| <<<Back 1 day (to 2013/07/01) | 2013/07/02 |
kens | chrisl I sent you an email with a proposed change to setpagedevice, would you mind taking a quick look please ? Its a bit big to post here (*lots* of comments), but not really worth pushing to Casper for review. | 08:31.18 |
chrisl | I'm reading it now..... | 08:33.23 |
kens | Thanks | 08:33.27 |
chrisl | 'Cept my mail client has decided to ignore me for a while..... | 08:34.35 |
kens | I thought you stopped using Thunderbird.... | 08:34.48 |
chrisl | Yeh, I'm using Claws Mail - but *something* changed a few weeks back, and now, every now and again, when it communicates with the servers, it just semi-freezes | 08:36.03 |
kens | weird.... | 08:36.13 |
chrisl | Yeh, I can't blame the mail servers because it started to happen with both gmail and spamcop at the same time. | 08:37.45 |
kens | Hmm, time to roll back a version or two ? | 08:38.00 |
chrisl | That's more hassle than it's worth - it might not be Claws Mail, but one the library dependencies. Although, oddly, it happens on both Ubuntu 12.10 *and* 13.04 | 08:39.26 |
kens | Yay for shared libraries.... | 08:39.43 |
chrisl | After I noticed this problem, I went back and tried TBird again - it doesn't suffer the same problem, but (part of?) the same old IMAP bug which made me stop using it in the first place, still occurs | 08:40.47 |
kens | is sticking with Eudora.... Obselete APIs never change | 08:41.09 |
chrisl | kens: so, those setpagedevice changes look sane (for sufficiently small values of "sane"). I can't help bug wonder if the DPS code really needs all these differences. | 08:42.07 |
kens | Well its the fact that it insists that teh 'savedinitialgstate' not have any local VM references, to avoid dangling pointers. | 08:42.41 |
| And since the gstate includes a pointer to the page device dictionary, we can't allocate that in local VM. | 08:43.07 |
chrisl | But that confuses me, since we're in the initial VM state, nothing there should disappear until we're shutting down anyway | 08:43.40 |
kens | But (WordPerfect bug) because we copy all child dictionaries to local VM, and then put them in the parent, the parent *must* always be allocated in local VM.... | 08:43.44 |
| chrisl, I make no promises on tha sanity of the DPS code :-( | 08:44.16 |
chrisl | Hmm, and that Wordperfect bug is just *wrong*...... | 08:44.49 |
kens | I agree totally, and the 'fix' is even worse, since it implictly relies on the parent dictionary being allocated in local VM | 08:45.16 |
| But OK, if you don't see anything totally loopy in there, and assuming the cluster test is OK, I'll commit it. | 08:46.10 |
chrisl | kens: that Wordperfect "fix" breaks the definition of currentpagedevice in the PLRM | 08:46.42 |
kens | It does work for -dSAFER now, as well as for setpagedevice when the VM alloction mode is global. Heavens only knows if it still works for broken WP files. | 08:46.48 |
| chrisl I know :-( | 08:47.31 |
| It specifically says its read-only | 08:47.43 |
chrisl | Personally, I would have been tempted to just kill the WP "fix" | 08:47.54 |
kens | chrisl I did try that, but it sort of spiralled off into oother problems. | 08:48.15 |
chrisl | I wonder if Distiller/CPSI return a writeable local dict..... | 08:48.51 |
kens | Let mer see | 08:49.00 |
chrisl | Jaws returns a read-only dict in local VM, even if the current VM mode is global, which is interesting | 08:51.28 |
Robin_Watts | just got a mail from Virgin. Apparently I'm silver again. | 08:51.50 |
kens | Distiller returns a local VM writeable dict | 08:51.53 |
Robin_Watts | Not quite sure how... | 08:52.00 |
kens | personal trips Robin_Watts ? | 08:52.09 |
Robin_Watts | The boston trip was BA | 08:52.22 |
chrisl | kens: that's a pile of crap, then :-( | 08:52.32 |
kens | No idea then. | 08:52.34 |
| chrisl yes I agree totally, but now we're compatible :-) | 08:52.46 |
Robin_Watts | and while I have a personal Virgin flight coming up, that hasn't happened for a while. odd. | 08:52.47 |
chrisl | I'm still silver | 08:53.25 |
| According to the web site, anyway | 08:53.44 |
kens | I feel left out :-( | 08:53.54 |
chrisl | I'm sure it won't last - it's probably because I was bit behind you guys to start with | 08:54.49 |
kens | I think youre right yes | 08:55.00 |
Robin_Watts | bug 694395. That sounds a lot like a commercial app to me. | 08:59.24 |
kens | is currently swamped with bugs, just a moment | 08:59.51 |
| Oh its MuPDF, that's OK then | 09:00.10 |
| and yes, it certainly does | 09:00.19 |
| I suppose it could be an in-house app, but.... | 09:00.29 |
| Also 'I want to load 100 PDF files, and I run out of memory' sounds like the response should be 'and ?' | 09:01.08 |
Robin_Watts | yeah. | 09:01.17 |
paulgardiner | So it's not the bloke that we did the one-page-per-doc stuff for? | 09:02.52 |
Robin_Watts | no. | 09:03.06 |
paulgardiner | He's probably creating a cache per doc | 09:04.54 |
pipitas | Is it possible for Ghostscript to do PDF->PDF conversion where embedded images do NOT get compressed with JPEG/DCT compression, but use flate/zip (or ASCIIHex or anything else but DCTEncode) instead? | 09:05.34 |
