| <<<Back 1 day (to 2014/03/26) | 2014/03/27 |
chrisl | kens: openjpeg and luratech are for jpx not jpeg | 09:41.57 |
kens | D'oh | 09:42.03 |
| well the findings are still correct | 09:42.10 |
| What are we using for JPEG ? | 09:42.20 |
chrisl | libjpeg | 09:42.25 |
kens | Fair enough | 09:42.34 |
chrisl | That's for GPL and Artifex releases | 09:42.49 |
kens | Still seems to me Acrobat is fixing broken JPEG dtaa though, and the correct answer is 'don't use broken files' | 09:43.10 |
| Thanks for sorting out logikos last night | 09:43.35 |
chrisl | That wasn't hard - personally, I don't think a worked example is that important. If you know Postscript, it's fairly clear what to do, and given that you are injecting Postscript, you kind of have to know it! | 09:44.34 |
kens | chrisl true, but... AN example wouldn't be hard to do, I may add one when I finish pounding my head of 6 colour ICC profiles | 09:45.03 |
chrisl | Well, if you get to it before 5pm this evening, my pastebin post should still be available | 09:45.45 |
kens | I'll try and remember to grab a copy :-) | 09:46.10 |
chrisl | Although, like I said, it's not exactly difficult | 09:46.32 |
kens | No, but I figure there will be other people who will want to use this feature | 09:46.54 |
| At least some won't know PostScript at all | 09:47.08 |
paulgardiner_lap | Morning. Any recommendations for IRC client on the MAC? | 10:25.07 |
Robin_Watts | chatzilla (asa firefox extension) | 10:26.13 |
kens | I think henrys uses collquoy | 10:26.54 |
| http://colloquy.info/ | 10:27.21 |
Jogux | is using ircii. but it probably wasn't a good choice :@-) | 10:32.32 |
tor8 | paulgardiner_lap: colloquy is a good macosx irc client | 10:32.36 |
paulgardiner_lap | great thanks | 10:33.29 |
jogux_mac | gives colloquy a go :) | 10:35.04 |
spanners | irssi in a terminal :-P | 10:42.20 |
jogux_mac | the dns for twiki.ghostscript.com seems⦠unhappy (I canât resolve it). squish.net agress: http://dns.squish.net/traverses/b6a6944982d8a9cb890894b4af8f3b72 | 10:56.49 |
| agrees, even | 10:56.55 |
jogux_mac | wonders if bumping the zone serial number on casper might help, if anyone has access to do that :) | 10:58.38 |
Robin_Watts | jogux_mac: I have root on casper. | 10:58.59 |
| I think peeves is our secondary though and I don't have root on that, so that may not have been updated. | 10:59.27 |
jogux_mac | is hoping that tickling the primary correctly would solve the issue on the secondary | 10:59.47 |
Robin_Watts | where is the serial number? | 11:00.31 |
jogux_mac | just a sec | 11:00.42 |
| bah, colloquy just aim a whole paragraph because it started with a / | 11:04.45 |
| ate, even | 11:05.00 |
| file is /etc/bind/master/ghostscript.com | 11:05.03 |
Robin_Watts | All irc clients do that, pretty much. | 11:05.15 |
jogux_mac | change 2014030502 to 2014032700 then /etc/init.d/bind9 restart probably | 11:05.23 |
Robin_Watts | Jogux: Sorry, yes, I have that file open. | 11:05.27 |
jogux_mac | robin_watts : yeah, usually they keep it in the line buffer though :-( | 11:05.33 |
Robin_Watts | jogux_mac: This is ubuntu, so: sudo service bind9 restart | 11:06.10 |
| done. | 11:06.13 |
jogux_mac | ta! | 11:06.17 |
| yep, that did it, secondary is correct now, thanks | 11:06.43 |
| whoever added twiki mightâve forgotten to update the serial number | 11:07.02 |
Robin_Watts | ah, great! | 11:07.04 |
| I added twiki. | 11:07.12 |
| and my bind-fu is weak. | 11:07.32 |
jogux_mac | ah :-) trick is to always update the serial number whenever you change anything, as thatâs what the secondaries look at to figure out if they need to refetch the zone | 11:08.14 |
jogux_mac | still fails to get my mac to resolve it even after flushing everything. boggle. | 11:09.48 |
Robin_Watts | will remember that, thanks. | 11:10.39 |
| jogux_mac: presumably caching ? | 11:10.57 |
jogux_mac | yeah, but Iâve flushed our local dns serverâs cache (and thatâs definitely okay now), and dscacheutil -flushcache on the mac, and restarted firefox. | 11:11.22 |
| although itâs the middle step thatâs failing, as âpingâ canât resolve it either. | 11:11.45 |
Robin_Watts | jogux_mac: but there will presumably be caching at your ISP level. | 11:11.59 |
jogux_mac | nah; our local bind is recursive. | 11:12.09 |
| and using host against that shows it can resolve it okay | 11:12.23 |
| ahha. killall -HUP mDNSResponder fixed it. boggle. | 11:13.50 |
| I was about to ask if ghostdocs is private now. but I see it is. could I have access to it perhaps? :-) | 11:19.02 |
Robin_Watts | ghostdocs where? | 11:19.19 |
jogux_mac | http://twiki.ghostscript.com/do/view/GhostDocs/WebHome | 11:19.26 |
Robin_Watts | I thought I'd added you to the relevant group. Let me check. | 11:19.44 |
| Try now ? | 11:20.23 |
jogux_mac | thatâs it, thanks | 11:20.51 |
jogux_mac | is going to throw my iOS build / release notes on there; presume thatâs okay. | 11:21.01 |
Robin_Watts | absolutely. | 11:21.20 |
jogux_mac | right, here we go: http://twiki.ghostscript.com/do/view/GhostDocs/BuildiOS | 11:27.16 |
| (formatting is not elegant, apologies) | 11:27.47 |
| paul_gardiner: http://twiki.ghostscript.com/do/view/GhostDocs/BuildAndroid ;-) | 11:28.17 |
| (hint, hint ;-) ) | 11:28.42 |
paulgardiner_lap | jogux_mac: :-) | 11:29.37 |
jogux_mac | suspects we should move that release tracking spreadsheet somewhere else; possibly itâs own .git, or a google spreadsheet if we like those. | 11:30.38 |
kens | OK So I registered with Twiki, but I don't have access. | 11:30.43 |
Robin_Watts | kens: I need to add you. | 11:31.23 |
kens | Would you mind then please ? | 11:31.34 |
| Do you need info from me ? | 11:31.53 |
Robin_Watts | kens: I don't. Just pondering whether we should all follow the same format for our names... | 11:32.16 |
kens | Umm, I just used my IRC one, maybe not the best choice | 11:32.33 |
Robin_Watts | kens: yeah, not sure whether it matters. | 11:32.58 |
| but if it does, probably best to correct it now. | 11:33.07 |
kens | has no idea | 11:33.10 |
| Yes indeed, any idea how I can change it ? | 11:33.27 |
Robin_Watts | I can do it I think. | 11:34.13 |
kens | is rummaging through options, but not finding anythign yet | 11:34.36 |
Robin_Watts | 5 mins. | 11:34.41 |
kens | OK | 11:34.46 |
paulgardiner_lap | robin_watts okay I've registered | 11:35.28 |
Robin_Watts | kens: Can you close your twiki windows please? | 11:44.10 |
kens | Robin_Watts : I thought I had | 11:44.18 |
| In fact I have | 11:44.27 |
| I'll close firefox | 11:44.31 |
| there | 11:44.38 |
