| <<<Back 1 day (to 2014/04/07) | 2014/04/08 |
robin_watts_mac | heads home. Will probably be home just after the meeting finishes. | 04:23.28 |
chrisl | ray_laptop: It was just coincidence that I happened to be writing some other (personal) mails when Len's arrived - figured I might as well jot him a note while I was there. | 06:36.44 |
| ray_laptop: the code 532 uses doesn't use UFST for TrueType - when I did that work, I hadn't got downloaded TTFs working reliably. | 06:38.17 |
norbertj | chrisl: ray_laptop: me too. I also force the fontrasterizer to FreeType in pl_load_tt_font | 08:12.44 |
| chrisl: ray_laptop: until it works | 08:13.35 |
chrisl | norbertj: TBH, the ufst 6.x code is better wrt downloaded TTFs, *but* my testing shows no advantage to using it - in fact, in general, it seems very slightly slower than Freetype - especially as, I know you disable TTF hinting for performance, and you can't really do that with UFST | 08:14.47 |
norbertj | chrisl: I also install additional TTF files, (to be loaded at startup) and that I did not get to work with UFST. With FreeType it does. | 08:20.41 |
chrisl | norbertj: Yes, it mostly can be made to work with a little tweaking, but as I said, I haven't pursued it because it doesn't seem to have any benefits. I did say before that if it was something you feel you'd get value from, I would look at it again, but as it's not something I would generally recommend, I don't currently plan to devote much time | 08:24.02 |
norbertj | chrisl: as quality/performance with FreeType is also ok, and the comparison against the legacy printer is OK, its fine with me. | 08:28.21 |
chrisl | norbertj: FWIW, I also found that Freetype is more tolerant of slightly out of spec fonts - of which there are a depressingly large number out there! | 08:30.13 |
chrisl | reboots - back in a mo..... | 08:33.06 |
betaboon | hello :) | 08:39.24 |
jogux_mac | hi | 09:03.53 |
chrisl_r500 | Good morning | 09:06.20 |
chrisl | My broadband is going to be down for a few hours - will be back when it returns - or I get bored enough to use the mobile..... | 09:35.21 |
jogux_mac | wow, this latest openssl security bug sounds pretty nasty ( http://heartbleed.com/ ) | 11:47.15 |
tor8 | jogux_mac: yeah, I wouldn't want to be the developer responsible for that bug today... | 12:17.43 |
jogux_mac | hehe. I wouldn't like to be the sysadmin for a hosting company responsible for renewing 4 billion potentially compromised ssl certificates ;-) | 12:18.39 |
tor8 | jogux_mac: yeah. bad stuff all around. | 12:19.29 |
henrys | meeting in 50 minutes - a short one hopefully ... | 12:49.34 |
kens | Because Robin and Paul aren't here ? >:-) | 12:49.58 |
henrys | nor michael | 12:50.23 |
kens | OhMichael too. Will be a sparesely populated meeting then. No wonder its been so quiet | 12:50.47 |
henrys | tor8: we looked at goodreader right? and didn't think it used mupdf? | 12:56.30 |
jogux_mac | goodreader hugely predates mupdf iirc. I remember buying it back in 2008 or so, and it had been around a while by that point :-) | 13:00.43 |
| hm. actually, mupdf has been around much longer than I thought :-) | 13:06.18 |
betaboon | is there a way to have something similiar to the shrinkwrap of mupdf but instead of adjusting the window size adjusting the content ? | 13:09.16 |
henrys | jogux_mac: any guess what an android build with the "Good" SDK would take? It looks like we want to make the Good business a priority now. | 13:13.52 |
jogux_mac | let me just check. | 13:14.13 |
| is there a reason you're only asking about android? | 13:14.18 |
tor8 | henrys: I don't remmeber | 13:14.24 |
henrys | jogux_mac: there was a working iOS release but no android so ... | 13:14.45 |
