| <<<Back 1 day (to 2014/04/08) | 2014/04/09 |
henrys | never mind the embollix ;-) | 00:07.10 |
| long day in the gl/2 minefield | 00:08.18 |
M_Odes | Hi | 03:36.45 |
| i would like to know two things | 03:37.36 |
| Related to de MuPDF viewer | 03:37.45 |
| It's anyone here that can answer my questions? | 03:38.11 |
ray_laptop | I was going to tell M_Odes the answer to his second question: "no" | 05:26.07 |
| but he didn't ask the first quesion, so "I don't know" | 05:26.34 |
| ;-) | 05:26.43 |
kens | chrisl am I correct in thinking that the only way I can cluster test a new file is by actually comitting the change ? | 08:46.06 |
Robin_Watts | kens: No. | 08:46.21 |
kens | Robin_Watts : is up and about early | 08:46.31 |
| OK Robin, so how ? | 08:46.37 |
Robin_Watts | As long as you have "git add"ed the file, you're file. | 08:46.39 |
kens | OK I haven't added the file, I'll do that, thanks | 08:46.49 |
Robin_Watts | too early to be able to type, it seems. | 08:46.50 |
| Something is unhappy with my machine. reboot time. | 08:51.32 |
kens | OK build seems to be good this time, thanks Robin | 08:54.24 |
chrisl | kens: do you use gitpush? | 08:55.53 |
kens | chrisl git gui mostly | 08:56.03 |
chrisl | for cluster runs | 08:56.10 |
kens | oh yes, I use gitpush yes | 08:56.26 |
chrisl | Right, clusterpush.pl doesn't require the file to be added | 08:56.48 |
kens | I'll need to do a manual commit this time since I've changed a makefile and don't want the tabs to be lost | 08:57.09 |
chrisl | "commit -n...." | 08:57.32 |
kens | Yeah, that's what I mean, I normally use git gui for that | 08:57.50 |
chrisl | I do think it's poor that git gui doesn't let you add options to operations like commit | 08:58.24 |
kens | Its less than ideal yes | 08:58.35 |
| Not normally a problem for me but it is in this case | 08:58.47 |
kens | doesn't often modify make files | 08:59.00 |
chrisl | doesn't often use git gui...... | 08:59.27 |
kens | Hmm xkcd is bang up to date this morning :-) | 09:24.38 |
jogux_mac | mm :) | 09:25.13 |
kens | is off out, back in a couple of hours | 09:26.51 |
paulgardiner | jogux_mac: can you tell me again your thoughts on what would be involved in having the iOS SO library support ARC? | 09:39.45 |
jogux_mac | I think I said less than a day. basically try to do the conversion, fix up whatever stops the conversion running, then try again. | 09:40.43 |
| I think if we start with the SOL sample project (which I think it what said customer cared about) that should be self contained. ish. maybe. | 09:41.14 |
| slightly worried there might be a snowball effect where we end up needing to convert all our iOS aliens | 09:41.29 |
Robin_Watts | paulgardiner: I thought jogux had said that a rebuild of the lib should pretty much get us it for free. | 09:42.10 |
| My impression (which may be wrong) was that the automatic reference counting is something that happens in the wrapper layer. | 09:42.47 |
jogux_mac | I think there is an issue in the alien where there's one or two bits of the obj c <-> c interface that aren't ARC safe and require attention. | 09:42.47 |
| paulgardiner: as we're on the subject :-) how about you review all the iOS stuff in my private repo? ;) | 09:43.09 |
paulgardiner | jogux_mac: sure. Will do | 09:43.30 |
jogux_mac | paulgardiner: also, you saw the BT keyboard stuff to get back to that customer? | 09:43.31 |
| or at least you will do when you do the reviews :-) | 09:43.39 |
paulgardiner | jogux_mac: yeah, just notice that in the logs | 09:44.00 |
| jogux_mac: good to have that done, thanks. | 09:44.43 |
jogux_mac | paulgardiner : I'm rather annoyed at the autorepeat not working. :( hoping Apple might come back to me with an idea for an easy fix... | 09:45.18 |
paulgardiner | jogux_mac: yes. That' a pain, but I think the customer that complained will be very pleased with what has been achieved. | 09:46.56 |
jogux_mac | we might want to give them a testflight build | 09:47.54 |
| s/them/him/ | 09:54.31 |
tor8 | paulgardiner: there's some oddities going on with the mupdf windows viewer and the save dialog on exit | 10:20.30 |
| (check the log yesterday) | 10:20.35 |
paulgardiner | tor8: yes. I was that in the logs and was just waiting for you to log in before replying | 10:32.21 |
| I've seen it before, and my assumption (pretty sure) is that it is a widget being updated by some JS that runs on opening the document. | 10:33.25 |
| I'm not sure how we are supposed to distinguish between widget updates that need saving and ones that don't, but probably an acceptable fix would be to reset the "changed" flag after running the "start up" JS. | 10:35.55 |
