standard reading the fail logs) | 14:21.26 |
tor8 | nsz: I've run several test suites, with various degrees of success. a lot of them have plain out bugs and don't test what they aim to test (but succeed anyway due to lax implementations) | 14:22.47 |
| and others depend on non-standard features | 14:23.00 |
rayjj | wow! the "taxes and fees" are 46% of the total | 14:31.30 |
kens | They always are | 14:31.44 |
| Well, for us anyway | 14:31.50 |
rayjj | I'm used to domestic where they are about 20% | 14:32.10 |
| so, I have to travel on Tue to get there Wed afternoon for a Thur meeting :-( then I have to stay over Sat night or they want $500 extra. Seems like I get a free day Sat in London (or somewhere over there) | 14:34.08 |
kens | rayjj : Tha'ts why us Europeans usually stay over a Saturday night in the US | 14:34.36 |
jogux | $500 extra is relatively cheap tbh :) | 14:35.04 |
kens | Though the differential is usually higher for us, | 14:35.08 |
| TOr said his last trip the difference was only about 30% (IIRC), its usually double or more for us if we don't stay over Saturday, but I should check next trip to the US | 14:35.52 |
henrys | any thoughts on the best way to get from london to strasbourg? We looked at one airline that is booked until april. Seems like we are going to have to use a different airport too - not Heathrow | 14:39.29 |
kens | You can use the Eurostar, or if you need to go from Gatwick we'd be happy to put you and Sabrina up for the night (or more if you like). Same for anyone else | 14:40.09 |
| The Eurostar is about a 5 hour trip as I recall (train journey) and you have to go via Paris | 14:40.43 |
| But as I remember its quite comfortable | 14:41.05 |
jogux | henrys: if you use expedia.co.uk, that lets you use âany london airportâ as a âflying fromâ. if that helps. should search everything except ryanair/easyjet probably. | 14:41.31 |
Robin_Watts | henrys: https://www.google.co.uk/flights/#search;f=SXB;t=LCY,LGW,LHR,LTN,STN;d=2014-12-02;r=2014-12-05;tt=o | 14:42.02 |
| Eurostar may be the best option, as it looks like direct ain't an option. | 14:43.11 |
kens | Yeah I'm not seeing a direct flight option | 14:43.29 |
Robin_Watts | Also, whichever london airport you fly into you've then got to get into the centre. | 14:43.37 |
| Heathrow has the 'express' (35 quid each) and the tube (less, but takes longer, smells more) | 14:44.05 |
| Stanstead has an 'express' too, but it takes longer. | 14:44.18 |
tor8 | jogux: easyjet and norwegian are much better than ryanair though! but still best avoided for long trips :) | 14:44.20 |
kens | Gatwick also has an express, and regular trains, and is easier than Heathrow | 14:44.26 |
Robin_Watts | Luton is impossible to get to. | 14:44.34 |
nsz | tor8: yes webkit/v8/mozilla tests were hard to run | 14:44.36 |
tor8 | nsz: I've mainly been using the sputnik tests | 14:44.51 |
kens | To reiterate, if anyone needs/wants to fly from Gatwick we'd be happy to provide accomodation. We can sleep quite a few now Melanie isn't here :-) | 14:45.04 |
henrys | kens: thanks | 14:45.19 |
nsz | there are several issues in math and float handling | 14:45.20 |
| subnormal numbers are incorrectly rounded | 14:45.29 |
| and Math.round is wrong | 14:45.37 |
| but there were a long list of other issues | 14:45.47 |
henrys | argues with sabrina - I wanted to experience the train ride to paris she wants to fly direct | 14:46.00 |
tor8 | nsz: Math.round uses libc's round() function | 14:46.07 |
nsz | yes and that's wrong | 14:46.14 |
| js round is different for whatever reason | 14:46.24 |
kens | henrys the train ride isn't that exciting from London | 14:46.34 |
nsz | and the standard makes an incorrect note about the behviour | 14:46.34 |
pedro_mac | tor: I quite like Norwegian - best of the budget airlines Iâd say | 14:46.35 |
tor8 | which makes ECMAScript a stupid and annoying specification... :) | 14:46.35 |
nsz | (which demonstrates a significant lack of understanding of ieee754 semantics by the editor of the standard) | 14:47.20 |
tor8 | pedro_mac: Norwegian would be fine if it weren't for the extremely short seat pitch | 14:47.22 |
