| <<<Back 1 day (to 2013/11/24) | 2013/11/25 |
Amr_ | hello, I am asking about MuPDF for Android, can MuPDF merge pdf files into one file and open it? | 09:23.30 |
kens | As far as I'm aware, no, MuPDF doesn't mergre PDF files at teh moment | 09:24.32 |
Robin_Watts | Amr_: As Ken says, no. not currently. | 09:26.58 |
kens | chrisl did you see the3 mail from Takane ? | 11:28.11 |
| Oh, it didn't go to everybody | 11:28.38 |
| Forwarded it to tech.... | 11:29.12 |
chrisl | I'll read it when I'm done on the bike..... | 11:30.31 |
kens | OK | 11:30.38 |
paulgardiner | tor7: Did you have to use XCode 5 to test MuPDF on an iOS 7 device? | 11:43.34 |
tor7 | paulgardiner: no, sorry. | 11:44.34 |
paulgardiner | So there's a trick to using Xcode 4? | 11:45.05 |
tor7 | my sister's computer died so I lent her my old mac laptop | 11:45.07 |
| paulgardiner: you can install Xcode 4 and Xcode 5 side by side | 11:45.17 |
paulgardiner | yeah. That's what I'm doing. Just wanted to check it was really necessary. | 11:45.40 |
| ... that there wasn't some sort of trick to using Xcode 4 instead. | 11:46.03 |
tor7 | I think you need Xcode 5 to target ios7, but more importantly, to run the ios7 simulator | 11:46.11 |
| it might be possible to compile with xcode4 and use the simulator from 5 | 11:46.29 |
paulgardiner | Okay. No probs. Downloading it now | 11:46.33 |
| I need the 7 SDK in any case for access to javascriptcore | 11:47.20 |
chrisl | kens: how can we be expected to give a price with no clue as to the problem?? | 12:16.14 |
kens | :-) | 12:16.43 |
Robin_Watts | chrisl: I read the email as asking for a generic support contract. | 12:17.05 |
kens | I wa interested that they would want to licence GS commercially too. So I guess the 'Eagle' rip can't distill either | 12:17.11 |
| Robin_Watts : there was mutterings about licencing GS commercially | 12:17.32 |
Robin_Watts | kens: Indeed. | 12:17.39 |
| but we (well, scott/miles) have pretty standard pricing for supported GPL/supported commercial licenses. | 12:18.06 |
chrisl | Robin_Watts: well, I guess that makes more sense - but I sort of read it as the licensing was conditional on resolving the issue...... | 12:18.18 |
Robin_Watts | chrisl: I think that's down to Scott/Miles to negotiate. | 12:19.19 |
chrisl | Still, looks like it could be a good deal..... | 12:19.46 |
Robin_Watts | but I'd figure on 'sign the deal, then we'll try and fix it. If we can't solve it, then you can terminate the deal early'. | 12:19.47 |
kens | I was mostly amused that a PS/PDF tip manufacturer wanted to licence GS for PDF production (possibly) and PDF->JPEG | 12:20.53 |
chrisl | I wonder what "Eagle" is based on nowadays..... | 12:22.58 |
kens | I was told it | 12:24.05 |
| was rewritten by students | 12:24.12 |
chrisl | Yes, I heard that as the "official line", but there were rumours of there being more involved than that.... but then that came from GG, so..... | 12:25.25 |
kens | Its hard to know, but it seems like it can't be a distiller either :-) Or posibly can't render PDF to JPEG (which would be somewhat surprising) | 12:26.17 |
Robin_Watts | kens, chrisl: So, am I right in thinking that the gs device interface relies on 'raster' being a multiple of bitmap_raster? | 12:29.58 |
kens | I have no idea | 12:30.08 |
Robin_Watts | i.e. all the calls assume at least word alignment? | 12:30.14 |
| Certainly that was the impression I had. | 12:30.20 |
chrisl | Robin_Watts: I think is normal, yes. | 12:31.18 |
Robin_Watts | It's certainly "normal". The question is whether it is required or not. | 12:31.39 |
chrisl | I'm not sure, but I think it is required, based on vague memories of the threshold array rendering code | 12:33.04 |
kens | heads off to lunch | 12:33.21 |
Robin_Watts | bug 694601 is caused by the thresholding code calling clip_copy_mono with a raster of 2. | 12:33.46 |
| The reporter is saying that on big endian platforms it needs to be changed to be a raster of 4. | 12:34.15 |
| but I think that actually, it should be 4 on ALL platforms. | 12:34.24 |
| and we are just getting away with it on x86 etc because that can cope with unaligned accesses. | 12:34.39 |