| When using this commandline, I get much much of JPEG artefacts than I like: | 09:05.44 |
| gs -o out.pdf \ | 09:05.55 |
| -sDEVICE=pdfwrite \ | 09:05.55 |
| -dPDFSETTINGS=/prepress \ | 09:05.55 |
| -sColorConversionStrategy=LeaveColorUnchanged \ | 09:05.55 |
| -sColorConversionStrategyForImages=LeaveColorUnchanged \ | 09:05.55 |
| -dCompatibilityLevel=1.4 \ | 09:05.55 |
| -dGrayImageResolution=100 \ | 09:05.55 |
| -dColorImageResolution=100 \ | 09:05.56 |
| -dGrayImageDownsampleThreshold=1.0 \ | 09:05.56 |
| -dColorImageDownsampleThreshold=1.0 \ | 09:05.57 |
| in.pdf | 09:05.57 |
| Hence, I tried to modify the command by adding other params to the command: | 09:06.04 |
| -sColorImageFilter=FlateEncode | 09:06.15 |
| -dColorImageFilter=/FlateEncode | 09:06.16 |
| or even | 09:06.28 |
| -c ".setpdfwrite <</QFactor 0.0 /Blend 1 /ColorTransform 1 /HSamples [2 1 1 2] /VSamples [2 1 1 2]>> setdistillerparams" | 09:06.28 |
| but to no avail... | 09:06.28 |
kens | QFacto and co are not distiller parameters | 09:06.45 |
| Those are arguments to teh DCTEncode filter, and will have no effect as Distiller params | 09:07.06 |
pipitas | kens: Well, QFactor is mentioned in a footnote of Ps2pdf.htm :-) | 09:07.39 |
kens | You mean note 7 ? It describes the default image parameters dictioanry | 09:08.31 |
pipitas | I realized that they are arguments to DCTEncode, but I tried to make it so that the JPEG artefacts go away. Didn't work though... | 09:08.40 |
| I took it from note 7 and thought I could modify these defaults with my commandline. | 09:09.16 |
kens | Note 7 says that they are the default image dictionary parameters | 09:09.34 |
| You need to set them to the ColorImageDict parameter | 09:10.16 |
pipitas | Yes, defaults which automatically get applied if nothing else is on the commandline. I *thought* my command would change that. Seems not. | 09:10.26 |
| How would I modify the ColorImageDict on the commandline then? | 09:10.55 |
kens | No, because you have set them to 'distiller params' and they aren't vvalid parameters, you need to do <</ColorImageDict <<........>> >> setdictillerparams | 09:11.07 |
pipitas | Aaaaah! | 09:11.18 |
kens | However, avoiding the JPEG encoding is the only way to go | 09:11.34 |
| <</ColorImageFilter /FlateEncode>> setdistillerparams ought to work I think | 09:12.24 |
pipitas | Let me try this | 09:12.37 |
kens | is disappearing under a mountain of questions today | 09:13.37 |
pipitas | No, doesnt work... | 09:13.46 |
kens | Well, I can't really comment further without seeing an example file | 09:14.06 |
pipitas | Wouldn't all these three be equivalent? | 09:14.49 |
| -sColorImageFilter=FlateEncode | 09:14.51 |
| -dColorImageFilter=/FlateEncode | 09:14.51 |
| -c ".setpdfwrite <</ColorImageFilter /FlateEncode>>setdistillerparams" -f | 09:14.51 |
kens | Probably | 09:15.06 |
pipitas | I can file a bug and attach an example | 09:15.08 |
kens | That's more likely to ge my attetion than anything else | 09:15.24 |
pipitas | Thanks, kens. Will do. | 09:15.42 |
kens | Hm, there's an Avon mobility who seel stair lifts in Bristol :-) | 09:33.23 |
| But I suspect avonmobility.com is more likely the people | 09:33.49 |
Robin_Watts | kens: Indeed. They have 2 windows store apps, but the link is broken to it, I think. | 09:34.22 |
kens | THey have 2 android apps as well, both crossword thingies | 09:34.42 |
| THis is their Android link: | 09:35.27 |
| https://play.google.com/store/apps/developer?id=Avon+Mobility+Solutions#?t=W251bGwsbnVsbCxudWxsLDEwMiwiY29tLmF2b25tb2JpbGl0eS5jb3JydXB0Y3Jvc3MiXQ.. | 09:35.27 |
| actually 5 apps in there, despite what ther web site says | 09:35.44 |
kens | wonders what he was doing before this VM allocation thing cropped up | 09:36.50 |
| Time for a caffeine top up I think | 09:37.56 |
| pipitas, a sudden thought. You will also have to set /AutoFilterColorImages to false to prevent getting DCT encoded output | 09:39.08 |
| Because otherwise pdfwrite will use DCT as it produces smaller output | 09:39.29 |
pipitas | Good hint, kens! Thanks. This sounds like it's the key! Will try⦠because so far all the resulting /Filter settings where /FlateDecode, even when I had specified /LZWEncode. | 09:52.35 |
kens | There are so very many distiller parameters its hard to remember them all | 09:53.00 |
| chrisl I messed up with my Distiller test, accidentally used rcheck instead of wcheck. Distiller behaves the same as Jaws, always returns a read-only dictionary allocated in local VM. | 09:57.41 |
| So the WP bug is just mad. If we ever get more problems I'll just take the time to kill it | 09:58.04 |
chrisl | kens: as I said, I would not be surprised if even WP had had to fix that one. Or maybe Adobe has an Idiom...... | 10:08.57 |
kens | chrisl, yes that's a possibility too. As I sadi, if it comesup again, I'll revisit it. (I hope it doesn't.....) | 10:09.37 |
| That's an inteeresting new bug : "ghostscript images are Morgaine;.e" | 10:12.10 |