jogux_mac | ahha. colloquy does have a line buffer for previous lines, but you use option-cursor-up to acces it. | 11:47.08 |
Robin_Watts | kens: OK. I deleted you. Reregister :) | 11:50.36 |
kens | ok | 11:50.40 |
| What name/nick should I use for consistnecy ? | 11:50.50 |
Robin_Watts | KenSharp | 11:51.53 |
kens | OK | 11:51.57 |
| OK, better now ? | 11:53.09 |
Robin_Watts | Looks good. | 11:53.27 |
| both you and paul should be able to look at GhostDocs ? | 11:53.34 |
kens | I'll try now | 11:53.41 |
| Yes, seems OK | 11:54.00 |
Robin_Watts | Jogux: hehe. BuildiOS confuses the wiki. it appears as Buildi OS. | 11:54.24 |
jogux_mac | Robin_Watts : oh, where does it appear like that? | 11:54.42 |
kens | Yes I noticed that, assumed it was a typo | 11:54.42 |
| In Recent Changes for one | 11:54.54 |
Robin_Watts | Under "RecentChanges" | 11:54.54 |
jogux_mac | The wiki really wanted me to make it BuildIos, but I wasnât sure | 11:54.57 |
| d'oh. | 11:54.58 |
Robin_Watts | kens: The way twiki works, anything in CamelCase is assumed to be a WikiWord. | 11:55.21 |
| That appears as a link to another page. | 11:55.28 |
jogux_mac | Shall I just make it BuildIos? | 11:55.33 |
| or any other suggestions? | 11:55.40 |
Robin_Watts | jogux_mac: BuildIOS ? | 11:55.51 |
| but whatever, we can cope. | 11:56.04 |
jogux_mac | nods. okay. itâs now http://twiki.ghostscript.com/do/view/GhostDocs/BuildIOS | 12:02.41 |
tor8 | just make it BuildIos, IMO | 12:02.46 |
| stupid wordmark capitalisation is stupid. | 12:03.13 |
Robin_Watts | tor8: I can make a MuJS web if you feel that will be helpful. | 12:07.33 |
Robin_Watts | runs. back in a bit. sooner if it rains. | 12:09.37 |
logikos | hi kens, after troubleshooting i have several of one model of hp printer (laserjet 4000n) which when using ps2write i get this error: http://paste.ubuntu.com/7162341/ printed on a page from the printer | 12:13.26 |
| if i use an older version of ghostscript and pswrite it works | 12:13.52 |
| prints fine | 12:13.55 |
kens | logikos, what happens if you don't try and insert the page commands, what happens if you try to run the file through Ghostscript instead of sending it toyour printer ? | 12:14.11 |
logikos | seems to take alot longer to create the .ps but it does work | 12:14.12 |
kens | Don't use pswrite, its nasty | 12:14.19 |
logikos | what do you mean if i dont try and insert the page commands | 12:15.52 |
kens | If you don't use the /PSDocOptions, (thereby not insrerting your own PostScript) does hte outptu work on that printer ? | 12:16.21 |
logikos | oh sorry, that printer does not support that | 12:16.39 |
| i was not useing it when i got the error | 12:16.45 |
| gs -sstdout=/dev/null -sstderr=/dev/null -q -sDEVICE=ps2write -sPAPERSIZE=letter -dNORANGEPAGESIZE -o"|rlpr -h -H10.0.6.3 -Praw" ./invoice.pdf ./test.pdf | 12:17.05 |
kens | OK, so you are saying that the output of ps2write doesn't work on a specific printer ? | 12:17.11 |
logikos | that will produce the error | 12:17.11 |
| even if i output to a .ps file | 12:17.17 |
| then in a seperate command send it to the printer with lpr | 12:17.25 |
kens | OK what if you send that file to the printer which *did* work, does that work OK ? | 12:17.32 |
logikos | yes | 12:17.41 |
| the other printer works fine | 12:17.44 |
| its a very fancy printer | 12:17.50 |
| i can go look at the model number if it maters | 12:18.00 |
kens | Well that 'sounds like' a buggy PostScript implementation on that priner | 12:18.04 |
logikos | i have several units of that model printer | 12:18.15 |
| all throughout the plant | 12:18.19 |
kens | First thing to do is send me the PostScript file | 12:18.26 |
logikos | replacing all the printers isnt an option right now | 12:18.28 |
kens | So I can modify it and we cna try to debug the problem (if you don;t mind working with us on this, it may take a little time) | 12:18.48 |
logikos | k, i'll re-create it again this morning and re-test to varify the issue is still there | 12:19.01 |
| then send you the .ps | 12:19.04 |
kens | OK send it to ken.sharp at artifex.com | 12:19.20 |
chrisl_r500 | IIRC, the error was on the image operator - possibly worth trying different image data encoding options? | 12:19.24 |
logikos | i dont mind | 12:19.25 |
kens | The error seems to be on setpagedevice from the pastebin | 12:19.43 |
| configurationerror on setpagedevice | 12:19.58 |
| I suppose hte printer *does* support A4 paper ? | 12:20.38 |
chrisl_r500 | Then I suggested using the option to tell ps2write not to change the page size, which prompted the error on image | 12:20.45 |
kens | Oh, so there are 2 problems | 12:20.59 |
logikos | of course | 12:21.00 |
chrisl_r500 | I assumed that the printer's policies were to throw an error if it didn't have the right paper size | 12:21.28 |
kens | chrisl I was wondering that, which is why I asked :-) | 12:21.41 |
| an error on image is more likely to be a buggy implementation :-( | 12:22.08 |
chrisl_r500 | I think it was am ioerror on image, which is why I thought about tweaking the encoding options | 12:25.24 |
kens | Yes, makes sense, can try that later. | 12:25.37 |
jogux_mac | robin / paulg : should we delete/make inaccessible the various epage.gitâs before anyone ends up getting confused? | 12:26.11 |
chrisl_r500 | kens: We've had filter problems with HP printers before | 12:26.17 |
kens | chrisl I know :-( Mayeb we should find out what configuration options Till is currently using with HP printers | 12:26.46 |
chrisl_r500 | kens: I can't remember for sure, although I think the HPs were lesser offenders | 12:28.39 |
kens | Probably true, I was just wondering if we could find out what options CUPS uses. | 12:29.09 |
logikos | kens: email sent | 12:29.35 |
| the email will be from todd at rmtohio.com | 12:29.47 |
chrisl_r500 | kens: one thing I noticed: if you ask ps2write for uncompressed fonts, you get the "bare" TTF - no ASCII encoding | 12:29.47 |
kens | OK thanks, I'll look at it in a bit | 12:29.49 |
| chrisl_r500 : Yes, I guess that's true | 12:30.09 |
chrisl_r500 | Not with any urgency, but it would be nice to allow uncompressed, but still ASCII85 encoding for TTFs | 12:30.52 |
kens | I'm not sure we have enough options to allow for that separation at the moment, and I'm not keen to add yet more options unless we really need them | 12:31.36 |
kens | heads for lunch | 12:32.05 |
| logikos I got the mail | 12:32.22 |
| My mail client doesn't like your PNG :-P | 12:32.46 |
chrisl_r500 | Hmmm, http://www.openprinting.org is s..l..o..w.. :-( | 12:33.24 |
logikos | the png isnt important | 12:40.06 |
| they make me put it in the signature its just the company logo | 12:40.21 |