chrisl | Isn't it already 4 on little endian platforms? | 12:34.54 |
Robin_Watts | chrisl: Search for "LAND_BITS" | 12:35.09 |
| LAND_BITS is defined to be 16 globally. | 12:35.20 |
| in gsiparam.h | 12:35.24 |
chrisl | Hmm, okay, that's not the part of the thresholding code I was thinking of..... | 12:38.34 |
| Robin_Watts: but it also depends on how you define "word"..... | 12:39.56 |
Robin_Watts | where do you see 'word' ? | 12:40.10 |
chrisl | From your question above | 12:40.25 |
| We might might a raster of 4 or 8 | 12:41.11 |
Robin_Watts | chrisl: right, sorry, I was being informal there. I'm searching for a formal definition. | 12:41.29 |
| bitmap_raster aligns to either 4 or 8 depending on platform, currently (though in theory it can be 1,2,4 or 8) | 12:41.53 |
| In gxbitmap.h: | 12:42.33 |
| * Most graphics library procedures that process bitmap data (such as, for | 12:42.43 |
| * example, the "device" procedures in gdevm*.c) impose two requirements | 12:42.45 |
| * on such data: an alignment requirement, and a padding requirement. | 12:42.47 |
| * Both requirements arise from the fact that these library procedures | 12:42.48 |
| * attempt to process the bits in units of align_bitmap_mod bytes. | 12:42.50 |
| But I was hoping for a definition within the documentation. | 12:43.07 |
chrisl | Are you still that naive? ;-) | 12:43.23 |
Robin_Watts | :) | 12:44.10 |
chrisl | The only worry is that LAND_BITS is clearly, and explicitly set to 16 - so is that a mistaken, and anachronistic constant, or a genuinely intentional setting. | 12:45.40 |
Robin_Watts | LAND_BITS can safely be configured to more than 16. | 12:49.30 |
| I'm thinking that perhaps we should set it to match bitmap_raster | 12:49.49 |
chrisl | Robin_Watts: yeh, I was just wondering: for the "main" rendering code, Jaws used rasters aligned to "words", but for the glyph rendering it used half-word alignment, to reduce wastage on smaller (internal only) bitmaps, I just vaguely wondered if this was intended as something similar | 12:52.52 |
Robin_Watts | chrisl: This is for handling landscape bitmaps. It's how many bits (rows) it collects into a buffer before sending it through. | 12:53.58 |
| so any multiple of 16 is fine for the code; 16 is optimally small, and (in tests) fastest. | 12:54.38 |
Guest38103 | hi is this mupdf forum? | 12:54.40 |
Robin_Watts | Guest38103: Sure. | 12:54.45 |
Guest38103 | Here is my question. We built an app that lets users read epubs bought from our website. the app is on android and ios | 12:56.22 |
| now we are planning to let users side load their own docs to read. This might include pdf. so we were thinking of using mupdf for that. Not sure about AGPL . Any suggestions | 12:57.07 |
| this support is only for user docs as we don't sell pdfs | 12:57.32 |
| so in other words it is strictly for user benefit not commercial use | 12:57.53 |
Robin_Watts | Guest38103: If you link MuPDF into your app, then you will either need to abide by all the terms of the AGPL, or you will need a commercial licsense. | 12:58.33 |
Guest38103 | how can we get commercial license? do you by any chance know how much does it cost? | 12:59.01 |
Robin_Watts | The AGPL would force you to release your source code (for all your app) to customers. | 12:59.19 |
| To get a commercial license, mail sales@artifex.com and describe your project in as much detail as possible. | 12:59.45 |
Guest38103 | yes, that is something we don't want to as we use some custom code to handle DRM for the books sold by us | 13:00.06 |
Robin_Watts | Scott will reply with questions to try and judge your situation. You answer them as best you can. Then he'll give you a quote. | 13:00.20 |
Guest38103 | sure. will do. Thanks for the help! | 13:00.22 |
Robin_Watts | no worries. If you don't get a reply from scott, then come back here and talk to us - sometimes mail seems to have a problem getting through. | 13:00.49 |
| He is generally quite fast at replying. | 13:00.55 |
| m"cb00k | 13:01.01 |
Guest38103 | sure. | 13:01.29 |
tor7 | Robin_Watts: time to change a password? ;) | 13:01.45 |
Robin_Watts | tor7: yeah. at least I mistyped it :) | 13:01.58 |