| What on Earth is that supposed to mean ? | 10:12.22 |
| Oh he's changed the title | 10:13.01 |
| No sample file, no command line, refers to 2 differenrt versions of Kubuntu and if his bug report is to be believed is using an old version anyway.... | 10:15.57 |
Robin_Watts | paulgardiner: There is a fix on robin/master | 11:07.32 |
paulgardiner | Okay | 11:07.39 |
| Looks okay to me. You are setting copying obj from old to new twice, I think, but that does no harm. | 11:13.25 |
Robin_Watts | I am? | 11:14.46 |
| oh, I am. | 11:15.24 |
| I will remove that line. | 11:15.28 |
| Thanks. New version up. | 11:17.21 |
paulgardiner | Yes. Looks good. | 11:23.22 |
Robin_Watts | ta. | 11:25.11 |
| paulgardiner: Any thoughts on bug 694388 ? | 11:50.47 |
paulgardiner | Not sure. I believe that having multiple calls to requestLayout within a single event handler is pretty harmless, since it only sets a flag, and we do need a call to layout whenever anything moves. | 12:04.22 |
| On the other hand, if everything still works without the calls, might be worth going with. | 12:04.49 |
Robin_Watts | paulgardiner: I believe the final suggestion there is to leave the ones in all the leaf functions and remove the one in the more central place. | 12:05.33 |
| so we won't do layouts on UP/DOWN events. | 12:05.56 |
paulgardiner | Where I guess I should know what UP/DOWN events are. :-) | 12:07.11 |
| Ah yes. | 12:08.23 |
Robin_Watts | I was assuming you did, cos I didn't. | 12:08.48 |
paulgardiner | Just realised it's mouse up and down. :-) | 12:09.07 |
Robin_Watts | ah. | 12:09.42 |
| not relaying out on every up/down could be good then. | 12:09.56 |
paulgardiner | I would avoid calling layout when tapping, but it wouldn't lead to less layout calls when panning or zooming. | 12:11.13 |
| Lost layoutRequest calls, but not less onLayout calls. | 12:11.46 |
| s/Lost/Less/ | 12:11.54 |
| Still, I'd hope that all would still work without the layoutRequest call in onTouchEvent. | 12:12.56 |
| It would be interesting to check that we get just one onLayout call per animation step when doing inertial panning. Trouble is, the latest firmware seemed to break debug output on my S2 | 12:24.10 |
JimmyB | hey all., anyone here got experience running Ghostscript using c#? | 12:26.02 |
kens | JummyB not really no | 12:31.44 |
| JimmyB: * | 12:31.50 |
malc_ | paulgardiner: any hope of seeing annotation problem resolved? | 12:32.54 |
paulgardiner | malc_: Which annotation problem are you referring to? | 12:33.52 |
JimmyB | oh thats to bad, I have some issues with the c# implementation, but Ill looks elsewhere then, thanks tho :) | 12:36.43 |
Robin_Watts | Jimmy B: what C# implementation? | 12:37.02 |
malc_ | paulgardiner: after f07fedd90e5d08fba4687283a29760d08cde8d0c me and sebras were having significant issues with some pdfs | 12:37.04 |
paulgardiner | What sort of issues? | 12:38.33 |
malc_ | paulgardiner: "error: Annotation object not a dictionary" | 12:39.12 |
| over and over again | 12:39.23 |
Robin_Watts | paulgardiner: http://ghostscript.com/irclogs/2013/06/28.html | 12:39.34 |
| You discussed this with sebras at the top of that log. | 12:39.49 |
paulgardiner | Oh yeah. Thanks for the reminder. We decided how to fix that. | 12:39.54 |
| I'll have a quick look now. Should be easy. | 12:40.06 |
Robin_Watts | paulgardiner: Was there anything else I was supposed to be looking at before going back to pdfwrite? | 12:47.31 |
| I can't remember anything... | 12:47.41 |
paulgardiner | Not that I'm aware of | 12:47.52 |
Robin_Watts | oh, no, it was progressive stuff I was doing. | 12:48.06 |
| and that has to build on tor8's stuff. | 12:48.50 |
kens | chrisl can you make sense of what the reporter is saying in bug #694392 ? | 12:56.27 |
paulgardiner | Robin_Watts: there's a fix for the annotation problem on paul/master | 12:57.11 |
chrisl | kens: the colour conversion one? No, he's gibbering | 12:57.39 |
Robin_Watts | paulgardiner: I'm mid merge at the moment, will look asap. | 12:57.51 |
kens | chrisl yes the one where he talks about ps2write, then reopens the bug and says its pdfwrite (or does he ? I#'m confused) | 12:58.08 |
paulgardiner | Robin_Watts: no probs | 12:58.27 |
chrisl | kens: yeh, no idea. The command line he gives is *definitely* ps2write, and in the first comment states that pdfwrite "works" <shrug> | 12:59.29 |
kens | Yeah, well if he doesn't answere real quick its getting closed again. | 12:59.46 |
| Same for the German chap who has supplied a screenshot.... | 13:00.12 |
chrisl | That looks like just a color management change | 13:00.35 |
kens | chrisl it could be, probably is, but hwo can I tell ? :-) Given he's comparing 9.07 against 8.71 it could be nearly anything | 13:01.14 |
henrys | chrisl:about the new font names, they did the same with me for the PCL fonts and I just changed the names back after checking with them. You can do what is easiest. | 13:52.48 |
kens | Hmm, this looks like a typo to me in gxcpath,c cpath_is_rectangle(): | 13:57.32 |
| if (pcpath->path_valid) | 13:57.32 |
