| | <<<Back 1 day (to 2014/04/29) | 2014/04/30 |
| marcosw | ray_laptop: for the logs - the commit warnings check is run on a different version of gcc than the cluster nodes use. This is quasi-intentional (it just worked out that way but I decided it wasn't a bad thing). | 01:31.38 |
| ave_ | hello to all | 11:01.40 |
| Robin_Watts | ave | 11:01.46 |
| | </latin gag> | 11:01.59 |
| ave_ | i've got a problem with gs consolle and pdf extraction | 11:02.03 |
| | :) | 11:02.05 |
| | yes a latin gag :) i know mr. Caesar | 11:02.19 |
| Robin_Watts | making a pdf from other input? Or extracting "something" from a PDF ? | 11:03.05 |
| ave_ | i'm using -dFirstPage=1 -dLastPage=1 and gs always process and extract all pages of source | 11:03.07 |
| | i've a pdf | 11:03.15 |
| Robin_Watts | ave_: Can you give the full command line please? | 11:03.24 |
| ave_ | and i want to extract only the first page | 11:03.27 |
| Robin_Watts | You are putting -dFirstPage and -dLastPage before the filename, right? :) | 11:03.39 |
| ave_ | yes here: gswin64c.exe -dSAFER -dBATCH -dNOPAUSE -dLastPage=1 -o pdf_conversion -sDEVICE=pdfwrite c:\temp\spool.prn | 11:04.18 |
| | or gswin64c.exe -dSAFER -dBATCH -dNOPAUSE -dFirstPage=1 -dLastPage=1 -o pdf_conversion -sDEVICE=pdfwrite c:\temp\spool.prn | 11:04.36 |
| Robin_Watts | What version of gs ? | 11:04.45 |
| chrisl | FirstPage and LastPage don't work with Postscript | 11:04.49 |
| ave_ | ohm damn :( | 11:04.59 |
| | i don't knew this | 11:05.28 |
| chrisl | Postscript is a stream, not a random access file format, so the params really don't make sense for it | 11:05.35 |
| Robin_Watts | ave: So... you can pdfwrite the whole file, then use pdfwrite again with -dFirstPage -dLastPage to write out just the page you want? | 11:05.35 |
| | chrisl: I believe they work for PCL. We just ignore early/late pages. | 11:05.55 |
| ave_ | yes Robin, i can do that.... the only matter is that i'm processing huge files :( | 11:06.09 |
| Robin_Watts | So it could work for postscript in the same way. | 11:06.13 |
| chrisl | Robin_Watts: I'm not sure that's true any more, I thought henrys killed that | 11:06.31 |
| Robin_Watts | ave_: Even if we extended gs to work in the way I described, it would still need to process all the pages. | 11:06.40 |
| | chrisl: Can we do pdfwrite with -o out%d.pdf ? | 11:06.59 |
| chrisl | Yes | 11:07.05 |
| ave_ | Yes Robin | 11:07.08 |
| Robin_Watts | ave_: So, you could do: | 11:07.18 |
| | gswin64c.exe -o out.%d.pdf -sDEVICE=pdfwrite input.ps | 11:07.36 |
| | and then you'd get each page out as a separate pdf. | 11:07.43 |
| | Would that help ? | 11:07.50 |
| chrisl | You can also achieve it using an EndPage procedure in Postscript *but* we'd still end up interpreting the entire file | 11:08.13 |
| ave_ | yes that help Robin | 11:08.43 |
| | thanks a lot for the support | 11:08.50 |
| Robin_Watts | No worries. | 11:08.55 |
| ave_ | now in italy we go to have lunch ;) best regards and thanks a lot | 11:09.10 |
| Robin_Watts | Here in the UK, I'm going to have a shower. | 11:09.25 |
| ave_ | Yo man ;) | 11:09.47 |
| henrys | chrisl: first/last page does work in pcl, there was some good reason it can't be added to postscript and I'll be damned if I can remember what is was. | 11:38.43 |
| chrisl | henrys: Ah, I thought when a customer was using it with PCL you told them not to, and decided to remove it - I guess not | 11:39.54 |
| | As I said, it's not hard to implement in Postscript, it just doesn't save any time - we still have to interpret the whole file | 11:40.34 |
| henrys | chrisl: yes it is slow in PCL too, but I thought there was something else in PS, maybe not | 11:41.52 |
| chrisl | henrys: probably because it's straight forward to make it happen in PS without changing GS, and really doesn't have much benefit - you can't just stop interpreting Postscript | 11:43.56 |
| kens | THere's no reason we couldn't do it in PostScript, as chris rightly says the mechanism (/EndPage) is already there, but it saves nothing. | 11:43.56 |