jogux_mac | ah, I was about to guess at that ;-) | 13:15.01 |
tor8 | betaboon: W and H shortcuts I think | 13:15.36 |
henrys | jogux_mac: when it was working Good was selling SOT on iOS and a competitor on droid | 13:15.40 |
jogux_mac | right. | 13:16.15 |
| looks like good were given an android build of SO library about a year ago; wonder if the intention was that they'd do the work. | 13:17.01 |
henrys | kens:I gave the wrong time for the meeting it's in an 1:15 from now | 13:17.56 |
| marcos is out too. | 13:20.16 |
| I suggest we just cancel this one | 13:20.40 |
kens | Ah, that's better :-) | 13:20.41 |
| Well I have nothing to say myself anyway | 13:20.56 |
henrys | tor8: are you waiting on robin and paul for the release? | 13:21.31 |
tor8 | henrys: the meeting time keeps changing! I can never keep it straight this way :) | 13:21.32 |
| henrys: I pushed out a release candidate yesterday; if people could download it and see if the windows binary and source build out of the box I'd be grateful | 13:22.14 |
jogux_mac | henrys : looking through the changes that were done for iOS it doesn't look trivial. First guess would be it's at least a weeks work; I think I'd need to look for at least an hour to make a list of what would need to be done. | 13:22.51 |
tor8 | henrys: but I think we're in fairly good shape for a release soon as robin and paul get back | 13:22.55 |
henrys | tor8: okay I guess robin usually does the windows build and test? | 13:23.56 |
jogux_mac | we've not even got a test environment for good at the moment :( | 13:24.00 |
tor8 | henrys: I do the windows build, but let robin (and usually kens) test it for me | 13:24.24 |
| just to make sure the windows build doesn't rely on anything robin has set up locally (since I never use MSVC normally, it's a clean slate) | 13:24.45 |
betaboon | tor8: thanks. just noticed the version i am running is old and doesn't support those. gotta recompile :D | 13:24.47 |
tor8 | betaboon: try the 1.4 release candidate :) | 13:25.00 |
| jogux_mac: you could try building the source on a mac (just unzip and type "make" or "make HAVE_X11=no" if you don't have an X server) | 13:25.50 |
kens | tor8 I never noticed a release candidate, where can I download it ? | 13:25.51 |
henrys | jogux_mac: is the test framework something lost in transition? | 13:26.29 |
tor8 | ketn: mupdf.com/downloads/archive/mupdf-1.4rc1-windows.zip | 13:26.41 |
kens | OK I'll fetch it now | 13:26.59 |
jogux_mac | henrys : never having used good I'm not 100% sure - I'm guessing it's probably just a case of registering developer accounts, getting a server setup, etc. | 13:27.15 |
betaboon | tor8: just compiled from git :) | 13:27.28 |
jogux_mac | tor8 : I'm not sure I'm following; was that a request for me to help? | 13:29.28 |
kens | tor8 I just had a dialog from MuPDF I've never seen before, very reminiscent of Acrobat. 'File has unsaved changes. Do you want to save' | 13:30.17 |
| THat's the RC of course | 13:30.24 |
tor8 | jogux_mac: only if you want and are not busy. | 13:30.24 |
| kens: yeah, I think that's new from paulgardiner's forms work | 13:30.43 |
| if you edit any forms it'll prompt you to save the changes | 13:30.54 |
kens | I hadn't edited the file | 13:31.00 |
tor8 | but if you didn't edit any forms, then that's probably a bad thing to see... | 13:31.09 |
kens | I don't believe its a form file either | 13:31.12 |
| Just a moment | 13:31.16 |
| Ourt test file BEAZH3P31.pdf, I'mfairly sure that's a QL test | 13:31.46 |
| Acrobat doesn't offer to save on exit, so I'm guessing the file is OK | 13:33.00 |
henrys | jogux_mac: I can test the mac | 13:33.14 |