jogux_mac | have we done ios ad-hoc builds of mupdf before? (tor8?) Iâm thinking weâll want to do adhoc builds of smartoffice for testing so curious about whatâs been before/how. | 13:03.16 |
tor8 | jogux_mac: yes, I think so (if I still remember what adhoc build means in apple's odd terminology) | 13:05.45 |
| we have added all our iDevices to the signing stuff and then you can install an archive using itunes, IIRC | 13:06.20 |
jogux_mac | thatâs the one | 13:06.32 |
tor8 | but it's been a while, and I'm sure everything that I remember has changed. | 13:06.34 |
| (I don't have a Mac anymore, ever since I handed off the iOS development to Paul) | 13:06.51 |
jogux_mac | tor8: ahhh. Okay, didnât know that. Iâll bug Paul. thanks :) | 13:07.16 |
tor8 | jogux_mac: the developer account we have is still in my name, but talk to paul and he should be able to give you all the details you need | 13:09.35 |
| and I'm sure we've got another account for picsel than the one we use for mupdf | 13:09.54 |
jogux_mac | tor8 : yeah, thereâs a proper artifex account with (most) of the picsel apps in it | 13:15.07 |
| I guess we should move mupdf over into that same account | 13:15.20 |
Robin_Watts | jogux_mac: back. I think we should delete the epage.git's *after* we've moved stuff across successfully :) | 13:38.10 |
jogux_mac | I thought we had moved stuff across? | 13:38.29 |
Robin_Watts | jogux_mac: Stuff has been moved in pauls repo, but not in golden. | 13:40.13 |
| We should review all the commits before pushing them into golden. | 13:40.21 |
jogux_mac | ah. right. ok. | 13:40.28 |
| When you say âweâ⦠| 13:40.32 |
Robin_Watts | golden being /home/git-private/sot.git | 13:40.33 |
jogux_mac | (Paul/I have reviewed our commits btw) | 13:40.54 |
Robin_Watts | I mean, each person that did the commit should look back at it, check they are happy with it, and then get someone to review it. | 13:40.59 |
| OK, so it's just me :) | 13:41.09 |
| I'll try and get that done today. | 13:41.14 |
henrys | reads about all the stuff students (who've never seen our code) are going to do to mupdf and ghostscript. What could possibly go wrong ? ;-) | 13:41.30 |
jogux_mac | actually maybe Paul has older commits - Iâve certainly reviewed the ones he did a week ago or so for the releases. | 13:41.32 |
Robin_Watts | My earliest commits were done in a hurry to get demo builds out. I'd like to check they are sane. | 13:41.56 |
jogux_mac | henrys : oo, neat - google code of summer stuff? | 13:41.58 |
Robin_Watts | henrys: where? | 13:42.08 |
henrys | http://www.linuxfoundation.org/collaborate/workgroups/gsoc/google-summer-code-2014-openprinting-projects | 13:42.23 |
tor8 | henrys: most interesting. | 13:51.34 |
paulgardiner_mac | henrys: jogux_mac pointed out to me earlier that we might need bigger discs in some of the cluster nodes in Marcosâs garage to run the ghostdocs DA tests. I wondered if we should just go ahead and upgrade up front before we actually run into the problem, on the basis that bigger discs canât be a bad thing. I donât know how much hassle is involved though | 13:53.25 |
henrys | paulgardiner_mac: I imagine that would be something marcosw would dread in proportion to the number of machines affected. Why are we doing this, can we cut back testing on the flash player ? ;-) | 13:57.46 |
Robin_Watts | henrys: You are probably right in that we don't need all the testdocs. | 14:02.37 |
| Each of the uber servers has 4x250GB discs. | 14:03.07 |
| each about half full. | 14:03.13 |
| so we could put 100Meg of stuff on each and cross mount. | 14:03.35 |
| Or we could buy a single large disc and put it in each server. | 14:03.49 |
| Each node can accomodate up to 3 discs (so 12 discs per server) if memory serves. | 14:04.10 |
| I think marcosw has previously said he's not averse to putting a large disc in each server if required. | 14:04.31 |
| henrys, chrisl: Do you want to register on the TWiki too, so I can put you on the GhostDocs access list? | 14:05.13 |
henrys | Robin_Watts, paulgardiner_mac, there certainly isn't a money issue so whatever you can talk marcosw into doing is fine. | 14:05.48 |
| Robin_Watts: yes I want to register | 14:06.07 |
Robin_Watts | henrys: go to twiki.ghostscript.com and use the register link under "Twiki Users" | 14:06.55 |
| I think we should encourage people to use full names on TWiki rather than irc nicks. | 14:07.37 |
henrys | no https registration page ? ;-) | 14:09.00 |
Robin_Watts | henrys: possibly I could move the whole lot over to https... but that would require apache hacking. | 14:09.47 |
ray_laptop | I know peeved is not in the same class as teh uber servers (it is "only" a dual Xeon config), but it _does_ have 5 more 2.5" slots for hard drives | 14:10.31 |
Robin_Watts | 2.5" ? | 14:10.45 |
| laptop drives? | 14:10.50 |
ray_laptop | robin_watts: yes. | 14:10.57 |
Robin_Watts | ray_laptop: I wasn't dissing peeves. Just the testdocs probably want to be co-located with where ATS runs. | 14:11.15 |
ray_laptop | robin_watts: peeved, not peeves, but that's in case we wanted more CPU's running part of the ATS | 14:11.51 |
Robin_Watts | ray_laptop: sorry, typo. | 14:12.03 |
| ray_laptop: If you would like to register for the twiki, please go to http://twiki.ghostscript.com/ and use the Register link under "TWiki Users" | 14:12.38 |
ray_laptop | robin_watts: OK. And ffrom what I read, I should register as RayJohnston ? | 14:13.35 |
Robin_Watts | please. | 14:13.44 |
henrys | Robin_Watts: so I can get RSS notifications for all changes? | 14:14.24 |
Robin_Watts | henrys: Yes. | 14:14.32 |
| henrys: You should be able to view the GhostDocs web now. | 14:14.50 |
tor8 | Robin_Watts: is it really supposed to be that slow? it's taking over 2s to load each page... | 14:15.42 |
Robin_Watts | tor8: I can't explain that. | 14:16.11 |
henrys | Robin_Watts: the "Home" menu doesn't have ghostdocs web | 14:16.33 |
Robin_Watts | henrys: Indeed. Because it's not public. | 14:16.43 |
ray_laptop | robin_watts: OK. Done | 14:17.01 |
Robin_Watts | ray_laptop: You should be able to view the GhostDocs web too. | 14:17.16 |
ray_laptop | robin_watts: yes, I can | 14:17.30 |
tor8 | Robin_Watts: 2570 ms to get the first page, then 64 requests and it isn't done until the 5s mark... | 14:17.54 |
Robin_Watts | I could probably arrange to have the GhostDocs web show up in the menu, but just not be readable. | 14:18.04 |
tor8 | 4739ms to get the front page... | 14:18.42 |
Robin_Watts | tor8: If you can think of something for me to tweak to improve it, then please let me know. | 14:18.52 |