| .. although there are related problems that that wouldn't solve, such as calc.pdf requesting save when exited. | 10:36.44 |
tor8 | paulgardiner: yeah. maybe flag if a change was initiated with a user event? | 10:38.03 |
| but I think that's too big of a change for the release | 10:38.10 |
paulgardiner | initiated by a user event would still mean calc.pdf requesting saving | 10:39.02 |
jogux_mac | wonders if there's a way to stop the twiki making me login every day, or whatever it's doing that means I keep on having to relogin :-) | 10:42.14 |
tor8 | paulgardiner: maybe check the NoExport flag for the widgets? | 10:44.51 |
| in calc.pdf the display widget has /F 4 | 10:45.02 |
paulgardiner | tor8: ah that might do it. I thought that flag was concerned with exporting forms but it might relate to saving too. | 10:46.23 |
tor8 | I reckon it ought to apply to both | 10:46.38 |
| brb, need to relog | 10:46.58 |
paulgardiner | tor8: resetting the "changed" flag after running the initialising js would be an easy change and would fix the main annoyance. | 10:49.11 |
tor8 | paulgardiner: kens: I can't find the test file kens mentioned yesterday | 10:49.35 |
| paulgardiner: I don't get a save dialog on x11, so I guess that's a windows specific feature? | 10:50.11 |
paulgardiner | windows android ios | 10:50.25 |
tor8 | I'm hoping to do something about that before next release... bring the x11/windows viewers into parity and make them both usable. | 10:51.43 |
| but gui programming has this tendency to make me procrastinate... | 10:52.17 |
| paulgardiner: is there any reason that doc->dirty = 1 is not done inside pdf_dirty_obj? | 11:09.29 |
| paulgardiner: and I can't see where toggling a checkbox would trigger a save | 11:12.06 |
jogux_mac | robin_watts : I can't find the mips toolchain on any of the VMs I've checked so far :( | 11:14.34 |
paulgardiner | tor8: I think pdf_dirty_obj has a different purpose. It's called to record that the appearance stream is out of date with the value... I believe. | 11:21.36 |
tor8 | not that the numbered object needs to be saved in an incremental update? | 11:30.28 |
paulgardiner | tor8: Not as far as I can tell/remember. I test the dirty flag in pdf_update_appearance. | 11:35.53 |
| objects that need saving are those in the last xref array | 11:36.38 |
| tor8: I think you may be correct that checkbox toggles aren't setting the document-dirty flag | 11:37.34 |
Robin_Watts | jogux_mac: OK. I can try to set up that toolchain on peeved. | 11:40.29 |
| Or I could try to set it up on one of the ATS VMs. | 11:40.42 |
| Which would be preferable? | 11:40.52 |
jogux_mac | I think just do what you need now | 11:40.58 |
Robin_Watts | k. | 11:41.05 |
jogux_mac | setting up ATS is proving frustrating, with all the data being in different places and all being remote. I'm pondering if I should just ask someone to buy a 4T USB drive, stick the entire NAS contents on it and ship it to me, then hopefully setup is fast and then we can sort out the permanent install at a more leisurely pace. | 11:42.38 |
| robin_watts : I had a quick check through the wiki and couldn't find anything obvious about that toolchain or customer. | 11:43.00 |
Robin_Watts | or we could ship a jogu to the NAS :) | 11:43.35 |
| The customer wants: mips-gcc-4.3.3-sourcery_g++_lite | 11:45.20 |
| but I can't see exactly that version on the toolchain download options. Di I go for 4.3 or 4.4, do you reckon? | 11:45.42 |
jogux_mac | is there any chance 4.3 is 4.3.3 | 11:47.39 |
Robin_Watts | could be. | 11:47.54 |
| I'd better get 4.3 as it probably has to match the customer. | 11:48.07 |
jogux_mac | there might possibly be an archive on dataserver or similar I guess. | 12:00.58 |
kens | chrisl ping | 12:10.54 |
chrisl | kens: pong | 12:11.15 |
kens | Can ytou try the 9.14 release with the bug from #695146 please ? | 12:11.40 |
| On Linux | 12:11.43 |
| When I download the binary and try it it tells me it cannot execute the binary | 12:11.58 |
Robin_Watts | jogux_mac: Can't see it. Ass. | 12:12.29 |
chrisl | kens: did you untar it on Windows? | 12:12.45 |
kens | chrisl no I downloaded it on LInux and used Fedora's archive menager | 12:13.03 |
| My suspicion is that its a broken package in some way, possibly system shared libraries, but I'd like to be sure that our binaries are OK with it. | 12:13.55 |
| Hmm, I suppose this might be a 32-bit system | 12:14.24 |
chrisl | It would give an error saying missing libs if that was the case | 12:14.27 |
kens | (my VM that is) | 12:14.28 |