| return NULL; | 13:57.32 |
| Surely that hsould be if (!pcpath->path_valid) ? | 13:57.51 |
Robin_Watts | kens: let me look. | 14:07.51 |
kens | thanks Robin_Watts | 14:08.02 |
| It always returns NULL for me, which causes pdfwrite to emit clip paths as a sequence of lines, rather than a rectangle | 14:08.30 |
| I'm just running a cluster test now, but I don't anticipate any differences | 14:09.18 |
Robin_Watts | yes, that looks like a typo. | 14:10.17 |
kens | Ah, thanks for looking | 14:10.40 |
Robin_Watts | The next if accesses pcpath->path.bbox, which presumably it should only do if pcpath->path is valid :) | 14:10.49 |
kens | That's what I assumed yes | 14:11.03 |
Robin_Watts | blame goes back to igor in 2005. | 14:11.14 |
kens | do you have a URL for the commit ? | 14:11.27 |
Robin_Watts | http://git.ghostscript.com/?p=ghostpdl.git;a=commit;f=gs/base/gxcpath.c;h=76ea9c2f50f5b6a8cedc3d3c04cb9342326d2a80 | 14:11.35 |
kens | wow, a big change | 14:11.54 |
| ah, he added the whole routine | 14:12.13 |
| I'm reasonably sure only pdfwrite will be using it | 14:12.23 |
| This should improve file sizes slightly in some cases. | 14:12.38 |
| thanks for the confirmation Robin_Watts | 14:13.51 |
Robin_Watts | The commit says it was a fix for pdfwrite. | 14:14.35 |
kens | Indeed it is | 14:14.42 |
| And teh new routine *ought* to emimt rectangular clips as 're' but because of the typo, it never recognises rectnagles, and so continues to use a sequence of lines. | 14:15.32 |
| The bug I'm looking at predates that commit, so if this works I'm going to close it. I will check and see how much difference it makes (I'm guessing very little) | 14:16.15 |
| good grief, the original file is a 165 page document | 14:21.40 |
| crumbs, the change makes a huge difference. Acrobat is 1.094 MB, the old PDF is 2.019 Mb, the new one is 1.335 Mb | 14:24.10 |
| aha, of course we have some differences :-( | 14:24.45 |
Robin_Watts | large sections missing ? | 14:25.22 |
kens | don't know, about to do a bmpcmp | 14:25.31 |
| I suspect small differences due to rounding, but we'll see | 14:26.00 |
mvrhel_laptop | there is something weird going on for me with git, mupdf and the winrt solution file. when I do a switch to the lcms branch the winrt solution is missing and when I switch back it is still missing | 14:26.50 |
| back to master that is | 14:26.56 |
kens | Is the solution file being tracked in git or did you put it in a directory which is excluded ? | 14:27.30 |
mvrhel_laptop | it was being tracked by git | 14:27.44 |
kens | Then that exhausts my tiny knowledge :-) | 14:28.01 |
| Is the solution file supposed to be in regular GS on master ? | 14:28.33 |
| And if so, where is it ? | 14:28.39 |
mvrhel_laptop | oh it moved from some reason | 14:28.53 |
kens | Oh wait, this is MuPDF | 14:28.55 |
mvrhel_laptop | s/from/to/ | 14:29.00 |
| oh this was part of tors rearrangement | 14:29.28 |
kens | :-) | 14:29.34 |
Robin_Watts | mvrhel_laptop: It should be in platform/winrt now ? | 14:29.35 |
kens | I was about to suggest that as the reason | 14:29.42 |
mvrhel_laptop | Robin_Watts: yes | 14:29.43 |
| I had been running from my old one all this time... | 14:29.59 |
| I probably should do a git clean here | 14:30.10 |
| how do you do the smiley face with the crooked smile on here | 14:30.46 |
Robin_Watts | :-/ ? | 14:31.14 |
mvrhel_laptop | yes | 14:31.17 |
Robin_Watts | dunno | 14:31.21 |
mvrhel_laptop | funny | 14:31.28 |
Robin_Watts | : - / | 14:31.29 |
kens | forward slash | 14:31.35 |
mvrhel_laptop | I do like the rearrangement | 14:31.57 |
henrys | [:) | 14:32.13 |
mvrhel_laptop | sigh I have been working in my old winrt folder though. need to fix a few things now... | 14:33.05 |
henrys | oops I wrote to chrisl and he's not here. | 14:33.10 |
kens | I'm sure he'll read teh logs when he comes back | 14:33.27 |
mvrhel_laptop | oh my commit was ok though. it was in platform/winrt so all is well. | 14:34.10 |
| aha. there is some work to do. some of the paths in the winrt solution are broken | 14:36.57 |
| and lots of other stuff to update in the winrt solution | 14:42.36 |
Robin_Watts | henrys: Are we having a meeting today in the absence of tor? | 14:49.08 |
henrys | I'm fine skipping the mupdf meeting. I have to get up to speed on the fixes shelly and zeniko have done to estimate payments. | 14:50.16 |
| that was one thing I wanted to discuss | 14:50.26 |
paulgardiner | I wont be here for next week's meeting: I'm away on holiday 5th to 12th. | 14:50.29 |
mvrhel_laptop | oh ok. are we still having the gs meeting at 9? | 14:52.23 |
Robin_Watts | mvrhel_laptop: Did you have anything you wanted to discuss? | 14:52.47 |
mvrhel_laptop | I am getting the winrt solution fixed up now. | 14:52.59 |
henrys | well no tor isn't here ;-) I'll take any excuse. | 14:53.08 |
mvrhel_laptop | so I will have a commit for you in the next hour or so | 14:53.12 |
| I have the windows phone set up for developement now so I was going to get going on that | 14:53.34 |