henrys | Robin_Watts: how is RSS going to work if it isn't public? | 14:18.56 |
tor8 | 6 seconds to load my user page... I have no idea what's making the web server take so long | 14:19.13 |
| do we have mod_perl on the web server? | 14:19.21 |
Robin_Watts | henrys: everything is public except for GhostDocs. | 14:19.24 |
| rss is done on a per-user basis isn't it? | 14:19.39 |
| So you'll get notifications of any webs you can view. | 14:19.55 |
| tor8: yes, we have /usr/lib/apache2/modules/mod_perl.so | 14:20.56 |
henrys | Robin_Watts: as long as the notification itself is public, it's fine I'm worried that's a potential hole. | 14:20.58 |
Robin_Watts | henrys: I don't believe a public user can get notifications about the GhostDocs stuff. | 14:21.36 |
henrys | Robin_Watts: but the wiki itself looks great and I bet it will be very helpful. | 14:21.40 |
tor8 | Robin_Watts: I wonder, does the logging do reverse dns lookups? | 14:21.56 |
| (is it slow for anyone else or is it just me?) | 14:22.11 |
Robin_Watts | but feel free to look at it on a different browser (so you're not logged in) and see what the RSS says. | 14:22.12 |
| tor8: It's not speedy for me, certainly. | 14:22.22 |
| tor8: maybe, http://twiki.org/cgi-bin/view/Plugins/CacheAddOn ? | 14:23.52 |
| <IfModule mod_perl.c> | 14:24.54 |
| # Mod_perl preloading | 14:24.56 |
| PerlSwitches -T | 14:24.58 |
| </IfModule> | 14:25.00 |
ray_laptop | tor8: what should I look at to see if it is slow ? | 14:25.42 |
tor8 | Robin_Watts: is there a more minimalist skin for twiki? | 14:25.54 |
Robin_Watts | tor8: probably, yes. | 14:26.05 |
tor8 | ray_laptop: just go to any page. it takes multiple seconds before any page loads for me. | 14:26.12 |
Robin_Watts | Feel free to play in your sandbox. | 14:26.16 |
henrys | Robin_Watts: seems rss just works for public stuff - gs is okay ghostdocs not | 14:26.26 |
jogux_mac | itâs slow for me too. I think twiki in general is not lightweight :( | 14:26.27 |
Robin_Watts | henrys: So you don't get ghostdocs notifications at all? Even if you're logged in? | 14:26.59 |
ray_laptop | tor8: seems pretty snappy for me. Less than 1.5 seconds for any thing I click on | 14:27.13 |
henrys | Robin_Watts: I did it from a foreign browse I suppose it would work with a login. | 14:27.33 |
ray_laptop | is this running on an Amazon instance ? | 14:27.50 |
Robin_Watts | ray_laptop: It's on casper. | 14:28.07 |
henrys | Robin_Watts: it doesn't work when I'm logged in either. I use feedly. | 14:29.52 |
jogux_mac | it might be possible to expose http://twiki.ghostscript.com/do/view/GhostDocs/WebRss via http authentication; I think Iâve seen that done before and a reasonable number of rss clients support it. | 14:29.53 |
Robin_Watts | henrys: You may need to use some other notification method then. | 14:30.09 |
tor8 | ray_laptop: I wouldn't call anything >0.5s "snappy" :) | 14:30.11 |
henrys | Robin_Watts: that's fine. I was more concerned about a security hole. | 14:30.33 |
Robin_Watts | lunches. will fiddle afterwards. | 14:30.35 |
| tor8: mod perl doesn't appear to be enabled. That might help a lot. | 14:30.46 |
jogux_mac | http://twiki.ghostscript.com/do/view/GhostDocs/WebNotify may be useful; email notifications. | 14:30.50 |
ray_laptop | tor8: many of the times it is in the .5 second range | 14:31.00 |
henrys | jogux_mac: so are we going to import the contents of the picsel wiki? | 14:31.17 |
jogux_mac | henrys : I believe the current plan is to bring up the picsel wiki on a machine in marcosâs garage that we can access via some private method | 14:31.44 |
| henrys : then to migrate any useful content onto twiki.ghostscript.com by hand. | 14:31.57 |
tor8 | ray_laptop: odd, cause I've never had a <1.5s response on any of the pages.. which makes me suspect something else is going on | 14:32.12 |
jogux_mac | Thereâs a *lot* in the Picsel twiki, over 50% of it wonât be any use to us I reckon. | 14:32.22 |
henrys | jogux_mac: As one goal it would be nice if neophytes like michael and I could build, run and debug various systems without lots of hand holding. | 14:32.47 |
tor8 | what sort of markup does twiki use? is it Markdown (which seems to be what all the kids use these days) | 14:33.23 |
jogux_mac | henrys : That sounds good. eg: with http://twiki.ghostscript.com/do/view/GhostDocs/BuildIOS Iâd hope you can sit down a do an iOS build without anything else other than a huge amount of swearing at iOS development certificates. | 14:34.14 |
| Iâll probably write up a windows 7/8 + vs2013 desktop express page for building windows too, unless anyone beats me to it. [Iâm not going to encourage anyone to use vs2005, itâs a PITA to setup with the extras required for the picsel builds ime.] | 14:35.43 |
henrys | jogux_mac: probably michael will use vs 2013, android is probably important and I'd like linux but I can do 2013 too. | 14:37.42 |
jogux_mac | oh, linux is likely easy. but Iâd probably encourage you to use windows for debugging to start with, as the picsel windows app has a bunch of built in tools that arenât in the linux app. | 14:38.55 |
| (personally Iâd rather use linux. or at least mac os.) | 14:39.10 |
| for android I believe paul_gardiner is going to write up what he knows ;) | 14:39.38 |
ray_laptop | mvrhel_laptop: you probably want to register with the twiki as well (see the logs) | 14:56.58 |
mvrhel_laptop | ok will do | 14:57.09 |
ray_laptop | mvrhel_laptop: the recommended user name will be FirstLast (MichaelVrhel) | 14:57.43 |
| (as I learned from the logs) | 14:58.08 |
mvrhel_laptop | have to run daughter to school now. bbiab | 14:58.36 |
chrisl | Robin_Watts: done | 15:10.19 |
Robin_Watts | chrisl: should be enabled. | 15:15.51 |
| Jogux: I use VS2005 all the time for picsel stuff. What extras are you referring to? | 15:17.30 |
| tor8: mod perl is enabled already apparently. | 15:26.30 |
tor8 | Robin_Watts: where are the twiki log files? | 15:39.00 |
Robin_Watts | tor8: Not sure offhand. | 15:39.34 |
| See /etc/apache2/sites-available/twiki.ghostscript.com for where I guess? | 15:39.56 |
tor8 | Robin_Watts: in /usr/lib/apache2/twiki/data/log* apparently | 15:43.20 |
| it doesn't seem to be doing reverse DNS lookups on the IP numbers for logging, so scratch that thought | 15:43.39 |
| (sshd used to do that, which caused me no end of grief when it had to wait for a reverse dns timeout every time I used ssh) | 15:44.05 |
henrys | marcosw or anyone was a performance check done with the new lc's code. I just did a profile of tiger and I'm surprised to see lcms weighing in as much as it does. | 15:50.09 |