| chrisl Not missing ones, but not the ones we ship ;-) | 12:14.47 |
| It works fine for me on Windows | 12:15.02 |
chrisl | The binaries are built with everything we ship | 12:15.11 |
kens | I'm building a debug binary right now on Linux | 12:15.13 |
| chrisl that's what I mean. I suspect the reporter has used a package form 'somewhere else' that uses system libs. | 12:15.38 |
| I just want to be sure that his file works with *our* binary | 12:15.48 |
chrisl | The pre-built binary has the gray/brown backdrop :-( | 12:16.57 |
kens | THe binary I ubilt myself works just fine but of course that's *current* code, not the 9.14 release (not that there's a massive difference) | 12:16.58 |
| Oh dear..... | 12:17.06 |
chrisl | Using his command line | 12:17.29 |
kens | I didn't use his whole command line on Linux (I did on Windows) | 12:17.49 |
chrisl | Hmm, the 32 bit one is complete garbage..... | 12:18.15 |
kens | O.O | 12:18.22 |
| THe Windows ones are fine, using his command line | 12:18.34 |
chrisl | I'll do a proper build of the 9.14 release, and see | 12:19.05 |
kens | I'm just trying the current code with his command line | 12:19.31 |
| Current code worked OK for me | 12:20.12 |
| His command line with a couple of minor path adjustments | 12:20.25 |
chrisl | Built on my current system, the 9.14 release works just fone | 12:20.50 |
| fine, even | 12:20.53 |
kens | OK well its not just me then, at least we agree on that | 12:21.07 |
| I can't see anything blindingly obvious in the log | 12:21.17 |
| Oh except LCMS 2? | 12:21.28 |
chrisl | I tried 9.14 | 12:21.42 |
kens | boggles | 12:21.52 |
chrisl | But, he says he's using Windows 7, but puts a Unix command line.... doesn't list 32/64 bit..... | 12:22.51 |
kens | Yeah I know | 12:22.57 |
| I commented on that in my reply | 12:23.04 |
chrisl | I wouldn't be surprised if he is actually using the Linux binary from the website - this is why I was so against Linux binaries :-( | 12:24.31 |
kens | I think you may be correct. Can I leave this one with you ? I don't think there's anything I can do about it | 12:24.49 |
chrisl | I doubt there's anything I can do about it.... but sure, give it to me. | 12:25.29 |
kens | OK, I'm just downloading the 32-bit binary to see if that works for me :-) | 12:25.56 |
| I changed it to 'build process' which probably isn't strictly correct but assigns it to you | 12:26.41 |
| Ah, the 32-bit binary does work, I wonder if I opened the wrong VM.... | 12:27.57 |
chrisl | Work? You mean the output is correct? | 12:28.19 |
kens | No work as in the binary actually launches and gives me a GS prompt :-) | 12:29.09 |
| Which the 64-bit one didin't | 12:29.18 |
| You're correct that the output is garbage | 12:29.31 |
| Really, really strange garbage | 12:29.40 |
| It looks like some kind of jpeg/JPX decoding error | 12:30.24 |
chrisl | Yeh, but if the glibc ABI has changed, strange behaviour is to be expected | 12:30.32 |
kens | for the MuPDF developers, there's a couple of questions on Stack Overflow, this one looks like the developer might be competent enought to help: | 12:35.47 |
| http://stackoverflow.com/questions/22959846/writing-an-air-native-extension-using-mupdf-render-engine | 12:35.48 |
Robin_Watts | drafts something, thanks. | 12:36.42 |
chrisl | Hmm, well, something is awry, because the 32 bit exe produced garbage on the system I built it on, so :-( | 12:38.01 |
kens | My VM claims to be 64-bit, runs the 32-bit binary ut not the 64-bit binary, so I'm kind of puzzled as to what I actually build when I compile from source..... | 12:39.16 |
chrisl | at a terminal, do uname -a | 12:39.40 |
kens | [ken@Zen-Fedora14-64 ghostscript-9.14-linux-x86]$ uname -a | 12:40.20 |
| Linux Zen-Fedora14-64 2.6.35.14-106.fc14.i686 #1 SMP Wed Nov 23 13:57:33 UTC 2011 i686 i686 i386 GNU/Linux | 12:40.20 |
chrisl | So, that's an x86 install "i686 i686 i386" | 12:40.48 |
| Mine says: " x86_64 x86_64 x86_64" | 12:41.03 |
kens | Ah, I wonder why its not 64-bit.... Oh well | 12:41.04 |
| IN that case the 32-bit code works when I build it locally | 12:41.18 |
| But not the 32-bit binary from the download | 12:41.28 |
chrisl | I've suggested he try building it himself - cue the complaints..... | 12:44.36 |
kens | :-D | 12:44.44 |
kens | needs coffee | 12:45.51 |
paulgardiner | jogux_mac: commits look good to me, alt.hough I have to admit failure to check very deeply that the additional autorelease and release calls in the last one are okay. | 12:53.17 |
jogux_mac | if they weren't we'll find out soon enough ;-S | 12:53.43 |
paulgardiner | true | 12:54.31 |