| Also I did submit that windows 8 app to the store. for some reason in my first submission they claimed it crashed everytime. I think there was a mixup with how visual studio generated the upload packages for the various architectures. I fixed that and resubmitted | 14:54.44 |
| I beat on the thing heavily and could never get it to crash in x64, win32 or ARM | 14:55.09 |
| it also passed the testing tool that they supply here, which I did on the x64 and ARM | 14:55.30 |
Robin_Watts | mvrhel_laptop: Preparing stuff for submission always brings bugs to the fore. | 14:55.45 |
| Normally just after you think you've got it working, and done all the builds and uploaded it. | 14:56.03 |
mvrhel_laptop | yes :-/ | 14:56.29 |
| there it is... :) | 14:56.38 |
| so no gs meeting either today? | 14:57.22 |
Robin_Watts | yes, a gs meeting as normal. | 14:57.57 |
| I have emoticons turned off here. Otherwise I can never read code that's pasted. | 14:58.25 |
mvrhel_laptop | ok. I am going to be out tomorrow. I will send out email in a bit. | 14:59.07 |
| taking kids to kings dominion which is roller coaster after roller coaster. | 14:59.34 |
Robin_Watts | cool. | 14:59.44 |
| recreational queuing. | 15:00.09 |
mvrhel_laptop | yes. lots of lines | 15:00.36 |
| for a 3 minute ride | 15:00.49 |
henrys | mvrhel_laptop: have you looked at open xps which doesn't "open" in less than windows 8? | 15:05.57 |
mvrhel_laptop | henrys: mupdf handles openxps just fine | 15:07.25 |
Robin_Watts | henrys: MuPDF was broken w.r.t. openxps, but michael and/or tor fixed it a few weeks back. | 15:07.59 |
henrys | mvrhel_laptop: is it a large change? | 15:08.14 |
Robin_Watts | kens: I can't post to gs-devel, could you forward my mail please? | 15:08.29 |
mvrhel_laptop | henrys: tor had it working apparently for a while. I only made a change for the oxps type | 15:08.49 |
kens | Robin_Watts : oh right, I saw ther mail, I'll forward it on | 15:08.50 |
Robin_Watts | Thanks. | 15:09.00 |
henrys | Robin_Watts:the coverage for pcl link on the dash is broken - ghostscript is okay | 15:09.15 |
Robin_Watts | henrys: will fix | 15:09.41 |
| henrys: try now. | 15:12.03 |
henrys | Robin_Watts:the other thing I was thinking of was asking how michael sweet is doing with mupdf2raster and if he has any questions. His involvement could be significant, he is a big name in the community. | 15:12.14 |
Robin_Watts | henrys: Shall I send a follow up email asking how he's getting on? | 15:13.15 |
henrys | I was wondering if it would best come form michael since he was there. But I don't feel strongly about who sends it. | 15:14.08 |
| mvrhel_laptop: ^^^ | 15:14.20 |
| Robin_Watts: thanks coverage link fixed | 15:14.49 |
mvrhel_laptop | either way. Robin_Watts if you already have a dialog with him on this, I would suggest that continue | 15:15.10 |
Robin_Watts | Last email was between him/me/til/tech on the 1st June. | 15:15.27 |
mvrhel_laptop | I can take over it if you would like | 15:15.43 |
Robin_Watts | mvrhel_laptop: I'm happy for you to do it if you want, but if he has questions, it'll probably be me he needs to speak to, cos I wrote the new bit of code he'll be testing. | 15:16.20 |
mvrhel_laptop | then it would make sense for you to continue | 15:16.48 |
| no need for me to get in the middle | 15:17.01 |
Robin_Watts | ok. I'll keep tech in the loop so that if he comes back with questions while I'm away for any reason, you can jump in. | 15:17.29 |
mvrhel_laptop | brb | 15:33.27 |
chrisl | Robin_Watts, kens: Hmm, I rather figured if that chap on gs-devel couldn't fathom how to set compiler flags, he may not be someone we want playing with Ghostscript and threading..... ;-) | 15:55.01 |
kens | :-) | 15:55.18 |
| Oh, new sales man in Japan | 15:55.52 |
| That *will* be interesting | 15:56.07 |
kens | is highly confused now as to what path_valid means | 15:58.38 |
| If its 0 then we *do* want to write out the box as a rectangle. So what does that mean then ? | 15:59.08 |
mvrhel_laptop | back | 15:59.25 |
henrys | kens:right that's item 1 for the agenda | 16:00.06 |
kens | Oh meeting time | 16:00.18 |
chrisl | henrys: I'll change the font names so they match what we expect - I suspect anything else would cause havoc with the distributions | 16:01.47 |
henrys | One of the things that might be good for us (artifex) would be have the new sales guy hang out with us on IRC - I'd like to get more comms with sales and marketing. | 16:02.37 |
| but if you guys think this would be a time sink I understand. | 16:02.59 |
| let me know what you think | 16:03.10 |
kens | That sounds good, time zones could be difficult | 16:03.10 |
Robin_Watts | henrys: I don't see how it can be bad to have him here. | 16:03.38 |
henrys | true, timezones, I am going to talk to him today and I'll see what I can do. | 16:03.58 |
Robin_Watts | If he takes up too much time, we'll just ignore him. We do that for other people already ;) | 16:04.11 |
kens | Well he can get you and ray first thing his time, and us last thing in the day | 16:04.17 |