Robin_Watts | henrys: I didn't check performance when I pulled lcms2.6 in. | 15:51.20 |
| Can you time before and after the import to see if there is a measurable difference? | 15:51.50 |
chrisl | Can't you just check the cluster times? | 15:52.21 |
henrys | marcosw: has all that stuff - in theory if there was a big change it would have been caught. | 15:53.23 |
| Robin_Watts: I doubt I'd be able to discern any difference with tiger.eps, interested in seeing the cluster times. | 15:56.36 |
chrisl | We get a performance regression report, I'd guess every week - but I'm not sure how to interpret the numbers.... | 15:56.59 |
| TBH, I think the cluster is too noisy for that regression report to be any use..... | 15:58.31 |
| Robin_Watts: did you clusterpush the LCMS 2.6 commit before pushing it? | 15:58.59 |
Robin_Watts | chrisl: I did. | 15:59.38 |
chrisl | So you'll have overal cluster times in the report from that | 15:59.53 |
Robin_Watts | chrisl: oh, only gs lowres though :( | 16:00.43 |
| Do user runs include times? | 16:01.43 |
chrisl | Yes. have done for ages | 16:01.54 |
Robin_Watts | henrys: So the thing to do is to checkout to before the change and do a clusterpush. | 16:02.13 |
| Then the report will show the timing differences. | 16:02.26 |
chrisl | The commit runs include timings, too | 16:02.28 |
Robin_Watts | Let me try that now. | 16:02.28 |
chrisl | The cluster numbers for the LCMS 2.6 commit are: http://pastebin.com/fNgUkFTk | 16:04.39 |
Robin_Watts | Thanks. I have been failing to find them. | 16:04.53 |
chrisl | In the cluster mail for the commit run | 16:05.21 |
Robin_Watts | So no meaningful differences there. | 16:05.30 |
chrisl | No, *well* below the threshold we can reliably detect | 16:05.52 |
| Oh, tiger.eps is naughty - it has a showpage.... | 16:08.22 |
henrys | maybe with just running tiger I'm just seeing lcms startup costs. cmsReverseToneCurveEx weighs in 2nd at 600 dpi after the expected mem_true24_fill_rectangle | 16:09.06 |
Robin_Watts | henrys: Profile before and after the addition of lcms. | 16:16.52 |
| It seems likely that it is one off startup costs though. | 16:17.06 |
mvrhel_laptop | lcms start up costs should be independent of rendering resolution certainly | 16:38.32 |
Robin_Watts | mvrhel_laptop: You should be able to look at GhostDocs on the twiki too now. | 16:39.31 |
mvrhel_laptop | robin_watts: someone needs to add in the info for building on windoze... ;) | 16:40.59 |
Robin_Watts | mvrhel_laptop: Yes. | 16:41.21 |
mvrhel_laptop | robin_watts this will be a useful thing to have to get started | 16:41.23 |
| very nice | 16:41.30 |
Robin_Watts | I've got an epage solution for VS2005 here. | 16:41.40 |
mvrhel_laptop | cool | 16:41.45 |
| i am not quite ready yet. soon though. | 16:42.00 |
Robin_Watts | It's like the gs one in that it uses the underlying make system rather than allowing the solution to build for itself. | 16:42.27 |
| but it does get stuff like "Goto definition" working. | 16:42.46 |
| I need to tweak it a bit more though. | 16:42.59 |
mvrhel_laptop | ok. that is fine. | 16:43.05 |
| I can live with that | 16:43.26 |
Robin_Watts | paulgardiner_, Jogux: OK, I've pushed a rejigged history to robin/master. | 16:49.07 |
| oh, ass, no I haven't. I need to reclone on casper, sorry. | 16:49.42 |
mvrhel_laptop | robin_watts: quick question for you | 16:57.25 |
| in fz_text_page_s there is a next member of type fz_text_page* . The comment on this structure says that the text pages is a list of page blocsk with a bounding box. why would a fz_text_page have a next entry? or is this for the next page? page? | 16:59.05 |
Robin_Watts | mvrhel_laptop: Yes, next page, I think. | 16:59.57 |
mvrhel_laptop | hmm curious how they get linked together. I will dig further | 17:01.08 |
Robin_Watts | paulgardiner_, Jogu: OK. rejigged history pushed. | 17:02.21 |
mvrhel_laptop | oh I see in fz_text_begin_page | 17:03.39 |
| so I could do a span of a few pages that are currently visible | 17:04.04 |
| I suspect the best way to do this is to have my application just carry around an opaque pointer that is the text_device | 17:06.08 |
Robin_Watts | mvrhel_laptop: I don't follow. | 17:06.39 |
mvrhel_laptop | robin_watts if I want to extract out the text location etc I run a particular page through the text device | 17:07.37 |
| and extract what I find from the search, and delete the device | 17:08.27 |
| lets say that I have a page displayed and I scroll 1/2 to the next page | 17:09.13 |
| but I want to select some text across the two pages | 17:09.29 |
Jogux | robin_watts : re picselness - to build the windows tgv emulator, you need directx 2006 sdk and the platform sdk (for winhttp.h); I gave up trying to install them from clean onto win7. | 17:10.57 |
Robin_Watts | mvrhel_laptop: Possibly you could run the text extraction device once for each page, and keep the appropriate text_page with each page. | 17:11.38 |
mvrhel_laptop | robin_watts true. I guess that may make more sense | 17:12.07 |
| I could do that on a back ground thread during the initial rendering | 17:12.29 |
Robin_Watts | Then when you drag a box to highlight/cut/paste etc, you can find the boxes for the chars etc without having to repeatedly rerun the extraction. | 17:12.30 |
Jogux | huh, colloquy seems to muck up the irc logs - ' seems to come out as âÃâ¬Ã⢠| 17:12.41 |
Robin_Watts | mvrhel_laptop: I might be tempted to do it "just in time". | 17:12.47 |
| i.e. the first time you drag on a page, do the page fetch then. | 17:13.06 |
mvrhel_laptop | robin_watts: ok. so do it on the span of pages that I also just rendered at full res | 17:13.10 |
| or for the screen display | 17:13.20 |
| right | 17:13.26 |
| ok. that makes sense | 17:13.34 |
Robin_Watts | 90% of the time people are not going to be doing text extraction, so why burden the machine with doing it until it's needed? | 17:14.11 |
mvrhel_laptop | right | 17:14.16 |
Robin_Watts | It's not a slow process, it's just not a process you want to be doing many times a second. | 17:14.27 |
Jogux | robin_watts : what's different in your rejigged history? | 17:14.42 |
Robin_Watts | I squashed several commits together, so rather than having "do X", "fix typo in X", "fix other typo in X" etc, we have a neater history. | 17:15.23 |
Jogux | ahha, neat. | 17:15.36 |
| was that an implicit request for me to review? | 17:15.44 |
Robin_Watts | Please. | 17:15.50 |
| I am trying to find ways to improve the VS solution at the moment. | 17:16.18 |
| Currently it just builds 1 thing. | 17:16.25 |