chrisl | kens: Hmm, so on Ubuntu 7.04 32 bit, gs 9.14 produces garbage.... | 12:55.17 |
kens | bult from source ? | 12:55.31 |
chrisl | Yes | 12:55.36 |
kens | That's odd | 12:55.42 |
chrisl | That's what I use for the binaries. | 12:55.52 |
kens | Mine definitely works OK with that file | 12:55.53 |
| Of course, I'm using an old gcc | 12:56.27 |
chrisl | Well, that was 4.1.2 which is pretty old | 12:56.48 |
kens | Mine is 4.5.1 | 12:57.14 |
chrisl | I have to run an errand - I'll mess with this a bit more when I get back.... | 12:58.25 |
kens | OK | 12:58.29 |
kens | is cursing encryption | 12:58.41 |
henrys | paulgardiner: welcome back | 13:20.28 |
paulgardiner | henrys: thanks. Good to be back. Trying to catch up with all that's been going on in the last week | 13:22.46 |
jogux_mac | henrys : did you see my earlier comments about ATS/etc? Tempted to ask that the whole NAS is stuck on a 4G usb drive and shipped to me. Having all the data in different locations is just making figuring out how to fit the pieces together trickier/slower than ideal :( | 13:34.49 |
kens | another backup is always good to have | 13:43.36 |
henrys | jogux_mac: 4G? | 13:47.56 |
Robin_Watts_ | henrys: 4TB :) | 13:48.40 |
jogux_mac | henrys: errr. oops. 4TB :-) | 13:48.42 |
| the NAS has about 2.8T of stuff on it I believe. | 13:49.06 |
Robin_Watts | Something like: http://www.ebuyer.com/339426-hgst-4tb-touro-dx3-desktop-hard-drive-0s03400?utm_source=google&utm_medium=products&gclid=CO3pw-XF070CFXMRtAodSX8AmA | 13:49.08 |
jogux_mac | precisely. | 13:49.19 |
Robin_Watts | I shelled one of those for the 4TB drive in my machine here. | 13:49.31 |
henrys | ah 4TB now it makes sense | 13:50.05 |
Robin_Watts | jogux_mac: So I've got the tools installed, and I've checked out to the tag, and I've done the build... | 13:50.34 |
spanners | just use lzma on max settings | 13:50.37 |
Robin_Watts | and after it says "Linux TGV library was successfully built!" | 13:50.53 |
| it then says: Copying fileviewer + test files to /tmp/robin-picselapp | 13:51.11 |
| and then: cp: cannot stat `/home/robin/sauce/sot.git/epage/platform/tgv/apps/linux-alien/fileviewer': No such file or directory | 13:51.23 |
jogux_mac | that doesn't sound good. :( | 13:51.38 |
henrys | jogux_mac: that's fine, is there any way though we could slim this beast down, surely some of it must be obsolete or unused. | 13:51.57 |
jogux_mac | (my. obvious) | 13:52.02 |
| henrys : yeah; there definitely is ways to slim it all down; it's just easier to figure that all out once we have it up and running | 13:52.33 |
Robin_Watts | indeed. My epage/platform/tgv/apps/linux-alien has no fileviewer dir. Is this something that got lost in the conversion? Have you or pedro or paulgardiner seen this error before? | 13:52.42 |
jogux_mac | henrys : we're missing the NIS master at the moment - whether it's on one of the other VMs, or got missed, not yet sure, and it's difficult to poke around with the data to quickly check. | 13:53.18 |
Robin_Watts | henrys: Lots of it will be wasted space (cos lots of it is VMs), but it's trying to establish what's wasted etc with the data being remote that's hard. | 13:53.32 |
henrys | jogux_mac: marcos is out of town though not sure when he returns and we don't have anyone else in the bay area that can do it immediately. | 13:54.12 |
jogux_mac | henrys : ah. darn. | 13:54.27 |
| in theory marcos already copied it all onto a usb drive, but where that drive is I'm unsure. | 13:54.48 |
| robin_watts : it's not there on master either. | 13:55.05 |
Robin_Watts | yeah, my first build was on master :) | 13:55.20 |
jogux_mac | ah | 13:56.20 |
| I think it /might/ be the name the postlink is expecting the generated binary to have. | 13:56.32 |
ray_laptop | when does marcosw return ? | 13:57.31 |
jogux_mac | robin_watts : what's in platform/tgv/apps/linux-alien/target-info.mk (that's generated by the build) | 13:57.42 |
| robin_watts : actually may better to take this to skype. | 13:57.57 |
henrys | jogux_mac: anyway I'll send email to marcos and ask him about it. | 13:58.08 |
ray_laptop | henrys: do you know when marcosw gets back? I want to make sure that he got the drive I sent to him | 13:59.03 |
jogux_mac | henrys : there's plenty of other work to do still, so it's perhaps not urgent. | 13:59.14 |
ray_laptop | I don't see an email from him that says | 13:59.42 |
henrys | ray_laptop: no but I'll ask in the email | 13:59.43 |
ray_laptop | plowing though 2700 files of differences between fast 4-bit thresholding and gs' normal halftoning is NOT my idea of fun. But after 300 files, I did find one I need to look into. | 14:22.17 |
