| <<<Back 1 day (to 2015/10/29) | 20151030 |
mvrhel_laptop | done for the night | 00:50.51 |
maquefel | hi all | 08:33.06 |
| how to add new line in -c "(test text) show" | 08:33.52 |
kens | newline has no meaning in PostScript | 08:42.32 |
| You position the text where you want it using moveto to alter teh currentpoint, or translate to affect the CTM | 08:42.57 |
Robin_Watts | https://xkcd.com/1597/ | 10:19.52 |
kens | Yeah I was amused by today's one. I guess that makes you and Tor the 'friend' | 10:20.21 |
| We don't have a jpegscaled devcie do we ? | 10:34.41 |
| No we defintely don't | 10:35.06 |
Robin_Watts | We do not. | 10:35.08 |
| What the hell would be the point? | 10:35.15 |
kens | I agree, I'm answering the muppet in bug #696313 | 10:35.28 |
Robin_Watts | Render in lovely AA quality. Throw all the quality away when compressing to jpeg :) | 10:35.34 |
| yeah. | 10:35.39 |
kens | His problem is that his 'transparency' is not transparency at all its a dumb monster image drawn under the text, with simple on/off black pixels. At certain resolutions bits of it drop out. | 10:36.36 |
| If you then quantise it, reducing the high frequencies, the problem just gets worse | 10:37.07 |
| So its a dumb way to do a watermark combined with a dumb selectio of output format | 10:37.23 |
Robin_Watts | paulgardiner: See skype. | 10:38.18 |
paulgardiner | Yerp | 10:38.36 |
mjouk | Hi, is there a good reason why gs doesn't have an easy way to set arbitrary PJL commands in the output using, say, the pxlcolor device ? Otherwise, what about the specific case of @PJL SET FINISH=STAPLE ? | 14:49.01 |
henrys | mjouk: you mean setpagedevice stuff should be converted to appropriate PJL if the language (PCL, PXL) doesn't support the functionality. Great project let us know when you have something ;-) | 14:55.48 |
mjouk | I was thinking more of a command line option | 14:56.25 |
| I used gs to convert a bunch of pdf documents to pcl, and had to hack the pcl afterwards to add the staple stuff. It seemed a bit suboptimal | 14:57.06 |
henrys | mjouk: that would be out of character for PostScript, because it would skip error checking. You can do the setpagedevice command on the command line. We e do have some stuff that is supported for pxlcolor, I think Duplex or something like that. | 14:58.18 |
mjouk | Yes, you can set duplex but not the finish option. At least I couldn't see a way to do it. | 14:58.58 |
| It was a one off job, which I've done now, but I thought I'd see if there was any mileage in trying to make life easier for anyone who wants to do the same thing in future. | 14:59.48 |
henrys | mjouk: If you make a list of commands in a bug report (bugs.ghostscript.com) we will pay a "bounty" for somebody to do the enhancement. I know somebody that might be interested in doing it. | 15:03.20 |
| mjouk: I shoudn't say list of commands, I mean PJL that you want supported. | 15:04.13 |
mjouk | OK, I'll do that. I should say that I know the PJL which worked for me, but I don't consider myself any sort of expert in this so there might be better choices. DO you know anyone who might be able to offer wisdom in this area ? | 15:05.27 |
henrys | mjouk: we hope wisdom will be added to the bug report by whoever takes up the problem. | 15:06.52 |
mjouk | I've added bug 696314. Is the report OK ? | 15:24.13 |
henrys | mjouk: looks good I'll make it bountiable and copy in the guy that might be interested in doing it. thanks | 15:26.29 |
mjouk | Not at all. Thanks for your advice. | 15:26.55 |
chrisl | Really, prepending some text to a file is hard?? Jeez..... | 15:28.55 |
henrys | revitalizing my appreciation of math: https://www.youtube.com/watch?v=rudzYPHuewc | 15:29.16 |
| chrisl: don't you think we should have the error checking and everything of going through the interpreter? | 15:30.24 |
| chrisl: are you feeling better? | 15:31.17 |
chrisl | henrys: since PJL is just arbitrary text as far as the PS interpreter is concerned, we won't be doing much error checking..... | 15:31.41 |
| henrys: Yeh, thanks, just a minor, but persistent nagging headache the other day | 15:32.04 |
kens | henrys have you read the 'maths in The Simpsons' book ? I enjoyed it | 15:34.23 |
henrys | kens: I got half way through that and realized I hadn't been listening at all... | 15:34.59 |