| and the location of the GENROOT is hardwired into it. | 17:16.40 |
Jogux | I was pondering if we can make the FBS generate a native build file. | 17:17.06 |
Robin_Watts | We could add new projects to the solution, one for each target. | 17:17.10 |
Jogux | the iOS situation is borked, you can set breakpoints in the core code as they don't trigger, but if you step down it finds the source fine. | 17:17.36 |
| s/as they/but they/ | 17:17.44 |
Robin_Watts | Or we could just keep the 1 project and add lots of configurations. | 17:18.06 |
| I think the multi project solution is nicer personally. | 17:18.27 |
Jogux | what do you think of making the FBS generate VS solutions? | 17:18.39 |
Robin_Watts | If you can make it work, I'd be all for it. | 17:18.51 |
| The problem comes in that people have to learn not to change the solution itself, but to change the build system and regenerate. | 17:19.36 |
Jogux | yes, indeed. albeit we'd probably not checkin the solution, so that should be relatively obvious | 17:20.08 |
Robin_Watts | otherwise we end up maintaining 2 different build systems, and that way lies madness. | 17:20.09 |
Jogux | yes, exactly | 17:20.22 |
Robin_Watts | Having a native solution (or at least a makefile we can call make or nmake on) would get us parallel builds. | 17:20.46 |
| Do I understand from what you said earlier that VS2008 wouldn't require extra stuff to be installed? | 17:21.27 |
Jogux | I wouldn't guarantee that :-) | 17:21.27 |
Robin_Watts | you wouldn't guarantee parallel builds? | 17:21.54 |
Jogux | robin_Watts: my testing so far relates to vs2012 and vs2013; vs2008 possibly won't but I can't guarantee. | 17:21.54 |
| I wouldn't guarantee parallel builds | 17:22.04 |
Robin_Watts | We have a company license for 2008, I think. | 17:22.11 |
Jogux | my experience with 2012 and 2013 both relate to the free version | 17:22.34 |
Robin_Watts | AIUI, 2012 needs W8 to work ? | 17:22.42 |
Jogux | 2013 requires W7 afaik | 17:22.53 |
| 2012, efven | 17:22.56 |
| yeah, 2012 is fine on Win7SP1 | 17:23.19 |
| 2010 works back to XPSP3 | 17:23.42 |
Robin_Watts | has tried hard to keep VS2005 as the standard for gs and mupdf as that keeps compatability back to XP (no service packs) and possibly earlier. | 17:24.02 |
| but for epage that's less critical. | 17:24.14 |
Jogux | let's hope so :) | 17:24.33 |
| tbf I think very few windows builds of epage are shipped just now. | 17:24.42 |
| (excluding winrt) | 17:24.46 |
| it's mainly used for debugging/dev. | 17:25.04 |
Robin_Watts | Well, the customer that just got back in touch had a win32 and RVCT builds. | 17:25.11 |
| I'm assuming that the win32 was for their development guys to use/debug. | 17:25.24 |
Jogux | interesting. I hope so :) | 17:25.55 |
| did you find rvct stuff or do you want me to prod at that? | 17:26.12 |
Robin_Watts | I didn't find it. I assumed it would be part of the ATS work you're doing? | 17:26.29 |
Jogux | sadly not; afaik the ATS vms aren't setup with RVCT. whether we want to is I guess another question. | 17:26.52 |
Robin_Watts | Jogux: I think we do want to set them up. | 17:27.29 |
Jogux | although I'm fairly certain the answer to that question is "yes" | 17:27.32 |
| mm :) | 17:27.34 |
Robin_Watts | I'm not sure how the licenses would work otherwise. | 17:27.51 |
Jogux | there are both node locked and floating licenses. | 17:28.03 |
| you /can/ get a floating license via ssh port forwarding, so long as you have a copious supply of tea. | 17:28.27 |
| do we know where picsel.com email goes? | 17:29.01 |
Robin_Watts | Jogux: ISTR doing that for ARM ADS builds at some point. | 17:29.05 |
| Jogux: Ask mace :) | 17:29.13 |
Jogux | ah, righto :) | 17:29.25 |
| if we're lucky, we'll be able to find the ads license server logins, and rehost the node locked licenses | 17:30.00 |
Robin_Watts | Jogux: Are you still in contact with Malcolm S ? | 17:30.31 |
| He might be able to tell us where to look on the servers for stuff like this. | 17:30.45 |
Jogux | possibly; I think Alex knows. Also I found a potentially useful spreadsheet, but it's password protected. | 17:31.03 |
| and I've forgotten the password (or someone has changed it since I knew) :( | 17:31.25 |
kens | Goodnight folks | 17:38.52 |
logikos | hey kens cant leave | 17:40.14 |
| he was supposed to figure out the error... | 17:40.24 |
mvrhel_laptop | bbiab | 17:41.13 |
chrisl | logikos: he'll be in touch tomorrow - I distracted him by phoning him | 17:41.30 |
Jogux | robin_watts : where is your sot.git? | 17:53.08 |
Robin_Watts | /home/robin/sauce/sot.git | 17:53.43 |
| (obviously! :) ) | 17:54.07 |
logikos | chrisl: alright, i have another question though, right now i'm using this command: gs -sstdout=/dev/null -sstderr=/dev/null -q -sDEVICE=ps2write -sPAPERSIZE=letter -dNORANGEPAGESIZE -o"|rlpr -h -H10.0.6.2 -Praw" /opt/gs/settings.ps ./file1.pdf ./file2.pdf | 17:56.32 |
| is it possible to scale file2.pdf because right now its margins are to tight | 17:56.56 |
| and the printer is cuting off part of the document | 17:57.04 |
chrisl | logikos: you could try adding to the command line: -dFitPages=true -dSetPageSize=false | 17:59.32 |
Robin_Watts | Jogux: So, I have a plan! | 17:59.54 |
| Jogux: Currently, when using the solution, it does the build by calling scripts/tgvbuild.py .... or whatever. | 18:00.28 |
| I plan to make it call: scripts/solution_build.py scripts/tgvbuild.py ... | 18:00.59 |
| scripts/solution_build.py can then have logic in to set GEN_ROOT etc based upon it's position. | 18:01.31 |
| So we can get it to automatically put genroot in the same directory as epage and resources | 18:01.56 |
| Also, I want to make scripts/solution_build.py look for local_setup.py, and if it exists, run it. So we can all apply local customisations to the build setup there if we want. | 18:03.01 |
Jogux | that all sounds feasible I think. | 18:03.50 |
logikos | chrisl: tried, no change | 18:03.57 |
Robin_Watts | logikos: Are you really using letter paper? | 18:04.32 |
logikos | normal 8.5 by 11" | 18:06.27 |
| thats letter right? | 18:06.33 |
Robin_Watts | yeah. In europe we use A4. | 18:07.05 |
chrisl | logikos: okay, you could try -dPDFFitPage | 18:07.28 |
logikos | in addition to the other 2 you gave me or in place of? | 18:07.53 |
chrisl | Yes | 18:08.06 |
Robin_Watts | A4 is 8.27x11.7, so using letter when you mean a4 will result in squeezed margins. | 18:08.11 |
logikos | I varified, it is letter | 18:08.27 |
| i see your point though | 18:08.32 |
| i'm in the states | 18:08.35 |
| letter is very common here | 18:08.45 |