paulgardiner | tor7: when I did exactly the same thing, henrys commented "Entropy looks good, but you might want to work on usage" :-) | 13:15.37 |
Robin_Watts | :) | 13:15.55 |
tor7 | :) | 13:15.56 |
Robin_Watts | chrisl: Bug 694807 seems to be assigned to HinTak - is that a hangover from his time with us? | 13:41.35 |
kens | I guess so, its another daft bug anyway | 13:42.20 |
Robin_Watts | It's less daft than some others, possibly, but yes. | 13:42.44 |
| "People can get directory listings on your download site". Erm... yes? | 13:43.05 |
chrisl | I suspect it's because GhostSVG is dead, and no one "owns" it now, so it's never been changed. | 13:43.41 |
kens | We should take GhostSVG off Bugzilla too. | 13:44.08 |
Robin_Watts | It's an "SVN" bug, not an "SVG" one :) | 13:44.21 |
chrisl | Any objection to just closing it? | 13:44.50 |
Robin_Watts | chrisl: If you're satisfied with it, sure. | 13:45.01 |
kens | Suits me | 13:45.04 |
chrisl | Well, the vulnerability, AFAICT, is that can anyone can information that's public anyway..... | 13:45.39 |
Robin_Watts | well, the only potential problem is that people might be able to see old versions of the websites. | 13:46.00 |
| but it's a public svn server anyway, right? | 13:46.07 |
chrisl | Yep | 13:46.13 |
Robin_Watts | so, yeah, just close it. | 13:46.18 |
chrisl | I remain bemused at the number of people who seem to have time to fart about reporting that kind of stuff | 13:47.57 |
henrys | Robin_Watts: I wonder if a cluster push with an alignment assertion would be good for that bug you just fixed. Or do you think that is the only client that doesn't set up the data properly | 14:02.30 |
| ? | 14:02.32 |
kens | chrisl, that's pretty amusing from Adrian | 14:08.03 |
chrisl | kens: amusing/frustrating... one or the other :-( | 14:08.51 |
kens | Is it possible his executable is using system shared objects ? | 14:09.20 |
chrisl | Other glibc, I don't think GhostPCL uses much in the way of libraries | 14:10.13 |
kens | Not FreeType ? (just asking...) | 14:10.41 |
chrisl | It only links to three libs: libpthread, libm, and libc..... The GhostPCL builds don't have a built-in way to use the shared libs, like gs does | 14:11.30 |
kens | Fair enough | 14:11.37 |
chrisl | My suspicion is that he's using a different test file, a different command line, or (in error) a different executable | 14:12.23 |
kens | aha, and ooops. If I reuse a shading, I need to add it to the page resources..... | 14:15.40 |
henrys | chrisl: oh yes I accidentally removed that shared library feature ;-) | 14:16.54 |
chrisl | henrys: just wait - if the LInux distros want to start shipping GhostPCL/GhostXPS, see what happens then! | 14:18.02 |
henrys | chrisl: I've only been skimming the adrian discussion but this customer has a long history of getting their money's worth from us. | 14:20.34 |
chrisl | henrys: I'll try to be patient. The (Windows only!) "system sharing" option he offered isn't especially good for debugging a Linux/AIX executable..... | 14:22.54 |
Robin_Watts | henrys: sorry, was grabbing lunch. | 14:25.01 |
| You may be right, but where would I put the assertion? | 14:25.22 |
henrys | the first line of mem_mono_copy_mono() I thought | 14:28.38 |
Robin_Watts | ok, but there are many other places where rasters are passed. It would be nice if there was some central way of testing them all, but I can't see a way of doing that. | 14:29.21 |
henrys | Robin_Watts: no I don't see how we could do that | 14:30.12 |
Robin_Watts | Rather than doing: dev_proc(dev, copy_mono)(...) we should be doing: device_copy_mono(dev, ... ) | 14:30.54 |
| with an appropriate #define for device_copy_mono. | 14:31.08 |
| That way we could tweak the #defines. | 14:31.19 |
| but that train sailed a long time ago. | 14:31.26 |
| oh, bugger, I think I screwed up that commit. | 14:32.28 |
| align_bitmap_mod is in bytes, not bits. | 14:32.39 |
| I will fix. | 14:32.47 |
chrisl | I guess I could periodically run tests on my sparc - I need to look at that at some point | 14:32.55 |
henrys | chrisl: a sparc cluster node! | 14:35.12 |
| at chrisl's place | 14:35.26 |