kens | 487ed6d3b5fabbe21c23da288fbf020f49a28fae:-) | 15:35.16 |
henrys | no haven't see the simpson book | 15:35.24 |
kens | me wonderck the 'sent from my iPhone' text, I'd modify it to sent from my iThrone and wait to see if anyone noticed | 15:37.08 |
henrys | chrisl: I meant the device should properly support setpagedevice staple. I was thinking of making it bountiable for Hin-Tak. | 15:40.25 |
kens | So the PXL device should read the staple value from the page device dictionary and emit it as PJL at the head of the output ? | 15:40.54 |
henrys | chrisl: I do not want to do arbitrary crap from the command line | 15:40.56 |
kens | Duplex as well ? Media size, tray ? | 15:41.06 |
henrys | kens: it does most of those now | 15:41.18 |
kens | OH O | 15:41.25 |
| OK | 15:41.28 |
chrisl | Yeh, I suppose that makes sense | 15:41.31 |
| My reading of the chat was he wanted arbitrary PJL on a passthrough | 15:45.23 |
henrys | chrisl: that's what he wants, I'm fixing that now ;-) | 15:45.44 |
chrisl | :-) | 15:46.10 |
henrys | chrisl: since you typically edit the gs web page can you double the bounties when you get a chance? | 15:52.44 |
chrisl | henrys: sure, will do | 15:52.55 |
henrys | a just the person I wanted to badger rayjj | 15:53.36 |
| oh chrisl may know too, rayjj how does a pdf file find it's way to "gs_interpret" on a 532 printer? Is it spooled, given a temporary name? | 15:55.21 |
rayjj | henrys: it is spooled to disk | 15:55.51 |
| henrys: and, yes, it has a file name, so it gets read using the normal OS file functions | 15:56.34 |
chrisl | IIRC, it can also be held in ram? | 15:57.03 |
rayjj | chrisl: it is on the '6th gen' printer that doesn't have a disk | 15:58.55 |
henrys | rayjj: I was just thinking ahead with the language switch API, should we provide that spooling or leave it to the customer, etc.. | 15:59.02 |
kens | optional sounds best | 15:59.19 |
rayjj | henrys: if the input is not seekable, then we should spool it | 15:59.30 |
kens | we could do it to a RAM file system and tehn spool to disk if it exceeds a threshold for example | 15:59.40 |
chrisl | We already do that, don't we? | 15:59.40 |
kens | I can see that a big document conversion server might well want to give us a filename though and not have it copied again | 16:00.25 |
rayjj | henrys: but that's already in the PDF interpreter (or at least used to be there) -- it creates a Reusable stream | 16:00.29 |
| henrys: but, yes, we should not spool it if the source is seekable (i.e. a file) | 16:01.12 |
henrys | with 532 does then use .runlibfile on this spooled file? | 16:02.24 |
| with 532 does it use .runlibfile on this spooled file? | 16:02.35 |
chrisl | henrys: I've updated amounts on the bug bounty page, and also added a bit about reserving the right to evaluate on a case by case basis - if you want to check it for sanity | 16:02.58 |
henrys | chrisl: lgtm | 16:04.23 |
rayjj | hmm... the PDF interpreter isn't very smart. It just checks if the file is from stdin, and if so creates a tempfile (not ReusableStreamDecode) and runs it | 16:04.26 |
chrisl | Well, we can probably do a little better than that... | 16:05.29 |
| henrys: thanks | 16:05.31 |
henrys | chrisl: I'll make sure shelly get the new amount when he files his report with miles | 16:07.56 |
rayjj | henrys: cust 532 calls gs_main_init_with_args | 16:11.14 |
| with the last argument being -ffilename | 16:11.40 |
henrys | rayjj: why no gs_main_init_with_args() without filename then gsapi_run_file() for each file? leaks? | 16:14.18 |
| rayjj: that's a lot of setup each fiel | 16:14.41 |
| s/fiel/file | 16:14.46 |
mvrhel_laptop | hi rayjj | 16:16.52 |
rayjj | henrys: the setup time is not very much compared with the rest of the time. I've suggested that they could start the interpreter as soon as they begin spooling the PDF file, but they didn't want to do it (yet) | 16:24.15 |
| if they did do the startup when the file starts getting spooled, it would almost always overlap the startup. By starting gs each time, they don't worry about memory leaks | 16:25.24 |
| mvrhel_laptop: BTW, the patch looks fine -- I had never used \xff escape before, but it should be portable enough | 16:26.21 |
mvrhel_laptop | rayjj: ok thanks | 16:28.50 |
| hmm I guess my commit title was a bit over the character limit. And I thought I had done so well | 16:41.35 |