| Its got to be easier for him to chat and ask us any questiosn, than send emails back and forth | 16:04.49 |
Robin_Watts | henrys: Are you still keen on us all installing/using skype? | 16:04.55 |
kens | THough he may get confused if he sits in here :-) | 16:05.01 |
kens | has no way ot use skype | 16:05.14 |
Robin_Watts | kens: buying a webcam is not hard :) | 16:05.35 |
henrys | Robin_Watts:I am but I haven't done it myself so I'd like to get set up first before I "recommend" it. | 16:05.49 |
kens | I'd prefer not to use a webcam | 16:05.53 |
| Isn't voice enough ? | 16:06.06 |
Robin_Watts | kens: I have a webcam, and it's perfectly possible to use the webcam just for the microphone. | 16:06.16 |
henrys | voice is sufficient | 16:06.25 |
kens | would stick with just a mic then | 16:06.51 |
Robin_Watts | but IME buying a webcam is a simpler path to getting a worthwhile mic than trying to buy a mic. | 16:07.04 |
kens | I can use a gaming headset | 16:07.26 |
Robin_Watts | Ah, perfect. | 16:07.36 |
henrys | Robin_Watts:I'll go about setting up Skype than I'll nag ;-) | 16:07.43 |
Robin_Watts | If people want to do test calls, I am set up on skype. | 16:08.03 |
henrys | Robin_Watts:I'll probably get it going after the meeting today. | 16:08.36 |
| I was hoping to just use built in stuff on a Mac Pro | 16:09.08 |
kens | Macs used to come with mics, I don't know if they still do | 16:09.29 |
Robin_Watts | henrys: skype works fine with my mac laptops camera/mic. | 16:09.30 |
| And there is a skype mobile app for android and ios too. | 16:09.57 |
henrys | Yes I did an entire meeting remotely - talking head and it worked surprisingly well. | 16:10.10 |
Robin_Watts | We did a test call on saturday between me/scott/miles/ray all with video, and that worked OK. | 16:10.56 |
henrys | oh has miles signed up artifex? I was talking to him about that. | 16:12.17 |
Robin_Watts | Helen and I both have this one: http://www.ebay.co.uk/itm/Logitech-E250-Computer-PC-Desktop-Webcam-with-built-in-microphone-UK-Stock-/111088469489?pt=UK_Computing_ComputerComponents_Webcams&hash=item19dd6385f1 | 16:12.19 |
| henrys: No, I think Scott has a premium account. | 16:12.31 |
| You only need 1 person with a premium account in order for everyone to use video. otherwise you're limited to 2 video feeds. | 16:12.57 |
| You can do unlimited numbers of people in a voice only call for free. | 16:13.12 |
henrys | anyway the new sales guy seems to have something of a technical background, should I forward his resume to tech? | 16:14.14 |
mvrhel_laptop | so is he on board for sure? | 16:14.41 |
kens | Miles' mail says so | 16:14.49 |
henrys | mvrhel_laptop: yes | 16:14.57 |
Robin_Watts | is nosey enough to be interested in reading the resume. | 16:15.01 |
henrys | I'll send the resume upon request. I don't think it appropriate to broadcast. | 16:15.50 |
| 1 request so far. | 16:16.19 |
mvrhel_laptop | oh just got his email | 16:16.31 |
chrisl | I think I hate Skype already....... | 16:16.36 |
mvrhel_laptop | I would like to see his resume too please | 16:16.46 |
henrys | is ray_laptop goes on vacation tomorrow | 16:17.26 |
| I'm going through the online agenda does everyone have the pointer to that - should I add or delete anything? | 16:19.58 |
Robin_Watts | kens: I haven't looked at the context of that patch too much; is it possible that the code is only sending the clipping path as rectangles if there is not a path it can use ? | 16:20.18 |
henrys | Robin_Watts put a link on the dashboard | 16:20.19 |
kens | Robin_Watts : it seems we cna have a clipping path (but not a path) which has a rectngle, but no path elements | 16:21.09 |
| In which case path_valid is 0, but the rectangle *is* valid | 16:21.23 |
Robin_Watts | henrys: The one thing that I really miss with the workflowy thing is that when I took notes at the meeting on the old style text agenda, I could note the things that were mine. | 16:21.30 |
kens | I don't know how we get these | 16:21.34 |
henrys | Robin_Watts:do you have some top of your head idea of shelly and zeniko hours? | 16:21.43 |
Robin_Watts | Is it possible to put "[robin]" or something at the end of each line to say who currently "owns" that task? | 16:22.08 |
chrisl | kens: intersection of two rectangular paths? | 16:22.19 |
Robin_Watts | henrys: We agreed figures for the mupdf fuzzing, for them, right? | 16:22.29 |
kens | chrisl, no its created that way | 16:22.39 |
Robin_Watts | I haven't been keeping up to date with the work they've done on the jbig2dec etc since. | 16:22.57 |
chrisl | kens: does it come from the interpreter? Or created internally? | 16:23.29 |
kens | I don't know | 16:23.38 |
Robin_Watts | kens: suppose I make a clipping path from path A. And another from path B. | 16:23.43 |
kens | but there are many of them soo I assume from the interpreter | 16:23.49 |
Robin_Watts | Then I make C, the intersection of those two. | 16:23.51 |
| The way that works is that A and B are flattened to rectangles, and the sets of rectangles intersected to give C, AIUI. | 16:24.14 |