jogux_mac | tor8 : there's a whole bunch of deprecated / unused warnings, but it builds, runs and I can open a pdf file :-) | 13:33.41 |
kens | tor8 Ghostscript does warn of a problem with the xref, so the file may not be valid | 13:33.43 |
tor8 | jogux_mac: yeah, there's a handful of known "deprecated" warnings and some unused stuff in thirdparty libraries | 13:34.38 |
| apple has marked all of openssl as "deprecated" to scare developers into using their own proprietary mac crypto library | 13:35.06 |
kens | tor8 even if I allow MuPDF to 'save the changes' and then open the saved file, it still wants to save the changes on exit. I think there's somethign wrong there | 13:35.23 |
tor8 | kens: one thought might be if mupdf has synthesized any annotation appearance streams... but yeah, that sounds wrong. | 13:35.39 |
| I will have to get paul to investigate when he gets back | 13:35.53 |
kens | Sure, I didn't think it was something you;d want to tackle | 13:36.06 |
tor8 | kens: I tested on calc.pdf so expected the save dialog :) | 13:36.17 |
kens | I'm not getting it on 'most' files | 13:36.32 |
| It may be only on files that are damaged in some way | 13:36.43 |
jogux_mac | thinks it would be awesome if we could do a native mac ui for mupdf, and perhaps even stick it on the appstore. | 13:36.48 |
tor8 | kens: ohhhh, yeah. might be the repaired xref triggering the flag? | 13:37.05 |
| but you said you got it when re-opening the saved file as well... | 13:37.26 |
kens | tor8 I thought it might be that, which is why I allowed it to savbe, then opened the saved file, but that still triggeered the warning | 13:37.33 |
tor8 | jogux_mac: yeah. I'd say go for it, but I doubt henrys would let you escape from SmartOffice so easily ;) | 13:37.58 |
kens | Ah, the saved file still fires a warning in GS. So it looks lke the save didn't actually fix the problem | 13:38.10 |
jogux_mac | henrys : do you want us to look further into good/android? | 13:38.15 |
tor8 | kens: it might have done an incremental save | 13:38.24 |
| which would still have the same inherent problems | 13:38.32 |
jogux_mac | tor8 : Yeah, it seems unlikely ;-) | 13:38.34 |
kens | tor8 yes it looks like an incremental update | 13:38.53 |
| New trailer but the xref is unchanged | 13:39.03 |
tor8 | but it does warrant thinking about; we'd be the one and only XPS viewer on macosx | 13:39.05 |
henrys | jogux_mac: yes, is there something I should do here contacting them? | 13:40.37 |
jogux_mac | henrys : well... I guess first question might be - is there a Picsel good dynamics account that we've taken over and Miles has (or can get) the login info for? | 13:41.20 |
henrys | jogux_mac: so we need to fix iOS and do android. | 13:41.44 |
| jogux_mac: I'll look into that. | 13:42.10 |
jogux_mac | henrys : failing that we may need to register via http://www.ndm.net/mobile/Good/good-dynamics-for-developers etc. | 13:42.35 |
kens | tor8 apart from the odd 'saving changes' thing, the RC seems OK to me | 13:44.24 |
tor8 | kens: thanks. I guess we'll have to fix the saving changes thing, that looks disturbing | 13:46.17 |
henrys | jogux_mac: I just forwarded a support request we got so you have their info if you need it. I'm asking them about the account now. | 13:46.52 |
jogux_mac | henrys : ok, thanks. I'll leave it with you for now. (The android stuff is probably one for Pete, I find android distateful and he doesn't :-) ) | 13:49.28 |
henrys | ray_laptop: we decided to skip the meeting. | 14:31.45 |
ray_laptop | henrys: yes, I saw. Thanks | 14:33.19 |
| henrys: IMHO, we shouldn't give a refund to someone that reports a bug unless they give us enough information that might let it reproduce and/or fix it. And then only after we determine that we won't fix it in a reasonable time. | 14:37.51 |