| chrisl | Of course for the vast majority of users, there would be no problem in leaving the interpreter in an unknown state, but if anyone actually uses the server loop, then.... | 11:44.53 |
| kens | Actually, you could have /EndPage throw an error when it finishes processing the last page, that might be a bit quicker | 11:47.24 |
| chrisl | Wouldn't that interrupt the showpage? | 11:48.04 |
| kens | Wel, I meant the endpage *following* the last completed desired page | 11:48.21 |
| | So you would have to interpret on extra page | 11:48.38 |
| chrisl | You could also have a StartPage which threw an error after the last page | 11:49.19 |
| kens | Yes I guess that would work too. But you would have to not miind the fact that the job throws an error | 11:49.49 |
| chrisl | I think that would confuse a lot of users... | 11:51.21 |
| kens | Yep, which is why I would hesitate to suggest it | 11:51.38 |
| Robin_Watts | spins the git pull --rebase roulette wheel and prepares the jogu voodoo doll... | 11:51.43 |
| | boggle. no conflicts. | 11:53.13 |
| jogux_mac | guess iOS SOL maybe isn't popular then :-) | 11:54.12 |
| Robin_Watts | I didn't touch any of that, fortunately :) | 11:54.42 |
| | So I'm going to have a crach at pulling in the missing commits based on that list you made, assuming I can get into the review system to see them. | 11:55.12 |
| jogux_mac | well... I can't, so I'm not sure you will have any more luck :-) | 11:56.42 |
| henrys | jogux_mac: Is Miah still working on the test stuff or are you solo now? | 11:57.45 |
| jogux_mac | henrys : just me afaik - Miah hasn't popped up for a while | 11:58.33 |
| kens | Oh dear, I accidentally interrupted a cluster push, and it look slike all my code has gone pop | 12:00.06 |
| henrys | jogux_mac: miles is looking for another web resource, he just hasn't been responsive enough. It's probably best for Miles to get someone local. I assume someone in the SF Bay area knows how to set up web pages ;-) | 12:00.45 |
| kens | Ah, its OK it was in the stash | 12:00.56 |
| jogux_mac | henrys : fair enough. :-S I'm not entirely sure if we have logins to get at the current content btw... | 12:01.39 |
| chrisl | henrys: yeh, we've got a minor issue with a page on the artifex site, and no reply about how to resolve it :-( | 12:01.48 |
| jogux_mac | henrys : we're happy enough making simple tweaks if we can get hold of the logins if that helps. | 12:03.01 |
| henrys | jogux_mac: yes miles is trying to collect the login stuff. | 12:03.02 |
| jogux_mac | henrys : ah, great. | 12:04.34 |
| chrisl | kens: there's a pdf interpreter patch for your perusal for http://bugs.ghostscript.com/show_bug.cgi?id=695194 | 12:05.55 |
| kens | chrisl I saw it, I don't see a problem but I have the interpreter in pieces right now | 12:06.12 |
| | doign teh special op thing | 12:06.22 |
| chrisl | You don't see a problem with the input file, or with my change? | 12:06.47 |
| kens | your change, the file is clearly broken | 12:07.12 |
| chrisl | It passes the cluster, and seems pretty safe to me.... | 12:07.34 |
| henrys | chrisl: what's the web page issue | 12:07.54 |
| | ? | 12:07.55 |
| | P1 bug assigned to miles ;-) | 12:08.11 |
| chrisl | henrys: the license page still references and links to the GPL, instead of AGPL | 12:08.17 |
| kens | chrisl, yes, but like I said, I have tehat file in pieces being stitched back together, it shouldn't take long <famous last words> | 12:08.27 |
| Robin_Watts | chrisl: Can we not change the artifex website ourselves? | 12:08.45 |
| chrisl | Robin_Watts: we can, but neither Miles nor I could find where that particular text is defined | 12:09.24 |
| Robin_Watts | ah! :) | 12:09.37 |
| chrisl | http://www.artifex.com/page/licensing-information.html | 12:09.51 |
| | The box on the left - the main body text is easily found and edited, but the sidebars... I have no idea :-( | 12:10.39 |
| Robin_Watts | I can't remember how to get to the editing bit :( | 12:11.28 |
| henrys | I don't understand - I'm looking in developer view in chrome and see <p class="gpllink"><a href="http://www.gnu.org/licenses/gpl.txt" target="_blank">GPL License</a></p> | 12:14.08 |