jogux_mac | ray_laptop : crikey. well done for staying concious till number 300 :-) | 14:39.07 |
chrisl | kens: I think the problem with that file is down to openjpeg - there seems to be something in openjpeg 2.x that trips up older gcc versions on x86 | 15:11.55 |
kens | Oh, I guess there's not a lto we can do about that | 15:12.13 |
| except tell teh reporter to upgrade his gcc | 15:12.44 |
chrisl | I wouldn't want to do that without testing *just* an updated gcc, and that will take quite some time | 15:13.14 |
kens | I guess we do have to do it though, we kind of need to know what hte problem is, at least in outline | 15:13.40 |
chrisl | I'm somewhat disinclined to try to track down a bug in an obsolete compiler version :-( | 15:14.29 |
jogux_mac | if it's the optimiser, building without optimisation may help narrow it down? | 15:14.40 |
kens | wchrisl, of course! I merely meant we need to be reasonably sure tha'ts the problem | 15:14.57 |
chrisl | jogux_mac: Possibly, yes. Shame I'm having to use a VM, and builds take a while :-( | 15:15.27 |
jogux_mac | :( | 15:15.33 |
henrys | jogux_mac: he'll ship it next week. | 15:19.29 |
| ray_laptop: he got the drive he'll send you email when it's transferred | 15:19.53 |
jogux_mac | henrys : awesome, thanks. | 15:20.27 |
ray_laptop | henrys: OK. Thanks. I don't need the drive back, BTW. Unless he's just going to throw it away | 15:23.17 |
| the drive cost (Miles) $70, and FedEx was $30. | 15:23.56 |
chrisl | kens: a debug build shows the same problem, so it's not the optimiser :-( | 15:52.02 |
jogux_mac | :( | 15:59.25 |
Robin_Watts | ray_laptop: How big was the drive? :) | 15:59.58 |
ray_laptop | Robin_Watts: just 1Tb | 16:00.47 |
| Robin_Watts: about .4 inches thick, and about 3" x 4" | 16:01.21 |
Robin_Watts | boom tish! | 16:01.43 |
ray_laptop | Robin_Watts: took 45 hours to xfer the 600Gb of junk to it | 16:02.03 |
Robin_Watts | I was just wondering if it was big enough to be repurposed for sending data to jogu. | 16:02.09 |
ray_laptop | Robin_Watts: well, it's small and lightweight, albeit a bit slow. | 16:03.05 |
| It belongs to Artifex (or will when I put in my expense report for April) so whatever it's good for is fine | 16:03.52 |
Robin_Watts | ray_laptop: I think jogu needs ~3 Gig. | 16:05.00 |
chrisl | Robin_Watts: 3 Gig? Isn't that doable over the net? | 16:06.34 |
Robin_Watts | 3TB. | 16:06.52 |
| Sorry. | 16:06.55 |
chrisl | Ah, I did wonder..... | 16:07.04 |
| And not compressable? | 16:07.16 |
Robin_Watts | Anything bigger than a floppy is too large for me to comprehend. | 16:07.25 |
chrisl | trying not to crack bad jokes about "large floppies"...... | 16:07.56 |
ray_laptop | I didn't think to try compressing the testdocs -- I just did a copy, but afterwards I started a tar --xz and after 30 hours it was up to 400Gb compressed so I killed it | 16:09.10 |
| I don't know if the stufff for jogux is more compressible | 16:09.46 |
Robin_Watts | chrisl: It's 2.8TB of VMs etc. I suspect that it's not worth the time/effort to compress them. It'll just make them harder to work with when they arrive. | 16:09.58 |
jogux_mac | some of the VMs are already compressed disk images, some not :-S | 16:10.19 |
ray_laptop | the ones that aren't will probably compress well | 16:10.36 |
jogux_mac | the ones that are compressed are a pain as I couldn't figure out how to access them without starting the VM, though I may have missed something. :( | 16:10.52 |
ray_laptop | test 711/2692 also has a problem. 3 so far :-( | 16:18.42 |
kens | Goodnigth all | 16:34.53 |
jackiechan | Hi Is there a way to include mupdf headers in c++ headers? I tried using extern "c" but im always getting "_pdf_obj_read_state_s already defined in ..." | 16:39.43 |
Robin_Watts | jackiechan: The extern "C" thing sounds right to me. | 16:44.10 |
jackiechan | But its not compiling im getting this error 1>testApp.obj : error LNK2005: _pdf_obj_read_state_s already defined in main.obj | 16:44.57 |
Robin_Watts | Then you must be doing something wrong. | 16:45.31 |
| Like using C++ for a start :) | 16:45.50 |
jackiechan | Like using C++ for a start what is this mean? | 16:46.16 |
| thanks for helping but i dont understand this one | 16:46.27 |
Robin_Watts | extern "C" { #include "whatever mupdf header.h" } | 16:46.34 |
| It was a flippant comment; using C++ at all is a mistake IMHO :) | 16:47.14 |
jogux_mac | I'm confused as to why you'd get a linker error for what appears to be a typedef/struct. | 16:47.25 |
jackiechan | i dont have choice for this one i need to stick with c++ | 16:47.42 |