chrisl | henrys: we discussed that before - mine are too slow. By the time they'd built the exes, the other nodes would be finished testing! | 14:35.54 |
Robin_Watts | henrys: trying a cluster test now. | 14:38.09 |
henrys | chrisl: oh that's right | 14:38.19 |
| chrisl: we could schedule an overnight | 14:38.48 |
chrisl | henrys: what I suggested (at least twice) was that I run marcosw1's weekly test script, but I've never seen it to try | 14:39.48 |
henrys | chrisl: it must be frustrating being surrounded by old farts that forget everything ;-) | 14:40.45 |
chrisl | henrys: if I find myself with nothing to do, I'll nag about it! | 14:41.30 |
| My own test script(s) would be usable, but I wanted to tweak them so I could "pause" testing part way through - the sparc's a pretty noisy | 14:41.38 |
Robin_Watts | mvrhel_laptop: Morning | 14:54.24 |
mvrhel_laptop | good morning Robin_Watts | 14:54.37 |
Robin_Watts | How jet lagged are you feeling? | 14:54.50 |
mvrhel_laptop | Just the early morning is a bit hard right now | 14:55.07 |
Robin_Watts | I get that all the time :) | 14:55.27 |
| How was the trip? | 14:55.48 |
mvrhel_laptop | Oh I think the trip went well. All the meetings were quite positive | 14:56.10 |
Robin_Watts | Excellent. | 14:56.18 |
mvrhel_laptop | need to help get kids off to school then I will be back | 14:57.47 |
Robin_Watts | Interesting... DEBUG_TRACE_PS_OPERATORS is defined in iinit.c | 14:58.19 |
| That might potentially slow the PS interpreter down a tiny amount. | 14:58.33 |
henrys | Robin_Watts: that's why it is conditional right? | 14:59.02 |
Robin_Watts | henrys: It's not conditionally defined. | 14:59.17 |
| It's just defined. | 14:59.19 |
henrys | Robin_Watts: funny I remember a conversation with igor where we agreed it should be explicitly used. | 15:02.22 |
kens | What does it actually do ? | 15:02.59 |
Robin_Watts | kens: It determines whether -z! works or not. | 15:03.13 |
kens | Oh.... | 15:03.25 |
Robin_Watts | pulls out git blame... i bet it turns out to be me. | 15:03.52 |
| oh, no. it was ray. on saturday. | 15:06.04 |
mvrhel_laptop | brb | 16:13.31 |
Robin_Watts | mvrhel_laptop: Are you available for a quick consultation now? | 16:17.07 |
mvrhel_laptop | Robin_Watts: yes | 16:17.26 |
Robin_Watts | So, I hit a problem in the psdcmykog device the other day. | 16:17.41 |
| The device sets it up so that num_std_colorants = 4. | 16:17.59 |
| and max_colorants = 6 | 16:18.16 |
| and we give 6 named colorants. | 16:18.24 |
mvrhel_laptop | ok sounds correct | 16:18.51 |
Robin_Watts | when running a PDF file, gs can see that there are only 4 colorants on a page, and so sets num_components to 4. | 16:19.04 |
mvrhel_laptop | oh. you need to specify a profile probably | 16:19.30 |
Robin_Watts | This was causing me to only output 4 planes. | 16:19.39 |
mvrhel_laptop | otherwise there is no way you are going to generate the other colorants | 16:19.43 |
Robin_Watts | I've patched around it in the device by writing empty planes if num_components < 6. | 16:20.10 |
mvrhel_laptop | unless those spots occur on that page | 16:20.15 |
Robin_Watts | But the worry I have is this: | 16:20.18 |
| suppose I run without a profile, and I run a file that contains a spot colorant that isn't "Artifex Orange" or "Artifex Green". | 16:21.14 |
mvrhel_laptop | it should use the alternate tint transform for those | 16:21.45 |
Robin_Watts | will gs set num_components to 5 ? Cos if so, I'll then output that colorants plate into the 'Orange' position. | 16:21.53 |
mvrhel_laptop | if the name is not on in the list of supported colorants then it should not up the count | 16:22.18 |
Robin_Watts | mvrhel_laptop: OK. Suppose I have a PDF file that has "Artifex Green" in it, but not "Artifex Orange". | 16:22.48 |
mvrhel_laptop | oh | 16:23.01 |
| that wont work | 16:23.04 |
Robin_Watts | Is gs smart enough to put the Green colorant in position 5, not position 4 ? | 16:23.09 |
mvrhel_laptop | hmm is your device smart enough is the question.... | 16:23.25 |
Robin_Watts | mvrhel_laptop: Almost certainly not. | 16:23.36 |
mvrhel_laptop | I suspect that this will not work | 16:23.42 |
| without a profile | 16:23.52 |