| | why can't that be changed? | 12:14.08 |
| chrisl | henrys: so, where do I go to edit that? | 12:14.35 |
| | And remember, artifex.com is now a content managed site, rather than "static" html | 12:15.19 |
| henrys | chrisl: oh this is generated | 12:15.52 |
| chrisl | Yes, each page is "assembled" from various parts - and I have no idea where that part is defined | 12:16.50 |
| Robin_Watts | chrisl: I see the problem :) | 12:17.57 |
| | In theory it means that Miles/Scott can make changes without needing to speak html. (add press releases etc) | 12:18.23 |
| chrisl | Which, indeed, Miles had done at least once.... | 12:18.53 |
| henrys | In practice it means nobody can do anything ;-) The data has been jailed | 12:18.56 |
| Robin_Watts | henrys: No, most stuff is dead easy to edit. | 12:19.22 |
| jogux_mac | from a quick look, that licensing thing is a "block" in quickapps speak | 12:19.33 |
| | so I'd go hunting within the blocks administration page | 12:19.56 |
| chrisl | jogux_mac: that was my thought, too, but there's nothing editable in the blocks page | 12:20.15 |
| jogux_mac | oh :-( | 12:20.22 |
| chrisl | Whether you need some kind of super-admin rights to do that, or something, I don't know | 12:20.44 |
| jogux_mac | that's possible I guess | 12:20.55 |
| chrisl | Even so, only three blocks are listed: "Artifex Front End Theme", "Admin Default" and "Unassigned or Disabled" and nothing that mentions licensing | 12:21.45 |
| jogux_mac | I think my next step would to get a shell login and start grepping/searching the mysql db :-) | 12:25.11 |
| chrisl | I don't know where artifex.com is hosted | 12:25.40 |
| jogux_mac | godaddy apparently | 12:25.49 |
| | (going by whois) | 12:26.34 |
| Robin_Watts | chrisl: We have 2 godaddy accounts I believe. | 12:26.39 |
| | One has details in ~robin/DownloadsInformation.txt | 12:26.56 |
| chrisl | Ah, so not the downloads.ghostscript.com account? | 12:26.59 |
| Robin_Watts | yes, apparently it's in there. | 12:27.10 |
| | In case I ever hold the shotgun backwards or something, I try to put important stuff in text files in my home dir :) | 12:27.57 |
| chrisl | Ah, I thought it ended up somewhere else for speed reasons | 12:28.04 |
| Robin_Watts | chrisl: Hmm. I don't know. | 12:28.25 |
| chrisl | And it doesn't look like I have shell access to the account, either :-( | 12:29.16 |
| Robin_Watts | chrisl: Not sure any of us do. But we can probably backup the data and look? | 12:29.47 |
| chrisl | I'll have a poke around wit ftp | 12:30.06 |
| Robin_Watts | If it's a mySQL database you'll not get to that by ftp. | 12:30.24 |
| jogux_mac | worse case write a quick php script that calls mysqldump :-) | 12:31.19 |
| chrisl | The files/directories under "artifex" don't seem to have been updated this year, but that doesn't necessarily mean anything | 12:32.19 |
| henrys | chrisl: miles did request login information for picsel and artifex suggesting we are missing something, but I really haven't kept up with that stuff | 12:33.45 |
| jogux_mac | I believe picsel is being hosted on miah's hosting. | 12:34.45 |
| Robin_Watts | I thought picsel was on the same hosting it ever was - just that miah had paid for it on his ccard. | 12:35.23 |
| jogux_mac | robin: I recall an email saying miah had moved it to his hosting | 12:35.46 |
| Robin_Watts | ok, you're probably more up to date than me. | 12:35.58 |
| jogux_mac | I believe it's on a bytemark vm now - I think it was on fasthosts or something when Picsel was in admin? | 12:36.10 |
| jhabjan | I think I found out what the problem is (from yesterday) with gs and pipes | 12:38.27 |
| | pipes works fine when client handle is passed from the other process | 12:38.51 |
| | but when the pipes are used within a same process, gs closes client handle... and I think it shouldn't | 12:39.38 |
| Robin_Watts | jogux_mac, pedro, paulgardiner_lap, henrys: Where did we land on pushing the branch->trunk reapps ? | 12:42.49 |
| | I'm inclined to push them now and we'll worry about ATS when it's working. | 12:43.09 |