henrys | won't ask why folks don't have a column and row number in their editor's status bar ;-) | 16:46.27 |
kens | git gui doesn't | 16:46.51 |
| Not everyone is using an editor | 16:46.58 |
henrys | well no big deal, I do think a full featured editor would be less painful for you guys: git config --global core.editor C:\Program Files\Notepad++\notepad++.exe ... | 16:54.40 |
kens | Dopesn't work with git, we tried | 16:55.12 |
rayjj | with vim you have to use: set ruler | 16:56.45 |
henrys | kens: you were right about the vm I'm going to find a fresh iso, I just kept tripping over stuff. | 16:57.04 |
kens | :-( | 16:57.13 |
henrys | mvrhel_laptop: are you using windows 10? does VS work okay? | 17:45.21 |
mvrhel_laptop | henrys: I have windows 10 on one machine. VS 2013 is fine on it. I am waiting to get VS 2015 from a friend of mine | 17:45.59 |
| It has not become available at the store yet | 17:46.13 |
Robin_Watts | Artifex meets the criteria for using VS2015 community edition, for everything except SO stuff. | 17:52.46 |
| And for that we can use VS2015 Express legally. | 17:52.54 |
| VS2015 Express for Windows Desktop that is. | 17:53.04 |
mvrhel_laptop | Robin_Watts: I was holding out for VS2015 Pro | 17:59.41 |
Robin_Watts | From what I could see, if you buy that it's $500+ for 2 years subscription. | 18:00.22 |
| After which time, you don't own it any more. | 18:00.29 |
mvrhel_laptop | Robin_Watts: I would get for much less | 18:00.43 |
kens | MS crap, I certainly won't buy that | 18:00.55 |
Robin_Watts | Right, but still for only 2 years. | 18:00.58 |
mvrhel_laptop | but it does not look like there is much diff between Pro and Community | 18:01.06 |
Robin_Watts | Community == Pro with the limitations in who can use it. | 18:01.22 |
mvrhel_laptop | ok | 18:01.28 |
Robin_Watts | OpenSource software => you can use community. | 18:01.40 |
mvrhel_laptop | Pro has integrated git | 18:01.51 |
| which I dont use on 2013 pro | 18:02.05 |
Robin_Watts | So, I'm sure, does community, but *why* would you want that ? | 18:02.09 |
mvrhel_laptop | so don't care about that | 18:02.09 |
| Community doesn't look like it does | 18:02.27 |
| ok from looking that over, I guess I will just grab community | 18:03.20 |
Robin_Watts | mvrhel_laptop: I'm not sure you can install both community and express at the same time. | 18:04.40 |
mvrhel_laptop | I doubt it. I dont have express | 18:04.54 |
Robin_Watts | Which means you'll be in trouble if you want to work on SOT using it. | 18:05.01 |
mvrhel_laptop | I have not installed any 2015 products yet | 18:05.08 |
henrys | mvrhel_laptop: so community 2015 does work with windows 10? You've built and run gs with that? | 18:05.09 |
mvrhel_laptop | I thought SOT was using 2010? | 18:05.19 |
| henrys: I have not | 18:05.28 |
| I have not downloaded any 2010 VS products yet | 18:05.39 |
Robin_Watts | SOT is using 2010, but we've had requests from customers for 2012. | 18:05.39 |
| so it's possible we're going to end up having to track later versions too :( | 18:05.55 |
mvrhel_laptop | Robin_Watts: might as well go to 2013 | 18:05.56 |
| 2013 is like 2012 but with fixes | 18:06.06 |
Robin_Watts | yeah. | 18:06.12 |
| but they wanted 2012, so we installed 2012 express. | 18:06.24 |
mvrhel_laptop | henrys: I will put community 2015 on my windows 10 machine this weekend and have a go at it | 18:07.31 |
| henrys: above that was supposed to say "I have not downloaded any 2015 VS products yet" | 18:08.11 |
| working on 695192. This is starting to look like a possible bug in lcms. | 18:09.10 |
| It seems to be not doing absolute colorimetric mappings when the source is gray and the destination is rgb or cmyk | 18:09.42 |
| very odd | 18:09.49 |
ray_work | mvrhel_laptop: I had build problems with VS 2015. I did manage to get the profile version to build (in order to see if the profile tools would work for me). I did have build issues (obsolete switches) with other configurations, iirc | 20:23.15 |
| mvrhel_laptop: I didn't want to waste time on it, but maybe someone else will (chrisl?) | 20:23.52 |
mvrhel_laptop | ray_work, do you know how I build gs in windows but with --disable-compile-inits | 21:56.02 |
| looking at 695046 | 21:56.16 |
| Forward 1 day (to 2015/10/31)>>> | |