Robin_Watts | mvrhel_laptop: I performed some surgery on the psd devices and the devicen stuff while you were away. | 16:23.56 |
| It would be good if you could give the commits a quick once over to see if you approve or not. | 16:24.14 |
| I'd like to make the device behave properly without a profile, as well as with a profile. | 16:24.40 |
mvrhel_laptop | Robin_Watts: right. that should be straight forward. I think if we look at how the profile setting forces things to be 6 colorants it should be clear what we need to do | 16:30.08 |
| I have to review that real quick | 16:30.14 |
Robin_Watts | Thanks. | 16:30.25 |
mvrhel_laptop | Robin_Watts: are your commit changes already in the repository? | 16:30.47 |
Robin_Watts | mvrhel_laptop: Yes. | 16:30.54 |
| Let me get you some links. | 16:30.59 |
| http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=a1bee59ddf2bf82273359564d6cdb8e0549ae14f;hp=a65b8db506a553334de495a9b27a18811188539a | 16:32.09 |
| http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=7089d5416766a4a526b382e57e4436633d29bb9e;hp=a1bee59ddf2bf82273359564d6cdb8e0549ae14f | 16:32.28 |
| Those are the 2 main ones. You've looked over the psdcmykog device itself before. | 16:33.15 |
mvrhel_laptop | The psd stuff is defined in gdevdevnprn.h? | 16:35.39 |
Robin_Watts | mvrhel_laptop: Currently, yes. | 16:35.48 |
mvrhel_laptop | currently with your change? | 16:36.24 |
Robin_Watts | I did debate about pulling it out to another header file, but currently it relies on the device being a gdevdevnprn device to work. | 16:36.37 |
| yes, with my change. | 16:36.42 |
mvrhel_laptop | it just seems odd to have format specific stuff to be in the more general file gdevdevnprn.h | 16:37.57 |
Robin_Watts | yes, it is not ideal. | 16:38.10 |
| I should probably create a gdevpsd.h file and put it in there. | 16:38.27 |
mvrhel_laptop | Robin_Watts: and don't get me wrong. I am sure what you have done is a vast improvement over what was there | 16:38.46 |
Robin_Watts | no, you're right. It would definitely be better moved. | 16:39.10 |
mvrhel_laptop | ok so you opened up the psd specific functions to that they can be shared with the other devices. I don't like the hack in psd_write_header as I wish there was a way to do that in the device that was using it | 16:42.03 |
| i.e. the special stuff for Artifex Orange and Artifex Green | 16:42.44 |
Robin_Watts | mvrhel_laptop: The hack with the 'known' colorant names? | 16:42.46 |
| yeah. | 16:42.47 |
| I couldn't see an immediate way to do that better. | 16:43.17 |
| We could pass it in from the device to that function, but then either that function would be getting the colorant names from one places and the values from another... | 16:44.05 |
| or we'd have to pass both names and values in, and that's nasty, as the names will be in 2 places then. | 16:44.24 |
mvrhel_laptop | right. | 16:44.37 |
Robin_Watts | Can we 'preseed' the equiv colors with the values? | 16:45.03 |
mvrhel_laptop | probably | 16:45.26 |
Robin_Watts | or will that screw up the profile setting the values? | 16:45.31 |
mvrhel_laptop | Robin_Watts: the rest of the stuff looks good. let me review what i did when we do have an n-color icc output profile | 16:46.05 |
Robin_Watts | and at the point this function is called, we don't know how many planes are being used. | 16:46.11 |
mvrhel_laptop | so when we do have an n-color ICC profile, we end up taking the set up at line 371 in dgevpsd.c | 16:49.41 |
| gdevpsd.c I mean | 16:49.49 |
| However, by this time, some other code has already set us up with the names etc | 16:50.41 |
| let me find that | 16:50.43 |
| Robin_Watts: so gsicc_set_device_profile_colorants sets up the colorant names when there is a profile. I wonder if you can do something similar | 16:56.14 |
| actually you should not need to do all of this | 16:56.35 |
| it should be possible for you to just set up your device to have 6 colorants | 16:56.56 |
Robin_Watts | mvrhel_laptop: That would be ideal. | 16:57.23 |
mvrhel_laptop | of course, that may conflict with what I have set up for the ICC profile set up | 16:57.42 |
| Robin_Watts: if you want, I can see if I can get it working when you feel things are stable | 16:58.16 |