| chrisl i asked an either or question and you said yes, i'm assuming you just read the first part so i'm going with option 1 | 18:09.24 |
chrisl | Yes, sorry, I meant "in addition to...." | 18:09.57 |
chrisl | is very tired..... | 18:10.09 |
logikos | added -dPDFFitPage with no change | 18:11.24 |
chrisl | Okay, one more suggestion, then I'm out of ideas: try adding -dFIXEDMEDIA | 18:12.18 |
| logikos: I should have added that these options shouldn't be appended to the command line, they should all be before the call to settings.ps | 18:14.32 |
logikos | k, testing that, but i also have anohter question, right now i have a settings.ps file i'm adding to it for duplex printing, any idea on how i might add stapling to it also? | 18:15.05 |
| current test: gs -sstdout=/dev/null -sstderr=/dev/null -q -sDEVICE=ps2write -sPAPERSIZE=letter -dNORANGEPAGESIZE -dFitPages=true -dSetPageSize=false -dPDFFitPage -dFIXEDMEDIA -o"|rlpr -h -H10.0.6.2 -Praw" /opt/gs/settings.ps ./file1.pdf ./file2.pdf | 18:15.37 |
| no change | 18:16.13 |
| oh well on that, what about staples ? | 18:16.24 |
| i found a stackoverflow: http://stackoverflow.com/questions/18098070/inconsistency-between-postscript-encoding-when-injecting-staple-commands which sugests you add <</Staple 3 /StapleDetails <</Type 21 /Position (1PLU)>> >> systemdict /setpagedevice get exec | 18:16.55 |
| which i tried, it did not error but also did not staple | 18:17.04 |
chrisl | logikos: I'm not sure if stapling is a standard Postscript option, so it might be device dependent | 18:17.28 |
Robin_Watts | logikos: What does your current settings.ps look like ? | 18:17.49 |
logikos | http://paste.ubuntu.com/7164167/ | 18:19.17 |
| first line was for duplex (works) second line i hoped would add staple (not working) | 18:19.41 |
Jogux | robin_watts : not really sure what level of review is wanted. There's a few randomly commented out lines of code, and I'd have prefered to see more description than "add alternative font library with modified metrics" in that commit message, and/or commentary on why we're keeping the old ones around too. and I can't really figure out how much of the python build script is C&P and how much is necessary. (I suspect the line for win98 support may not b | 18:19.53 |
Robin_Watts | logikos: I might have tried: http://pastebin.com/6C8TWavG | 18:21.11 |
logikos | printer is hp color laserjet 4730mfp | 18:21.17 |
Robin_Watts | but my PS is largely guesswork. | 18:21.24 |
| Jogux: Be harsh :) | 18:21.33 |
logikos | will test it | 18:21.43 |
| ty | 18:21.45 |
Robin_Watts | I should fix those randomly commented out bits of code. | 18:21.58 |
| And most of the build script is cut and paste, but I thought I'd cut it down as much as I dared. | 18:22.33 |
| Possibly I should just overwrite the fonts with the new ones. | 18:23.29 |
logikos | robin_watts: printed: ERROR: rangecheck OFFENDING COMMAND: setpagedevice | 18:24.33 |
chrisl | logikos: do you know that the staple settings in your command are correct for your printer? | 18:25.14 |
Robin_Watts | logikos: I suspect my postscript string quoting is bad. chrisl may want to laugh at me. | 18:25.30 |
logikos | chrisl: no | 18:25.40 |
| and i'm having a hard time finding what commands would be correct for the printer | 18:25.52 |
chrisl | Robin_Watts: I glanced at it, it looked okay - otherwise I'd have said | 18:26.13 |
| logikos: you need to find a PPD for the printer | 18:26.21 |
Robin_Watts | logikos: Ah, right, sorry, didn't see "rangecheck" in that. | 18:26.45 |
logikos | the printer has a ppd built in i'm told | 18:26.46 |
chrisl | Printers don't have PPDs "built in" | 18:27.08 |
Robin_Watts | That means that in the setpagedevice call, one (or more) of the arguments in the the preceeding stuff is out of range. | 18:27.28 |
logikos | sorry was thinking LPD | 18:27.44 |
Robin_Watts | So either the /Staple X entry is wrong, or the /StapleDetails stuff is. | 18:27.59 |
| or both. | 18:28.04 |
chrisl | logikos: http://en.wikipedia.org/wiki/PostScript_Printer_Description | 18:30.34 |
Robin_Watts | AFAICT that stapler only has "on or off", so I'd try /Staple 1 and omit the /StapleDetails ? | 18:31.15 |
| but that's a shot in the dark. | 18:31.28 |
chrisl | Probably better /Staple 2 and omit the /StapleDetails | 18:32.15 |
logikos | so your sugesting: << /PSDocOptions (<< /Duplex true /Staple 2 <</Type 21 /Position \(1PLU\) >> >> setpagedevice) >> setdistillerparams | 18:32.36 |
| ? | 18:32.37 |
chrisl | "2" is Staple at the end of the job. | 18:32.39 |
Robin_Watts | logikos: No. | 18:33.00 |
chrisl | logikos: you need to remove the StapleDetails value, too | 18:33.10 |
Robin_Watts | Within a << >> you have pairs of entries. A key and a value. | 18:33.14 |
| /Staple is a key, 2 is it's value. | 18:33.26 |
| So if you remove the /StapleDetails key you must remove its value too. (<< /Type 21 ...>> | 18:33.54 |
logikos | testing with: << /PSDocOptions (<< /Duplex true /Staple 2 >> setpagedevice) >> setdistillerparams | 18:36.29 |
| didnt work | 18:37.44 |
chrisl | logikos: what's the printer make/model? | 18:37.48 |
logikos | hp color laserjet 4730mfp | 18:38.22 |
Robin_Watts | /Staple 2 /StapleDetails << /Type 8 /Portrait true /StapleOption (ONEANGLED) >> | 18:38.37 |
| That's for an LJ9000. | 18:38.44 |
Jogux | robin_watts , paulgardiner : so... Paul should rebase his tree to remove Robin's changes then push to golden? | 18:41.39 |
Robin_Watts | Jogux: yes. | 18:42.30 |
chrisl | logikos: like I said, you need to find a PPD for your printer, and that will (or should!) contain the Postscript runes to enable features like staple | 18:46.27 |
logikos | thanks | 18:49.15 |
chrisl | Okay, I'm taking off the the night - and I'm off tomorrow... back on Monday. | 18:52.48 |
Jogux | chrisl : have a good weekend :) | 18:53.14 |
chrisl | thanks | 18:53.25 |
Jogux | robin_watts : boggle. your changes break ios builds. just trying to figure out why :( | 18:56.02 |
| robin_watts : ahha - ccaf300f353c381570097142159c9c0a2dd74487 - the #endif is too early; I believe it should be after # endif /* _STDIO_H */ | 19:01.25 |
Robin_Watts | Jogux: Try a fix and throw it at me, and I'll merge it into my original? | 19:02.20 |
Jogux | dealing with this many git repositories is seriously testing my git foo :) | 19:06.13 |
Robin_Watts | Jogux: Sorry. | 19:06.32 |
| The way I have it here is to have my casper repo as "origin", the golden one as "golden", pauls as "paul" etc. | 19:07.13 |
Jogux | paulgardiner_lap : can we rebase your sot git tree without the changes in Robin's git tree (which I reckon will leave us with just stuff that has been reviewed) and push them to golden please? :) | 19:07.27 |