| but at least a full bug report should be required | 14:38.34 |
henrys | on android you have 15 minutes to get a refund without the developer's involvement. | 14:48.27 |
| and that uninstalls the app properly I suppose. | 14:48.46 |
Robin_Watts | re | 14:52.25 |
kens | welcome back Robin_Watts | 14:52.33 |
Robin_Watts | so, meeting was cancelled? | 14:55.29 |
kens | Yes, not quorate :-) | 14:55.40 |
henrys | Robin_Watts: couldn't have a proper meeting without you ;-) | 14:56.13 |
| Robin_Watts: did you have something to pass on? michael and marcos are also out. | 14:59.28 |
Robin_Watts | nope. | 14:59.39 |
henrys | and chris too I think | 14:59.56 |
kens | chris has lost broadband | 15:01.20 |
henrys | jogux_mac: I've sent email to the good contacts, but I'm betting we'll have to register again, I'll let you know as soon as I hear back. | 15:02.15 |
Robin_Watts | I thought that good sold servers to companies. | 15:02.54 |
| so it's not a question of registering with good, so much as them giving us a server? | 15:03.27 |
jogux_mac | robin_watts : I think registered developers get access to the server software. | 15:03.42 |
Robin_Watts | ok. | 15:03.48 |
jogux_mac | but I'm presuming there's an existing revenue stream that is associated with an existing account | 15:04.00 |
henrys | kens:just him or is his area out? | 15:10.09 |
kens | henrys I don't know, I was assuming just chris | 15:11.31 |
norbertj | henrys: hi, I noticed that I too :( had some typos in the commit (in message and in pcparse.c ) shall I correct them? | 15:45.14 |
henrys | norbertj: I think it's okay. Nice to have an editor pop up in git that has a checker if you can set that up. | 15:48.06 |
chrisl | Yay, broadband is back..... | 16:08.34 |
kens | welcome back chrisl :-) | 16:09.02 |
chrisl | kens: sorry I missed you mail earlier, but good you got it sorted | 16:09.31 |
kens | Yeah I was being dumb | 16:09.39 |
| It always takes me ages to sor out adding a new PS operator, especially when I have to add a C file too | 16:10.00 |
chrisl | Yeh, I don't find it very intuitive.... | 16:10.31 |
kens | Its just the sheer number of things you have to do, all of which have to be correct or it simply doesn't work (or in my case, initially, build...) | 16:11.06 |
| But I'm well on the way I think. THe revision 6 security handler hash generation is being done i C, not PostScript | 16:11.50 |
| Partly because that way I can steal gobs of code from MuPDF | 16:12.10 |
chrisl | You obviously lack the self flagellation gene that encourages writing that kind of crap in Postscript! | 16:13.40 |
kens | I started it, then decided it was too much like masochism | 16:13.59 |
| Copying around big chunks of data in PostScript string was just insane | 16:14.22 |
chrisl | Just because you can, doesn't mean you should! | 16:15.05 |
| The meeting was cancelled, then? | 16:15.40 |
kens | TBH I don't think it was terribly sensible to do the key genertion for revision 5 in PostScirpt either, but that's done now and I'm not changing it | 16:15.46 |
| Yes, lack of a quorum for meeting | 16:15.54 |
chrisl | Oh, okay, I thought the logs were rather short for a Tuesday! | 16:16.07 |
kens | Robin, Paul, Michael, Marocs and you all missing | 16:16.07 |
chrisl | So, apparently I now have 152Mb downstream / 12Mb upstream on my broadband.... I can't say IRC seems much quicker! | 16:17.00 |
henrys | chrisl: fiber? | 16:22.14 |
chrisl_r500 | henrys: cable modem | 16:22.35 |
| Fibre to the cabinet | 16:22.46 |
| TBH, it's highly questionable how useful it is, *but* I'm on an old, no longer available account which bundles cable modem, TV and phone, and changing to a lower lower speed internet would mean a new account, and not a great deal of money saving | 16:26.28 |