| | otherwise we're into potential rebase hell. | 12:43.24 |
| jogux_mac | yes, I think don't delay for ATS | 12:43.26 |
| | I don't know what we need to do wrt to review | 12:43.36 |
| Robin_Watts | pedro looked them over the other day and seemed broadly happy I think. | 12:43.46 |
| | If anyone else wants to have a look, I'd be grateful. | 12:43.55 |
| jogux_mac | robin : I think I'm too deeply stacked into ATS/cvs/blah just look :-S | 12:46.37 |
| Robin_Watts | I understand. | 12:46.51 |
| chrisl | So, the only place I can find the text from that block is in "artifex/tmp/cache/cake_blocks_ca......" which sort of confirms it is a block, gives no clue about where it originates :-( | 13:00.31 |
| jogux_mac | it must be in the database then, surely? | 13:01.04 |
| chrisl | I would assume so, but it mustn't be stored as plain text, so grep doesn't find that | 13:01.44 |
| | Unless the database is stored somewhere else entirely..... | 13:04.22 |
| jogux_mac | chris: it'll be in /var/lib/mysql somewhere | 13:04.41 |
| chrisl | I can't get there from here.... | 13:05.13 |
| jogux_mac | yeah, not a surprise :( | 13:05.20 |
| chrisl | I rather thought in a setup like this, the site would be entirely within our use area | 13:05.48 |
| jogux_mac | I suspect godaddy are providing a db instance as part of the setup | 13:06.18 |
| henrys | chrisl: anyway miles is going to get a local resource so it's probably best left with that, we've plenty of other stuff to do, looks like norbert has some more files... | 13:08.11 |
| chrisl | henrys: yes, but if he's right, I think they should go in a new bug, and be yours to look at..... he says it's not text this time | 13:09.06 |
| henrys | chrisl: yeah I know it's mine to look at. | 13:10.35 |
| | chrisl: the bug subject still fits so I won't ask him to create a new one. | 13:11.14 |
| chrisl | henrys: so we'll have two commits referencing different fixes for the same bug? | 13:12.04 |
| | I thought we were trying to avoid that kind of thing | 13:12.19 |
| henrys | chrisl: yeah you're probably right, how about if there is action on this problem I'll make a new bug. | 13:13.05 |
| | ? | 13:13.09 |
| | chrisl: nvm I'll make a new bug, painless | 13:13.34 |
| chrisl | henrys: It's up to you - it's just it's the kind of thing that I get complaints about with the changelogs | 13:14.11 |
| henrys | chrisl: I meant to ask you yesterday about the language switch build but forgot. Anything others can help with? | 13:22.28 |
| chrisl | henrys: I'm still banging on getting the three executables building without recursive nmake calls. I have a concern though, that we can't build a graphics library shared between the interpreters because Ghostscript needs 64 bit colour indices, whilst PCL must use 32 bit ones | 13:25.00 |
| henrys | chrisl: oh so zoltan wasn't kidding, I missed that development, why is it no longer configurable | 13:26.12 |
| | ? | 13:26.15 |
| | chrisl: that can't be right it would too slow for ray_laptops primary concern. | 13:26.53 |
| chrisl | henrys: I think it's because we switched to using pre-defined arch.h files for the Windows bulids | 13:27.01 |
| | henrys: I put link to a suggested fix on the bug | 13:28.39 |
| henrys | you can soldier on with 64 though and we'll deal with that later. Are there other problems? | 13:31.15 |
| | PCL doesn't *require* 32 | 13:31.32 |
| chrisl | The romfs being tied to the graphics lib is the biggest remaining issue, I think | 13:32.11 |
| henrys | chrisl: isn't 532 32 bit? have to ask ray about that when he gets here. | 13:34.52 |
| chrisl | henrys: yes, 532 only use 32 bits | 13:35.07 |
| | But they don't use a "Windows build" exactly and they are only up to 9.06 | 13:36.08 |
| | henrys: after you mentioned multithreaded autogen.sh/configure the other day, I had the idea that we could easily run the libtiff configure script in parallel with ours - unfortunately, it only saves about 10 per cent on the time.... on a four core CPU | 13:47.43 |
| henrys | chrisl: good thing we don't have to run it frequently | 13:49.56 |
| chrisl | It's about 15 seconds, for me, so not bad, really | 13:50.19 |