| if it helps this is testApp.h #pragma once #include "ofMain.h" extern "C" { #include <mupdf/fitz.h> #include <mupdf/pdf.h> } | 16:47.53 |
jogux_mac | do you have any other header files that #include mupdf header files? | 16:48.23 |
jackiechan | no | 16:48.38 |
| i dont have | 16:48.41 |
jogux_mac | I believe those should be "", not <>s - ie. #include "mupdf/fitz.h". | 16:49.06 |
| I don't know if that is significant. | 16:49.10 |
jackiechan | ok i will change it | 16:49.35 |
| just a sec | 16:49.38 |
jogux_mac | <> should only be used for #include files provided by system libraries. | 16:49.41 |
jackiechan | i used the example code provided by the mupdf website http://www.mupdf.com/docs/example.c | 16:51.20 |
| i changed it.. im still getting the same error | 16:51.41 |
Robin_Watts | jackiechan: Then we'd need to see more code I guess. | 16:52.13 |
jackiechan | https://gist.github.com/anonymous/10291372 | 16:53.02 |
| ok here you can see all the files i have in a single gist | 16:53.27 |
jogux_mac | what's in main.c / main.cpp ? | 16:54.45 |
jackiechan | ok let me clear that a bit | 16:55.28 |
| ok sorry about that | 16:56.37 |
| https://gist.github.com/anonymous/10291703 | 16:56.38 |
jogux_mac | what's in ofmain.h ? | 16:57.24 |
jackiechan | thats a library header | 16:57.49 |
| for OpenFrameworks | 16:57.56 |
| https://github.com/openframeworks/openFrameworks/blob/master/libs/openFrameworks/ofMain.h | 16:58.19 |
Robin_Watts | And the exact error messages you are getting please? (In a pastebin.com window?) | 16:58.40 |
jogux_mac | can you try removing the #pragma once ? | 16:59.18 |
jackiechan | Error8error LNK2005: _pdf_obj_read_state_s already defined in main.obj testApp.objemptyExample | 16:59.25 |
| sure i can remove #pragma | 16:59.44 |
| after i removed that pragma in testApp.h i got | 17:00.33 |
| Error11error C2371: 'fz_context' : redefinition; different basic typesc:\users\ali\desktop\of_v0.8.0_vs_release\addons\ofxmupdf\lib\include\mupdf\fitz\context.h201emptyExample | 17:00.34 |
| oh sorry thats had nothing to do with this | 17:00.52 |
| i got the same error : | 17:01.26 |
| Error2error LNK2005: _pdf_obj_read_state_s already defined in main.objC:\Users\Ali\Desktop\of_v0.8.0_vs_release\apps\myApps\ofxMuPDFExp\testApp.objemptyExample | 17:01.27 |
Robin_Watts | rather than using #pragma once, use: #ifndef TESTAPP_H\n#define TESTAPP_H\n ... current file contents ... #endif | 17:02.06 |
jackiechan | ok let me try | 17:02.21 |
jogux_mac | jackiechan: tell us more about that redefinition error - is that from the same project? | 17:02.40 |
Robin_Watts | and testapp.h includes ofMain.h, so you don't need to #include testApp.h in main.cpp | 17:02.44 |
jackiechan | yes its from the same project | 17:04.37 |
jogux_mac | robin_watts : other way around - don't need to #include ofmain.h I think? | 17:04.57 |
Robin_Watts | Sorry, yes. | 17:05.06 |
jackiechan | Its defaults file coming with all projects in openframeworks | 17:05.07 |
jogux_mac | jackiechan: tell us more about ofxmupdf | 17:05.07 |
| that appears to be a second copy of mupdf | 17:05.16 |
Robin_Watts | jackiechan: Please paste the entire output from the make process into a pastebin.com window. | 17:05.28 |
jackiechan | ofxmupdf : yes in the mean time i tried something while we have this discussion but it wasn't helpful. | 17:06.12 |
| ok sure | 17:06.26 |
| ok regarding to ofMain.h header its a standard setup in openframeworks that they have by default. You can check empty project here if you want https://github.com/openframeworks/openFrameworks/tree/master/examples/empty/emptyExample/src | 17:08.36 |
| btw im recompiling the app to get all compilation output (including openframeworks it self, its attached as dependent project) | 17:09.33 |
| output : https://gist.github.com/anonymous/10293034 | 17:10.24 |
Robin_Watts | I am confused in that main.cpp doesn't ever mention pdf_obj_read_state_s | 17:11.54 |
jackiechan | yes exactly | 17:12.14 |
| it should be somewhere in the mupdf headers | 17:12.30 |
| if i cut the extern "c" mupdf includes part and take it to testApp.cpp everything works fine | 17:13.29 |
| but that makes hard to design my app sine now i cant use any mupdf types in my c++ header, plus this should be working and i want to know whats happening here. | 17:14.31 |
jogux_mac | I do not understand it :( | 17:15.10 |
jackiechan | sorry which one? the last one? | 17:15.29 |
| or the whole thing? | 17:15.48 |
jogux_mac | I mean, I don't understand why you are getting this error. | 17:15.52 |
jackiechan | oh sorry | 17:15.58 |