Robin_Watts | That way when I push to origin, I don't inadvertently push to golden :) | 19:07.32 |
| I am tempted to move the epage solution from 'epage/epageoffice' to just being 'epagesolution' or something. | 19:09.15 |
henrys | ms office for iPad ⦠a bit late but there it is | 19:09.22 |
Jogux | robin_watts : I reckon there's a fix on my private-repos master. | 19:09.45 |
Robin_Watts | Jogux: Ta. | 19:09.54 |
Jogux | I think I should've just pastebinned a patch, that would've been way easier :-) | 19:10.21 |
Robin_Watts | I've got to go move a harpsichord. I will be back in a bit to try to address some of the review issues and take on your patch. | 19:11.12 |
Jogux | enjoy :-) | 19:11.23 |
Robin_Watts | So, you reckon we should just replace the fonts with the new ones? | 19:11.40 |
Jogux | I think I only got as far as "I don't understand why we have two sets of fonts". | 19:12.02 |
Robin_Watts | Cos I was afraid I'd bugger 'em up :) | 19:12.26 |
Jogux | or "why do we still have the old fonts in the tree if we're not using them anymore" :) | 19:12.34 |
Robin_Watts | but we can rescue them from git if we need them. Probably cleaner to just have the one set. | 19:12.47 |
Jogux | nods. | 19:12.57 |
Robin_Watts | That will require more git massaging than I will have time for tonight, so that will have to wait til tomorrow. | 19:13.04 |
Jogux | I think it is worth documenting how the new fonts came into existance somewhere too | 19:13.15 |
| whether commit message or wiki page or whatever | 19:13.30 |
Robin_Watts | yeah. | 19:13.44 |
| I have a horrible tool that I used to rewrite them. I might commit that to resources. | 19:14.10 |
Jogux | hehe | 19:14.16 |
| (that sounds reasonable to me) | 19:14.57 |
logikos | I found the docs for my printer and it says to enable the staple feature i should put this in the PJL header at the beginning of the job: @@PJL SET FINISH=STAPLE | 20:05.26 |
| any clue how i do that by chance? | 20:05.38 |
Robin_Watts | You could try putting exactly that at the start of the settings.ps file. | 20:06.00 |
logikos | i tried adding /Finish STAPLE to the post script just for kicks and it didnt work lol | 20:06.01 |
| k, i'll give it a try | 20:06.12 |
| GPL Ghostscript 9.14: Unrecoverable error, exit code 1 | 20:07.06 |
Robin_Watts | logikos: Oh, D'oh, sorry. | 20:07.29 |
| yeah, that won't work :) | 20:07.35 |
logikos | http://paste.ubuntu.com/7164708/ | 20:07.38 |
Robin_Watts | You could probably fix it at the pipe stage. | 20:07.46 |
| Where you have | blah, you could probably do: | cat settings.pjl - | blah | 20:08.24 |
| and then put the PJL magic in settings.pjl | 20:08.35 |
| assuming my memory of cat is correct. | 20:08.58 |
logikos | so you think i can do -o"| cat settings.pjl | rlpr -h -H10.0.6.2 -Praw" | 20:11.40 |
| ? | 20:11.41 |
| i dont think cat will work like that... | 20:11.52 |
Robin_Watts | logikos: No, you missed a - | 20:12.08 |
| | cat settings.pjl - | | 20:12.17 |
| That should (if I remember correctly send the contents of settings.pjl, then copy stdin to stdout. | 20:12.39 |
logikos | yeah i tested cat settings.ps | cat settings.pjl - | 20:13.54 |
| seemed to work as your expecting | 20:13.59 |
| doing -o"|cat /opt/gs/settings.pjl - |rlpr -h -H10.0.6.2 -Praw" caused the printer to spit out a blank page | 20:15.11 |
| no error, no document... | 20:15.16 |
| these are the docs and an example: http://paste.ubuntu.com/7164747/ | 20:16.57 |
| perhaps i need the other stuff arround the @PJL stuff | 20:17.12 |
Robin_Watts | maybe. I'm not a PJL expert. | 20:18.02 |
| You might need some PJL to put it into PS mode? | 20:18.13 |
logikos | i know nothing, i'm a php developer who was told to make a web based app that can print from the server | 20:18.36 |
| lol | 20:18.37 |
Jogux | :-) | 20:18.49 |
logikos | so i'm writing shell scripts and having php execute the shell scripts | 20:19.04 |
Robin_Watts | logikos: Hmm. You do know that gs 9.10 and above are AGPL rather than GPL licensed? | 20:19.07 |
Jogux | logikos : http://h20565.www2.hp.com/portal/site/hpsc/template.PAGE/public/kb/docDisplay/?spf_p.tpst=kbDocDisplay&spf_p.prp_kbDocDisplay=wsrp-navigationalState%3DdocId%253Demr_na-bpl01378-2%257CdocLocale%253D%257CcalledBy%253D&javax.portlet.begCacheTok=com.vignette.cachetoken&javax.portlet.endCacheTok=com.vignette.cachetoken has an example at the bottom | 20:19.42 |
logikos | robin_watts, nope didnt know that | 20:19.59 |
Robin_Watts | AGPL has an impact for web server setups. | 20:20.48 |
Jogux | albeit AGPL isn't a problem if the web app is internal to his company aiui. | 20:21.01 |
logikos | the server is on our intranet | 20:21.30 |
| its not public | 20:21.32 |
Robin_Watts | Then no problem, AIUI. | 20:21.40 |
logikos | good to know though | 20:21.48 |
| ty for the info | 20:21.50 |
| also the server version if 9.06 | 20:22.06 |
| although i did add 9.14 to another location to see if it would solve some of the issues i was having | 20:22.23 |
Robin_Watts | I thought we needed 9.14 for the setpagedevice stuff? | 20:22.56 |
| I might be wrong. | 20:22.58 |
logikos | we needed it for something | 20:23.35 |
| right now i'm useing the 9.14 version to send to the new printers for duplex | 20:23.45 |
| and the 9.06 version for the old printers becaues the old printers dont like ps2write | 20:23.58 |
| they are ok with pswrite | 20:24.04 |
Robin_Watts | Jogux: New version of commits on robin/master. I've rebased the solution and the fonts to the end. Everything else should be good to go, I think. | 20:38.50 |
| If you spot anything I've still cocked up, let me know. | 20:39.15 |
| actually... hold on... | 20:39.29 |
| There now. | 20:39.56 |
| I'll get the solution and the fonts in tomorrow. | 20:40.32 |
Jogux | okay, I'll try and check; might be tomorrow now, I'm in the middle of ios 7 fixes. | 20:41.07 |
logikos | Jogux: the example on that page has stuff above and bellow where they say the postscript should go | 20:49.34 |
| if i start with PJL do i have to end with it too? | 20:49.55 |
| testing with this in settings.pjl: http://paste.ubuntu.com/7164912/ | 20:51.05 |
Jogux | logikos : I'm afraid my knowledge of this area is a strict subset of that webpage | 21:01.17 |
| gahh. I'm relying on the stuff in 4 billion branches for my work to build. or at least that's what it feels like. | 21:22.11 |
Jogux | decides to give up for tonight (and twist Paul's arm into pushing his stuff to golden tomorrow) | 21:23.27 |
| Forward 1 day (to 2014/03/28)>>> | |