henrys | Robin_Watts:I'll figure out shelly and zeniko - I think the best thing to do is get a list of bugs from them first. | 16:24.27 |
kens | Robin_Watts : but that must generate more than one rectanlge (or more than one clip) in the general case | 16:24.41 |
Robin_Watts | kens: in the general case, yes. | 16:24.53 |
kens | I'm unconvinced its that | 16:25.37 |
| But it doesn't really matter | 16:26.36 |
henrys | if everyone wants names next to the items I'm fine doing it but does everyone have the problem? It doesn't seem like a very long list and you sort of know what's yours and I'd rather you read all the items anyway. You may have something to contribute. | 16:26.56 |
kens | Did I miss sopmething there henys ? | 16:27.25 |
henrys | yes Robin_Watts wants owners names next to agenda items. | 16:27.49 |
Robin_Watts | henrys: It's not that I haven't read the list, it's that I like to scan the list to check I haven't forgotten stuff that is assigned to me. | 16:27.54 |
henrys | we are looking at the agenda now. | 16:27.56 |
kens | Oh I see, missed that | 16:28.10 |
Robin_Watts | clearly not every item would be assigned. | 16:28.18 |
kens | I'm not bothered either way | 16:28.29 |
Robin_Watts | kens: cpath_init_rectangle sets a clippath up to just be a single rectangle. | 16:29.10 |
kens | Yes | 16:29.17 |
| Ah, I wonder if that results from cliprect | 16:29.33 |
| Not the way I'd have done it, but never mind. | 16:29.51 |
henrys | I'll do it if we agree that it is a requirement to read other folks items? I'm concerned about the change. | 16:30.00 |
kens | need to put dinner on, brb | 16:30.25 |
henrys | oh sorry we should be done. | 16:30.45 |
| that went fast | 16:30.48 |
mvrhel_laptop | I would not be opposed to having names on the items where applicable, but I can see my list from the forest too | 16:30.50 |
Robin_Watts | henrys: I certainly wouldn't have assumed that because some items are assigned to particular people that I wouldn't need to read the whole thing. | 16:30.59 |
henrys | okay as soon as I hear from everyone I'll update it. | 16:32.23 |
mvrhel_laptop | henrys: so wait, what do you need to hear from us | 16:32.40 |
chrisl | henrys: I'm okay either way | 16:32.53 |
henrys | mvrhel_laptop: I've heard from you ;-) | 16:32.55 |
mvrhel_laptop | ok | 16:33.02 |
henrys | anything else for the meeting? | 16:33.15 |
mvrhel_laptop | I do make my own to-do list during the meeting | 16:33.24 |
kens | is back | 16:33.51 |
henrys | Sorry I ran the meeting late kens | 16:34.09 |
kens | Not a problem | 16:34.14 |
henrys | see above. | 16:34.16 |
kens | henrys, I don't mind either way | 16:34.33 |
Robin_Watts | henrys: well, don't worry about it just for me then. | 16:34.49 |
henrys | Robin_Watts:if nobody cares it's fine to do. Miles would like it but I don't count him ;-) | 16:35.20 |
ray_laptop | hi, all | 16:35.43 |
henrys | ray_laptop:see above... | 16:35.47 |
ray_laptop | I got here before ken left :-) | 16:35.53 |
marcosw | sorry, lost I'm late too | 16:36.17 |
ray_laptop | my laptop was offline while I was trying to get help from the VMware people. | 16:36.58 |
henrys | maybe email + cron or something would have some effect. | 16:37.02 |
mvrhel_laptop | ok. well I am heading out for a bit unless someone has something for me | 16:37.07 |
henrys | nope I'm done | 16:37.19 |
ray_laptop | I upgraded to 9 and now I can't install either. | 16:37.26 |
| I'm reviewing the logs... | 16:37.35 |
| w.r.t. the .jp sales guy, early during his day is the end of the day for us on (or near) the Pacific, so we can just watch for him signing on to IRC | 16:44.04 |
kens | ray_laptop : yes I said that, and the end of his day is the beginning of ours | 16:44.26 |
ray_laptop | I often work in the evenings | 16:44.27 |
henrys | marcosw:anything with the plotter? I think the problem I assigned to you can go back to the customer I don't know that I can do anything else. | 16:46.50 |
| the plotter will just tell me I'm right or stumped - neither changes the code. | 16:47.23 |
marcosw | sorry, no. I ordered the other part I needed last week, but I haven't received it. I'll ask them if it's been shipped. | 16:47.27 |
ray_laptop | henrys: it's strange that that other package seems to work for them. I suspect, however, they only mention it when it produces what they want and we don't. There's probably an equal or greater number of files they get wrong | 16:50.43 |
henrys | ray_laptop:yes he has at least 5 different programs and he reports the one that works. 2 others we've worked with don't work. | 16:52.05 |
| whey is 693684 with ray_laptop and not Robin_Watts ? | 16:54.16 |
| s/whey/why | 16:54.23 |
marcosw | henrys: presumably when I get my plotter back up and the file does not print correctly that will be the end of it. | 16:54.35 |
henrys | well there is a fix in there that he needs to get. | 16:55.00 |
| marcosw ^^^ | 16:55.28 |
Robin_Watts | henrys: It's with ray because ray is applying the changes he's already done in this area for a customer. | 16:55.58 |
marcosw | henrys: I'm reading the bug now | 16:56.12 |