Robin_Watts | jackiechan: As a workaround... Can you use a .c file to call mupdf? Then have your c++ files call that .c file ? | 17:16.41 |
jackiechan | himm i dont know how that works i can try i think | 17:17.13 |
Robin_Watts | Did you say something about their already being an ofxMuPDF? | 17:17.15 |
jackiechan | yes in the mean time i tried something while we have this discussion but it wasn't helpful. | 17:17.30 |
jogux_mac | ah. I now understand why it's a linker error, I think; link time code generation is enabled. | 17:17.36 |
jackiechan | (ofxmupdf) | 17:17.46 |
| oh what is that | 17:18.08 |
jogux_mac | it means it does some of the compilation in the linker. It's not a problem. | 17:18.42 |
ray_laptop | jogux: other than obfuscating the cause of the error ;-) | 17:19.10 |
jogux_mac | ray_laptop : true :-) | 17:19.30 |
jackiechan | i checked it value and for the both projects its "No Whole Program Optimization" | 17:20.34 |
| i dont know if you find it useful but i can even send you the project zip it self | 17:21.11 |
| with the lib and headers for mupdf | 17:21.26 |
| and if you want you can quickly download openframeworks (80mb) | 17:21.49 |
| to compile try compiling it | 17:22.03 |
ray_laptop | jackiechan: that seems reduntantly redundant ;-) | 17:23.29 |
jogux_mac | jackiechan : yeah, link time code generation is being forced on because the libmupdf you have was built with it enabled. | 17:26.11 |
jackiechan | ok should i disable it? | 17:26.41 |
| and perhaps rebuild the library? | 17:27.02 |
jogux_mac | I don't know if that would help or not. But I am stumped. | 17:27.34 |
jackiechan | ok its gonna take couple minutes i will let you know | 17:28.08 |
jogux_mac | is it correct to say that you are using ofxmupdf, and that is your only copy of mupdf? | 17:28.35 |
jackiechan | jogux no i removed that from the project i just tried another implementation | 17:32.58 |
jogux_mac | the build log mentions "c:\users\ali\desktop\of_v0.8.0_vs_release\addons\ofxmupdf\lib\include\mupdf\fitz\system.h(85)" which suggestions it may not have been fully removed :-( | 17:33.54 |
jackiechan | yes thats the folder im including mupdf nothing other then that | 17:34.45 |
| i will remove that one as well | 17:35.17 |
jogux_mac | so you are using the mupdf that comes from ofxmupdf, but you are not using ofxmupdf itself? | 17:35.41 |
| ah, if you're removing that then that's fine. | 17:36.02 |
jackiechan | ofxmudf i just a folder sorry for confusion, its a seperate folder i created to put sources related to mupdf | 17:39.04 |
| it* | 17:39.24 |
jogux_mac | okay, I remain unsure as to what the problem is then :-( | 17:39.59 |
jackiechan | most recent build output | 17:41.38 |
| https://gist.github.com/anonymous/10295845 | 17:41.40 |
Robin_Watts | jackiechan: AIUI, you are including the mupdf headers both from testApp.cpp and from main.cpp ? | 17:43.18 |
jackiechan | no | 17:43.36 |
Robin_Watts | both testapp.cpp and main.cpp #include "testApp.h" right? | 17:44.00 |
jackiechan | yes | 17:44.07 |
Robin_Watts | Right, so both testapp.cpp and main.cpp are both including the mupdf headers. | 17:44.23 |
| So, move the extern "C" { #include "" ... } from testapp.h | 17:44.48 |
| and put it in testapp.cpp | 17:44.53 |
jackiechan | but i got #pragma once, or #ifndef stuff to prevent that? | 17:44.58 |
Robin_Watts | That way it's only included from 1 place. | 17:45.03 |
jackiechan | ok if i understand you correct | 17:45.35 |
Robin_Watts | jackiechan: The #pragma once stuff says that it's only included once per .cpp file compiled. | 17:45.36 |
| You have 2 .cpp files ,and so it's being included twice. | 17:45.48 |
jackiechan | so is that means i cannot put these c headers into my testApp.h? | 17:46.18 |
| and the only way to use them is to include them in cpp? | 17:46.46 |
Robin_Watts | jackiechan: I'm saying try it and see. | 17:47.12 |
jackiechan | yes i know that is working | 17:49.02 |
Robin_Watts | oh, well, you might have told *us* ! | 17:49.17 |
jackiechan | as i said its working when i include them in testApp.cpp | 17:49.20 |
| yes i told it already | 17:49.27 |
| i think you missed it :) | 17:49.37 |
Robin_Watts | I strongly suspect that the problem is your compiler having a bug to do with link time code generation. | 17:49.44 |
jackiechan | Himm maybe i need to update it or try it some where else.. maybe on gcc? | 17:50.24 |
Robin_Watts | Actually... maybe it's not even link time code generation that's at fault. | 17:50.42 |
jackiechan | i really appreciate all hard work thank you very much. if any solution comes into your mind i would love to hear it. please use nakipogluali at gmail dot com | 17:53.05 |