Robin_Watts | mvrhel_laptop: I feel things are stable :) | 16:58.26 |
| This is now being tested by marcosw. | 16:58.38 |
| as part of the overnights etc, I mean. | 16:58.55 |
| So, if you can see how to improve it, I'd be really grateful. | 16:59.22 |
mvrhel_laptop | Robin_Watts ok. in that case, just leave the color stuff with me if you want | 16:59.29 |
Robin_Watts | Consider it left :) | 16:59.41 |
| Thanks. | 16:59.48 |
mvrhel_laptop | ok. I will try to get it done before the meeting | 17:00.01 |
| Robin_Watts: when do you leave? | 17:00.07 |
Robin_Watts | If you have a profile that uses "Artifex Orange" and "Artifex Green" and maybe a test file, then you could commit those too into gs/examples/psdcmykog/ | 17:00.27 |
| and mention them in a comment in gdevpsdcmykog.c | 17:00.40 |
mvrhel_laptop | Robin_Watts: ok I will do that | 17:00.45 |
Robin_Watts | mvrhel_laptop: I fly on satuday. | 17:00.48 |
| or saturday. | 17:00.56 |
mvrhel_laptop | oh what day do you go to Maui? | 17:01.07 |
Robin_Watts | mvrhel_laptop: We get to Oahu on saturday at 10pm. | 17:01.26 |
| Then fly to Maui on the tuesday I think. | 17:01.35 |
mvrhel_laptop | ok | 17:01.40 |
Robin_Watts | Are you going via Oahu? | 17:01.51 |
mvrhel_laptop | yes. we are going this wed. | 17:01.59 |
| and then to maui on Monday | 17:02.05 |
| and then home on sunday | 17:02.14 |
Robin_Watts | right, we go home on the monday. | 17:02.27 |
mvrhel_laptop | ok that makes sense | 17:02.38 |
Robin_Watts | So, we could maybe meet up on sunday for a meal? | 17:02.50 |
mvrhel_laptop | that sounds like a good idea | 17:03.01 |
Robin_Watts | (in Oahu). | 17:03.05 |
| I will speak to the boss when she returns this evening :) | 17:03.29 |
mvrhel_laptop | yes. me too | 17:03.34 |
Robin_Watts | http://shorebirdwaikiki.com/ <- That's a nice place. | 17:04.35 |
| Where are you staying ? | 17:04.38 |
mvrhel_laptop | Robin_Watts: I see why you think its a nice place | 17:06.35 |
| ;) | 17:06.44 |
| We are staying at the Marriott | 17:07.11 |
kens | Goodnight all | 17:08.37 |
henrys | mvrhel_laptop, Robin_Watts wow thanksgiving vacation and then a week in Maui - does someone else want to lead the meeting ? ;-) | 17:09.29 |
| my brain should be empty when the meeting comes around. | 17:10.19 |
Robin_Watts | will be expecting henrys to be very laid back by the time the meeting rolls up. | 17:10.25 |
mvrhel_laptop | yes I was just thinking the same thing | 17:10.30 |
Robin_Watts | We don't do thanksgiving here, christmas is our equivalent. I have relations coming for xmas, so after every xmas holiday, I need a week in Maui to recover :) | 17:15.41 |
henrys | item 1 - surfs up | 17:16.04 |
| done | 17:16.14 |
Robin_Watts | item 1: Has the molasses cleared enough for the surf to be up :( | 17:16.25 |
| (Big molasses spill in Pearl Harbor in september. Killed lots of sealife. Worries were that it would also attract lots of sharks etc.) | 17:17.04 |
| Pearl Harbor = 5 miles from Waikiki. Sharks are bad. | 17:17.20 |
henrys | Robin_Watts: wow didn't hear about that. Miles will have a tail to grab | 17:17.56 |
Robin_Watts | hah! | 17:18.04 |
mvrhel_laptop | that is messed up. I had not heard about this either | 17:18.43 |
henrys | why sharks would be attracted to molasses is something I'll have to read about ... | 17:19.05 |
mvrhel_laptop | I think they are attracted to the dead fish | 17:21.36 |
Robin_Watts | molasses causes algae, algae kills sea life. Floating dead sealife = shark heaven. | 17:21.43 |
henrys | mvrhel_laptop: have you thought about windsurfing there? | 17:21.56 |
mvrhel_laptop | henrys: Maui has good windsurfing. I have gone there before. the trick is that Dec is big wave and not big wind time of year | 17:22.36 |
| Summer time is better for windsurfing | 17:22.49 |
henrys | mvrhel_laptop: yeah I do hope to watch those guys on the north shore, pretty amazing and we should be there right in the middle of it. | 17:23.33 |
Robin_Watts | has mailed the surf school I learnt with on waikiki. | 17:23.38 |
henrys | surfing that is. | 17:23.49 |
Robin_Watts | henrys: Are you staying on Oahu too then ? | 17:24.20 |