Robin_Watts | i.e. this is pretty much a reapplication problem, rather than a new code problem, I believe. | 16:56.16 |
henrys | Robin_Watts: thanks | 16:57.13 |
kens | OK I'm off, night all | 16:57.53 |
henrys | we need an email address for the new employee, marcosw can we just update our account and not be always worried about getting another email address, is it much more expensive? | 16:59.36 |
| sebras did you get you NDA from miles yet? | 17:18.19 |
| so if you type in # in the workflowy search entry field things should be clear | 17:25.41 |
Robin_Watts | ooh, nice. | 17:26.40 |
henrys | well it isn't clear how you guys are going to fix all that crap but it looks nice. | 17:28.07 |
| you know I never noticed it, sabrina pointed it out to me, you can't silence the phone key pad on the iPhone. Unbelievable | 17:32.43 |
Robin_Watts | henrys: That would be an option. Apple don't do options. | 17:33.25 |
| but why would you want to silence the phone key pad? | 17:36.40 |
| If you're going to be making phone calls from a place where the sound of the keys is an issue, surely you shouldn't be making calls from there anyway? | 17:37.23 |
ray_laptop | Robin_Watts: why is bug 694385 P1 (highest priority)? Your comment says marking it low priority | 17:40.15 |
| oops That's bug 694384 (typo_ | 17:40.38 |
Robin_Watts | ahem. I'll mark it P4. | 17:41.02 |
| ooh, we have a P5 now. | 17:41.06 |
| Is that some super secret special designation? | 17:41.25 |
ray_laptop | Robin_Watts: AFAIK we've always had P1 (highest) through P5 (lowest) | 17:42.25 |
| I think the default priority is P4 initially, and we rarely bother to downgrade from there | 17:43.05 |
Robin_Watts | fair enough. | 17:45.01 |
marcosw | Robin_Watts: Just ordered a big, heavy thing from Lufthansa WorldShop for you to carry over the Chicago for the next staff meeting. | 17:58.21 |
Robin_Watts | marcosw: OK... | 17:58.46 |
marcosw | btw, I was stalking you on Google Maps, where is there a white van in front of your house? Are you a white van man on the side? | 17:59.18 |
Robin_Watts | Hopefully you'll be bringing me a games console :) | 17:59.19 |
| marcosw: It's a grey van technically :) | 17:59.35 |
| And it's Helens, officially. | 17:59.42 |
marcosw | Robin_Watts: oops, sorry. | 17:59.51 |
Robin_Watts | She uses it for carting the harpsichord about. | 17:59.54 |
| but it is a very small games console. | 18:00.26 |
marcosw | is this the kickstarter thing that hangs off of a hdmi port? | 18:01.26 |
Robin_Watts | marcosw: yeah. | 18:01.52 |
| gamestick.tv | 18:02.08 |
| Supposed to be shipping mid august. | 18:02.36 |
marcosw | did you get a raspberry pi camera yet? They are sold out in the US. | 18:03.02 |
Robin_Watts | marcosw: My raspberry pi has been assembled into a case, and hasn't yet been plugged in. | 18:03.26 |
marcosw | I've compiled ghostscript on mine. I don't recall how long it took, but I decided not to add my raspberry pi to the cluster. | 18:04.26 |
| speaking of which, I wonder where mine is. It's too small to keep track of in my very cluttered office. | 18:05.18 |
Robin_Watts | marcosw: That's why I got it a case :) | 18:06.04 |
marcosw | everyone should bring their pi to the next staff meeting so we can have a lan party :-) | 18:08.59 |
ray_laptop | Robin_Watts: did you get the case or the box ? | 18:13.41 |
Robin_Watts | ray_laptop: I got a perspex case thing. | 18:14.10 |
| but I don't understand the question really - there are many different cases/boxes out there. | 18:14.37 |
ray_laptop | Robin_Watts: sorry, I was just looking at the Adafruit choices. There are a lot of others out there. | 18:20.11 |
Robin_Watts | http://www.kickstarter.com/projects/ciseco/pi-lite-white-bright-white-led-display-for-raspber?ref=category | 18:39.32 |
sebras | marcosw: is there a gps module for the pi? that might solve your problem. :) | 18:54.32 |
Robin_Watts | http://www.raspberrypi.org/archives/3916 | 18:55.51 |
sebras | Robin_Watts: well, that kit solves marcosw's problem in two ways: 1. gps-coordinates 2. the pi is no longer too small to lose track of! | 18:56.47 |
Robin_Watts | http://www.adafruit.com/blog/2013/01/24/adafruit-ultimate-gps-on-the-raspberry-pi-raspberry_pi-raspberrypi/ | 18:56.57 |
| sebras: Indeed. | 18:57.00 |
ray_laptop | sebras: maybe not quite enough positional resolution, but that would still help. Of course, unless he's running it on a battery all he has to do is follow the cord :-/ and if it's not running, the GPS module doesn't help too much | 19:34.32 |
henrys | ray_laptop:are you about? | 21:10.58 |
Robin_Watts | henrys: If it's just you wanting someone to skype, I'm here. | 21:26.44 |
henrys | Juichi.Takanezawa at artifex dot com - is there any cultural reason he can't have the same format as the rest of us? That was my question | 21:27.10 |
| or should he be Takanezawa.Juichi at artifex dot com | 21:28.43 |
| Robin_Watts: sent you a contact request | 21:38.53 |
Robin_Watts | accepted. | 21:39.52 |
henrys | thanks | 21:42.09 |
| Forward 1 day (to 2013/07/03)>>> | |