| pedro_mac: I'm 195 tall, which unfortunately throws me a CantFitInTheDamnSeatException on plenty of airlines | 14:48.00 |
nsz | http://www.ecma-international.org/ecma-262/5.1/#sec-15.8.2.15 | 14:48.48 |
| note 2 says floor(x+.5) is round(x) which is not the case whenever x+.5 involves a rounding according to the current rounding mode (nearest in js) | 14:49.42 |
tor8 | according to old mozilla javascript, round(x) = copysign(floor(x+0.5), x) | 14:55.51 |
henrys | kens:yes I watched a time lapse on you tube was hoping to see a lot of scenery while not in the tunnel, guess it isn't that scenic | 15:12.49 |
kens | Well, there's some at each end :-) | 15:13.06 |
Robin_Watts | henrys: My memory is that london->paris is very dull. | 15:13.14 |
| paris->strasbourg might be better? | 15:13.27 |
kens | Its better, but still somewhat dull as I recall | 15:13.41 |
| hours of farmland I seem to remember | 15:13.51 |
henrys | but it's something to do and the flights from paris -> strasbourg look good, we'll probably go with it and fly back | 15:15.28 |
kens | London -> Paris is fairly quick as I recall, a Paris->Strasbourg flight sounds good | 15:16.01 |
henrys | I've never been on a high speed train .. so probably more interesting to us than you. | 15:16.41 |
nsz | tor8: do you have reference for that? | 15:16.49 |
| or can i just find it in mozilla's codebase? | 15:16.59 |
tor8 | nsz: an old source code dump from before they rewrote it into c++ | 15:17.07 |
Robin_Watts | The trains aren't that high speed. | 15:17.16 |
tor8 | you can find it as jsref on various places on the web | 15:17.21 |
nsz | this is a huge mistake as far as precise arithmetics is concerned | 15:18.03 |
jogux | henrys: if you want high speed train and scenery, go to Japan ;) | 15:22.35 |
henrys | jogux I was in japan but always seemed to end up on pokey trains. | 15:24.41 |
jogux | :( | 15:24.47 |
nsz | src/regression/strverscmp.c:14: strverscmp("000", "00") < 0 failed | 15:38.41 |
| src/regression/strverscmp.c:15: strverscmp("00", "000") > 0 failed | 15:38.42 |
| src/regression/strverscmp.c:20: strverscmp("09", "0") < 0 failed | 15:38.42 |
| sorry | 15:38.48 |
mvrhel_laptop | good morning | 15:57.29 |
kens | Morning Michael | 15:57.50 |
dinamic_ | evening folks | 18:32.57 |
| i jsut stumble upon MuJS, and wants to know how i would store a C pointer (user data) with a state | 18:33.32 |
| i think the registry is what i want to use but i cant find out how to put my pointer into it.. | 18:34.10 |
Robin_Watts | dinamic_: The guy you want to ask is tor8, but he's gone for the night. | 18:41.33 |
| Try again tomorrow (in about 16 hours or so) | 18:41.42 |
dinamic_ | ah ok :/ | 18:41.50 |
| ill check back then.. | 18:42.13 |
Robin_Watts | you're quite welcome, I'm sure. | 18:42.52 |
mvrhel_laptop | bbiaw | 18:53.41 |
dinamic_ | i figured it out... | 21:53.42 |
| Way more easy to use mujs than spidermonkey, great project! | 21:54.31 |
| kudos to involved.. | 21:54.39 |
nsz | dinamic_: note that it is work in progress.. i started to use it recently and found various conformance issues in cornercases (they should be easy to fix though) | 22:01.04 |
dinamic_ | i will contribute patches when i stumble upon those corners... | 22:02.30 |
nsz | a way to stumble upon them is to run the official test262 suite :) | 22:03.47 |
dinamic_ | ah ok.. | 22:03.57 |
nsz | i see unicode handling issues (silly utf16 in utf8 of js), regex issues (silly cornercases), floating-point issues (even more sillyness) | 22:04.36 |
| so mostly the uglyness of js | 22:04.50 |
henrys | mvrhel_laptop: does your office subscription from joann work? Mine doesn't. | 23:05.14 |
Robin_Watts | henrys: Mine does. | 23:30.50 |
henrys | Robin_Watts: thanks I didn't ask you because I figured you had left for the day. Do you know if paul's is good too? | 23:37.07 |
Robin_Watts | I would assume so. | 23:41.56 |
| In what way doesn't yours work? It rejects the code? | 23:42.19 |
henrys | Robin_Watts: it doesn't recognize the user name | 23:42.44 |
Robin_Watts | henrys: ah, the username is a stupid long thing, right? | 23:43.05 |
| Let me dig mine up to compare with. | 23:43.11 |
| (but not in public) | 23:43.20 |
| Forward 1 day (to 2014/10/14)>>> | |