henrys | Robin_Watts: no just maui | 17:24.41 |
| I meant maui's north shore which is knows for surf - I don't think you are quite ready for that. | 17:25.21 |
Robin_Watts | henrys: right. oahu's north shore is hosting the World Surf Championship (or something like that) while we are there. | 17:26.27 |
| We are hoping to get up to see some of it. | 17:26.39 |
henrys | oh cool | 17:27.34 |
Robin_Watts | Morning ray. | 17:40.37 |
| In your commit on saturday you forced DEBUG_TRACE_PS_OPERATORS on in iinit.c | 17:41.16 |
| Was that deliberate? | 17:41.25 |
ray_laptop | Robin_Watts: OOPS. No, I was debugging to look into the problem I was having, and forgot that I had that in there. Then I was bit by git commit -a :-( | 17:42.13 |
Robin_Watts | ray_laptop: Been there, done that :) | 17:42.40 |
ray_laptop | Robin_Watts: I'll fix it right now. | 17:42.52 |
Robin_Watts | pure coincidence I was looking where to enable it myself :) | 17:43.15 |
ray_laptop | Robin_Watts: you're welcome ;-) | 17:43.36 |
| Robin_Watts: BTW, I had it defined in iinit.h since it is actually used in iinit.c _and_ interp.c | 17:50.04 |
Robin_Watts | yeah, I realised that. When you undefine it, leave it in as: /* #define DEBUG_TRACE_PS_OPERATORS */ | 17:50.53 |
| maybe with a comment. | 17:51.10 |
ray_laptop | I did | 17:51.16 |
Robin_Watts | cool :) | 17:52.04 |
ray_laptop | Robin_Watts: did we ever decide to make an limited account for cust 801 to access the code snapshots and git ? | 18:29.29 |
Robin_Watts | ray_laptop: No. | 18:29.47 |
ray_laptop | henrys: Robin_Watts: think I'll send them a code update. Do you think I should ask them (again) if they would like to be able to access the repository that has their device included ? | 18:31.45 |
Robin_Watts | ray_laptop: sending them a code update can only be a good thing. Mentioning it again seems reasonable too. | 18:32.16 |
ray_laptop | Robin_Watts: OK. To bring your private-repo up to date w.r.t. origin master, what is the 'correct' way ? I'm guessing "git push robin-private" but I don't want to mess it up | 19:12.47 |
Robin_Watts | git push robin-private master:master | 19:13.35 |
| at least, that's what I'd use. | 19:14.12 |
ray_laptop | Robin_Watts: OK. and that will update the xxxx branch as well ? | 19:14.19 |
Robin_Watts | no. | 19:14.23 |
| That will just push the local 'master' branch to be the remote 'master' branch. | 19:14.43 |
| you'd need to: git push robin-private xxxx:xxxx as well. | 19:14.58 |
ray_laptop | so, if I have a local xxxx branch, and I've pulled in the changes from master, I'd need to do both ? | 19:15.14 |
Robin_Watts | ray_laptop: Yes. | 19:15.25 |
| git push robin-private would probably work, but it might push all your other local tags too. | 19:15.48 |
| which would be bad if you had, say, customer 532 local stuff in a branch. | 19:16.06 |
| (and we share the repo with 801) | 19:16.14 |
ray_laptop | Robin_Watts: oops. I may have already inadvertently done that | 19:16.19 |
Robin_Watts | ray_laptop: Well, we should check for that before sharing :) | 19:16.35 |
ray_laptop | Robin_Watts: how would I check, and how would I clean out other junk branches ? | 19:17.06 |
Robin_Watts | ray_laptop: Urm... | 19:17.23 |
| git branch | 19:17.50 |
ray_laptop | Robin_Watts: that shows my local branches | 19:18.12 |
Robin_Watts | yeah, I'm just going to do that on casper. | 19:18.25 |
ray_laptop | Robin_Watts: can't I ask it what's on a remote repos ? | 19:18.39 |
Robin_Watts | git remote show robin-private | 19:19.38 |
| or: git ls-remote robin-private | 19:26.18 |
henrys | I wonder about getting them into git... | 19:27.24 |
Robin_Watts | http://www.youtube.com/watch?v=QMYfkOtYYlg | 19:30.16 |
maladmin | hi, i have 3 folders each with a collection of EPS images, i want a script to create a new image (pdf or eps) with folder1/image1 folder2/image1 folder3/image1 arranged in a 1x3 layout. Does anyone know what command can achieve this? | 19:31.10 |
ray_laptop | Robin_Watts: it looks like it has your branches, but doesn't have mine. That is good | 19:31.23 |
Robin_Watts | ray_laptop: It certainly doesn't have all of mine. | 19:31.55 |