henrys | Robin_Watts: should a bmpcmp cluster push cause a build? seems odd it wouldn't just use the last exe built. | 16:52.55 |
kens | henrys, there might be a cluster run in between, or another user job | 16:53.21 |
henrys | kens:mine were really close together, but yes that could be. | 16:54.10 |
kens | henrys they often are, but say you expected no diffs, when the log came in you looked at it and decided you needed a bmpcmp, in the meantime someone could have scheduled a test | 16:54.46 |
| THe cluster can't tell that the source hasn't changed | 16:55.04 |
chrisl | henrys: it's been discussed before: the cluster has no provision for retaining binaries, and it would (apparently) add a lot of complexity to allow it to happen | 16:55.43 |
henrys | chrisl: I was doing a bmpcmp for a couple files and it fired up ~30 builds, maybe something we should review as the cluster grows. | 17:03.37 |
chrisl | henrys: possibly - it might be worth seeing if marcosw can record how often a bmpcmp immediately follows a normal push - if that's fairly rare, it might now be worth the effort | 17:05.10 |
| s/now/not | 17:05.20 |
ray_laptop | since each user has their own cluster test code, and the bmpcmp always uses the Differences from the previous run, it seems like just doing 'make' before bmpcmp is enough | 17:11.20 |
kens | ray_laptop : we do, that's what Henry wants to avoid | 17:11.43 |
chrisl | Actually, I think we "clean" up everything between runs, so we have to rebuild everything. | 17:13.10 |
ray_laptop | kens: now we do a make clean (as chrisl said) | 17:13.49 |
| just doing a make would suffice, IMHO | 17:14.06 |
kens | Yeah, but as chris also says, we do a clean, because we can't be certain what has changed | 17:14.22 |
| At least, so I understnad it | 17:14.29 |
ray_laptop | since it is not at all likely that the source changes from a regression run to the subsequent bmpcmp will change something requiring a make clean | 17:15.00 |
kens | That's true for a user's directory, do we always do a make in the user's home ? | 17:15.46 |
ray_laptop | usually NOTHING changes, but even if a source file changed, make dependencies *should* suffice | 17:15.46 |
| kens: yes | 17:15.56 |
| kens: not the /home/kens but the /home/marcos/cluster/users/kens | 17:16.37 |
kens | Got to go and cook.... | 17:16.49 |
chrisl | The only thing might be possibly the rsync process could confuse make as to what has or hasn't changed | 17:16.54 |
ray_laptop | chrisl_away: great work on cust 532 issue. Thanks. | 18:26.17 |
henrys | jogux_mac: Brad said to just create a new account and call if you need help. | 21:39.44 |
jogux | henrys : right, okay, thanks. | 21:40.31 |
| this is higher priority than bug fixing, right? | 21:40.36 |
henrys | yes current plan is embollix to do that and artifex robin and paul continue on bugs | 21:41.21 |
jogux | ok | 21:41.36 |
henrys | emobix sorry | 21:42.18 |
jogux | I'm not sure if it's going to confuse Good that pete/I don't have artifex email addresses | 21:47.34 |
henrys | Jogux: I'll just set it up then send you what you need. | 21:49.00 |
jogux | ok, great, thanks :-) | 21:49.21 |
henrys | they just now gave me a different link : https://community.good.com/community/gdn | 21:49.35 |
| to use | 21:49.41 |
jogux | yeah, I just found that one too. Not sure how I got a different one earlier :-S | 21:49.57 |
henrys | Jogux: and now I'm to wait for a GDN partner group for the onboarding process, it all makes sense now ;-) | 22:14.26 |
Robin_Watts | I think we may now have to refer to them as embollix all the time :) | 23:02.17 |
| Forward 1 day (to 2014/04/09)>>> | |