Robin_Watts | So I would suggest that you write an interface class. | 17:53.16 |
| And keep all the mupdf operations within that single file. | 17:53.33 |
jackiechan | yes thats an option and then an impl class | 17:53.37 |
| thank you all very much! | 17:54.28 |
Robin_Watts | np. | 17:54.39 |
jogux_mac | I still don't understand that at all. :( | 17:57.40 |
Robin_Watts | jogux_mac: I think, for sanities sake, I should switch to building this on a 32bit linux box. | 17:58.41 |
| or a 32bit linux VM at least. | 17:58.46 |
jogux_mac | hehe, that may be a good idea :-) | 17:58.56 |
Robin_Watts | Is there an ATS VM running I can get to? | 17:59.12 |
jogux_mac | essentially. I'd need to create you an account. | 17:59.22 |
Robin_Watts | And do they have sufficient disc space? | 17:59.33 |
| (git repo + working set + generated dir) | 17:59.57 |
jogux_mac | I really hope so, if not that is another flaw in my continually failing cunning plan. | 17:59.59 |
| I'll copy your ssh keys from points | 18:01.29 |
Robin_Watts | Ta. | 18:01.34 |
jogux_mac | it's very laggy, not sure why | 18:03.09 |
| robin_watts : done. you want fermis; see http://twiki.ghostscript.com/do/view/GhostDocs/AutomatedTest | 18:04.15 |
| once on fermis, I guess try deb32-etch01 as root, with the password on there | 18:04.29 |
| or maybe debian32ats01. not sure which may be best. | 18:04.42 |
| sshing in as root has long delays due to ypbind/nis timeouts; be patient | 18:05.09 |
Robin_Watts | jogux_mac: I don't appear to be able to get to fermis. | 18:08.08 |
| Not as robin, anyway. | 18:08.14 |
jogux_mac | you're trying the right port? | 18:08.22 |
Robin_Watts | Yes. It asks for a password for robin. | 18:08.53 |
| do I actually have an account on fermis ? | 18:09.07 |
jogux_mac | yes | 18:09.22 |
| do you know what ssh host key did you get asked about? | 18:09.32 |
Robin_Watts | wierd. | 18:10.00 |
jogux_mac | it says it worked :-) | 18:10.04 |
Robin_Watts | Now it let me in. | 18:10.04 |
jogux_mac | I didn't change anything, honest. | 18:10.11 |
Robin_Watts | Ah. plancks doesn't have a robin? | 18:10.42 |
jogux_mac | no | 18:10.47 |
| plancks has nothing of interest at this stage :) | 18:10.56 |
Robin_Watts | I set up short cuts for both, and clicked the wrong one. | 18:10.57 |
| So, sorry! | 18:11.00 |
jogux_mac | np :-) | 18:11.04 |
Robin_Watts | So once on fermis, I ssh to root@ip_address ? | 18:12.24 |
jogux_mac | yes | 18:12.27 |
Robin_Watts | No root to host. | 18:12.53 |
| for deb32-etch-01 | 18:13.03 |
jogux_mac | can you possibly keep a note of everything you have to do (ie. install etc); I have a feeling we may end up doing this again :-) | 18:14.10 |
| I swear that was working earlier today :( | 18:14.15 |
Robin_Watts | jogux_mac: Will do. | 18:14.30 |
jogux_mac | but I was having multiple problems earlier | 18:14.36 |
Robin_Watts | Got ATS problems, I feel bad for you son. Got ATS problems, cos deb-etch won't run. | 18:15.21 |
| Sorry. | 18:15.31 |
jogux_mac | :) | 18:15.33 |
| is it laggy for you too? | 18:15.46 |
Robin_Watts | I can't ssh in, so I have no opinion :) | 18:16.27 |
jogux_mac | I got in ok, to .181 | 18:16.44 |
Robin_Watts | Right, 181 is looking better. | 18:17.07 |
jogux_mac | that's a wheezy machine, so is much newer than etch. | 18:17.20 |
| which one your build was done on when picsel last did it.... no idea. | 18:17.34 |
Robin_Watts | I'll look at setting up on that one then ? | 18:17.55 |
jogux_mac | yes | 18:19.05 |
| I can probably fix the etch one if it looks to be necessary (but probably not till tomororw) | 18:19.35 |
Robin_Watts | If you're happy for me to fiddle here, then I'm sorted for now. Thanks! | 18:19.55 |
jogux_mac | yes, very happy. we can restore it back to how it was easily :-) | 18:20.36 |
| assuming you don't find a way to remotely set computers on fire, like on the Arrow I watched last night. | 18:21.10 |
Robin_Watts | jogux_mac: I watched that one too! :) | 18:26.16 |
jogux_mac | :) | 18:39.01 |
jogux_mac | offs to pub. night all. | 18:39.06 |
ray_laptop | jogux: tip one for me :-) | 18:41.02 |
Robin_Watts | night jogux | 18:47.33 |
svilayphiou_ | hello, I drew cropmarks in Postscript and I want to paste them onto a PDF. But I get the error "Can't set GraphicsAlphaBits > 1 with a vector device pdfwrite." | 20:40.36 |
| this is my postscript file: http://dpaste.com/1775218/ | 20:41.41 |
| and then: gs -dNOPAUSE -sDEVICE=pdfwrite -dNOPAUSE -dBATCH -sOUTPUTFILE=output.pdf marks.ps input.pdf | 20:42.14 |
| any idea? | 20:42.18 |
| Forward 1 day (to 2014/04/10)>>> | |