ray_laptop | Robin_Watts: if I push xxxx:xxxx and you've changed their files, does it not undo your changes (how does it know which is newest) | 19:33.37 |
| Robin_Watts: well mooscript and ramfs_branch sure aren't mine (to name a few ) :-) | 19:34.47 |
Robin_Watts | ray_laptop: It will refuse the push unless it's a fast-forward. | 19:35.23 |
ray_laptop | Robin_Watts: OK. That's good. | 19:35.42 |
Robin_Watts | i.e. unless you're just adding commits onto the end of the history, it will refuse to take them. | 19:35.44 |
| Use -f to force it, but ideally we shouldn't need to do that. | 19:35.57 |
ray_laptop | Robin_Watts: OK. The git push robin-private master:master worked OK, and the git push robin-private xxxx:xxxx then told me "Everything up-to-date" | 19:38.34 |
Robin_Watts | fab. | 19:38.43 |
| I have to go fetch helen from the station in a mo. | 19:39.00 |
sebras | may I ask when your next irl meeting is? | 19:41.44 |
Robin_Watts | sebras: not this saturday, next saturday. | 19:42.09 |
ray_laptop | hmm... I can't seem to get into marcos' FTP site for cust 801 | 19:55.51 |
| and marcosw isn't here to ask about it. | 19:56.28 |
sebras | Robin_Watts: ok, thanks. | 19:56.31 |
ray_laptop | It connects OK, but it doesn't return the directory (LIST command) | 19:59.03 |
chrisl | ray_laptop: do you need to set it to passive transfers? | 19:59.48 |
ray_laptop | chrisl: it is set for passive transfers | 20:00.24 |
chrisl | Non-passive, then? | 20:00.41 |
ray_laptop | chrisl: non-passive fails immediately (passive waits for the 'network timeout' then gives the error) | 20:02.17 |
| chrisl: Failure is !Receive error Blocking call cancelled | 20:02.55 |
| I guess I can try from casper. It may be where I'm connected from (not my usual location) | 20:03.36 |
henrys | ray_laptop: just hangs for me. | 20:05.05 |
chrisl | Odd, that error does (according google) seem to related to passive transfer. Maybe an sms to marcos is in order | 20:05.50 |
ray_laptop | henrys: I guess I'll wait for marcosw to show up | 20:05.52 |
| I'll try calling him... | 20:06.08 |
chrisl | ray_laptop: if you want it out the way, you can always use your html_public directory on casper | 20:06.53 |
ray_laptop | chrisl: BTW, that error message was from passive mode | 20:07.07 |
chrisl | ray_laptop: yeh, I assumed that - but the two pages I looked at said to solve it, set passive on - weird | 20:07.48 |
ray_laptop | chrisl: this is has cust 801 private code in it, so putting it on a web site doesn't seem secure enough' | 20:08.54 |
| but maybe with a password required, that'll be good enough | 20:09.44 |
chrisl | ray_laptop: that's what I've done with private code in the past | 20:10.21 |
ray_laptop | particularly since marcos' FTP site is just a password (not sftp) | 20:10.22 |
| OK, I've uploaded the updates to cust801 to a web site on ~ray with username and password required (.htaccess method). Does anyone else need access for any reason ? (you can all just go directly to ~ray/public_html/... anyway) | 21:36.02 |
| and I've sent an email telling them about it, and asking if they want an account to access a private git repository | 21:36.54 |
| marcosw1: I didn't notice that you were here, but since I was having trouble with your FTP site, I just went ahead and created a private page for them to get updates from. They might still need to use the FTP site to upload stuff to us, I guess, so you might want to look into why it isn't working. | 21:45.13 |
Micha` | paulgardiner: Hi there! For the SMALL_FLOAT thing, thanks, it does make sense with > then :-) | 22:23.51 |
| Say, by the way, I've rewritten the set_markup_appearance for Highlights to mimick Adobe's, and it appears that adobe, poppler, and other all agree on the fact that the QuadPoints are not, as the standard specifies, given in counter-clockwise order. | 22:25.32 |
| More like in a God-why-oh-why order. From upper right, and in counter-clockwise, this would be 1st, 3rd, 4th then 2nd. | 22:26.28 |
| If it wasn't for the AP, the highlights produced by MuPDF would display wrongly in Adobe (and Poppler discarding the AP, it does appear wrongly in evince). | 22:27.23 |
| Forward 1 day (to 2013/11/26)>>> | |