| Proxios | Hello, I just received an email from a Ken at ghostscript | 14:29.16 |
| kens | THat's me | 14:29.20 |
| Proxios | Have you ever seen Ghostscript work in a Citrix environment? | 14:29.49 |
| kens | No, but it works in all kinds of other places. | 14:30.00 |
| | I don't have a Citrix server, so I can't test it myself | 14:30.11 |
| | I do see a bug report claiming slow performance using a Citrix server, so apparently someone has at least got it to work in the past | 14:30.55 |
| | http://ghostscript.com/pipermail/gs-bugs/2009-July/008348.html | 14:31.05 |
| chrisl | That was down to disk access, IIRC | 14:31.17 |
| kens | I think that's true yes | 14:31.30 |
| Proxios | ok that is fine. if we have a client interested in getting setup with 30+ licenses, where would they go to purchase | 14:31.30 |
| | or is this completely free? | 14:31.35 |
| kens | Ghostscript is licenced under the AGPL | 14:31.48 |
| | So if your/their usage is permitted under AGPL, then its free. | 14:32.09 |
| | (but unsupported) | 14:32.19 |
| | We don't normally sell licences to Ghostscript as such, its licenced on a royalty basis to people distributing it (eg printer manufacturers), but we do sometimes do 'support only' contracts | 14:33.32 |
| | You would have to talk to Scott about it (he's in Texas, so its failry early in his day yet) | 14:33.48 |
| Proxios | ok so you're saying for us to install this across our servers to deploy to the client | 14:34.03 |
| | it would come at no charge? | 14:34.11 |
| kens | I am not a lawyer, I cna't indemnify your usage | 14:34.20 |
| | If you are in doubt, you should seek professional legal advice | 14:34.32 |
| Proxios | regarding what? | 14:34.42 |
| kens | Whether your usage is compliant with the AGPL | 14:34.55 |
| Proxios | is there a way to tell? | 14:35.13 |
| kens | It sounds like it is, but firstly I'm not a lawyer and secondly I don;t know exactly what you are doing# | 14:35.14 |
| Proxios | they just want to use it as a pdf application, like adobe | 14:35.31 |
| kens | Reading the AGPL is a good place to start | 14:35.37 |
| Proxios | link? | 14:35.50 |
| kens | Umm, Google ? | 14:35.57 |
| | http://www.gnu.org/licenses/agpl-3.0.html | 14:36.13 |
| Proxios | Okay, I have never heard of that before, I thought it was something under your belt | 14:36.54 |
| | thanks | 14:36.55 |
| kens | No problem, its a variant of the Gnu GPL | 14:37.15 |
| | Like I said, I don't *think* you have a problem, but its not totally clear to me what you are doing, you really have to decide yourselves. | 14:37.57 |
| | If you are looking for a 'PDF pritner' you might do better working with one of our customers. | 14:38.29 |
| Proxios | they need a pdf software in order to get one of their other programs 'Made to Manage' working successfully | 14:39.36 |
| kens | Well, I haven't heard of that, but that's no great surprise.... | 14:40.00 |
| | Hmm some kind of ERP software. Presumably its 'PDF reporting' requires an additional bolt-on tpo convert the PostScript it prints into a PDF file. Naff but not unusual | 14:41.00 |
| Robin_Watts | Proxios: Essentially, AIUI, if people set up ghostscript unchanged on their servers, they don't have a problem. | 14:46.09 |
| | If people distribute servers to other people (i.e. images of software to run on servers) with unchanged ghostscript in them, they are fine, as long as the people receiving such distributions are free to update ghostscript themselves. | 14:47.22 |
| | If they make any alterations to ghostscript, or they link ghostscript into their own app, or they prevent people from updating ghostscript then the situation is more complex. | 14:48.55 |
| | (But IANAL) | 14:49.04 |
| Proxios | Ok guys thanks for the advise. | 14:55.55 |
| mvrhel_laptop | brb | 15:30.40 |
| | blah. I need to add in a data virtualization method into the wpf list view for gsview. otherwise things get a little wonky when we zoom in and have a large number of pages | 22:56.37 |
| | off to baseball game now. bbiaw | 22:56.48 |
| | Forward 1 day (to 2014/05/01)>>> | |
IRC Logs