| <<<Back 1 day (to 2013/08/01) | 2013/08/02 |
Guest29625 | hi. | 11:35.13 |
| i've got a question | 11:35.34 |
| what will happen, if i wanr to print a pdf data via ghostsrcript | 11:36.21 |
| will ghostscript create an image with 4000dpi and send this to my printer? | 11:37.10 |
| please. need help | 11:39.39 |
kens | Guest29625 : GS can render PDF files, and it can do so at arbitrary resolutions. What happens to the resulting bitmap is up to you, you cnawrite it to disk or send it to a printer, if you want to do the latter you are (probably) on your own though. Do you really have a 4000 dpi printer ? | 11:55.29 |
Guest29625 | got it. thank you kens. | 12:11.25 |
| the problem is, that the data which is recieved by the printer is 120mb per din a4 letter. | 12:12.10 |
kens | Well, 4000 dpi is quite a lot of data | 12:12.35 |
Guest29625 | the 4000dpi is wrong | 12:12.50 |
kens | :) | 12:12.54 |
Guest29625 | its my fault | 12:12.55 |
| the software creats an postscript file and than a pdf | 12:13.14 |
kens | OK | 12:13.20 |
Guest29625 | and this pdf will be printed via ghostscript | 12:13.48 |
kens | wionders what it uses to create the OPDF file, possibly Ghostscript :-) | 12:13.51 |
| What OS is this ? And what printer ? | 12:14.03 |
Guest29625 | its a very strange situation. i work in a software company in customer support | 12:14.43 |
| nobody here can explain, why the programmer choose this way | 12:15.28 |
| postscript->pdf->ghostscript->print | 12:15.47 |
kens | Well, that's fairly nonsensical, PS->Ghostscript is *much* more sensible | 12:16.09 |
| quicker and less prone to errors | 12:16.21 |
| You still haven't said which Operating System this is, nor what is doing the conversion to PDF | 12:16.43 |
| Nor how you are driving the pritner from Ghostscript | 12:16.58 |
Guest29625 | this is the way i prefer and try to convince them | 12:17.30 |
| its windows 7 | 12:18.02 |
| and is done by ghostscript i think | 12:18.35 |
kens | OK so what is converting the PostScript to PDF, and how are you printing the file ? Are you using the mswinpr2 device ? | 12:18.38 |
| If you are using Ghostscritp to convert the PS to PDF< then you should stop that, ther's no point at all | 12:18.59 |
| By the way, you haven't actually said what your problem is | 12:19.21 |
Guest29625 | sorry. the pdf is created by a java library | 12:21.17 |
| i asked the programmer | 12:21.38 |
kens | From PostScript ? I bet its using Ghost4J then | 12:21.40 |
Guest29625 | and yes. we are using mswinpr2 | 12:22.03 |
kens | SO what is the problem ? You haven't said | 12:22.14 |
Guest29625 | the problem is, that it takes a very long time to print the file | 12:22.49 |
kens | Well it will yes | 12:22.56 |
Guest29625 | the customer will print 8pages | 12:23.01 |
| so it will take more than 15minutes | 12:23.29 |
| because of its 850mb | 12:23.38 |
kens | The PostScript is rendered to a high resolution bitmaop (takes time) which is then sent to the printer (takes time) potentially wrapped up in some other printing language, which the printer then pritns one sample at a time | 12:23.41 |
Guest29625 | this is the way how we print the file | 12:24.35 |
| C:\Program Files (x86)\gs\gs\bin\gswin32c.exe -dPrinted -dNOPAUSE -dBATCH -dQUIET -dNOPROMPT -dDEVICEWIDTHPOINTS=842.0 -dDEVICEHEIGHTPOINTS=595.0 -sDEVICE=mswinpr2 -sOutputFile=%printer%PDFCreator -IC:\Program Files (x86)\gs\gs\lib;C:\Program Files (x86)\gs\fonts C:\Users\OFRA\AppData\Local\Temp\\TV2\GBR.pdf | 12:24.42 |
kens | You are sending the file to a pritner called 'PDFCreator', which presumably will create a PDF file..... | 12:25.18 |
Guest29625 | are yoz from germany? | 12:25.27 |
kens | No, but I can speak some German | 12:25.35 |
Guest29625 | oh ok...i just saw that the server which u use is in frankfurt | 12:25.55 |
kens | the IRC server may be there, I'm not | 12:26.08 |
Guest29625 | the printer in this example (pdf creator) is just an example | 12:26.24 |
| normaly we use a laser printer which is conencted to the company network | 12:26.41 |
kens | OK well as I said, what you are trying to do is always going to be slow, its a process which will produce and transmit a lot of data, which then needs to be consumed by a *usually pretty dumb) pritner. | 12:27.18 |
Guest29625 | ok thank you | 12:27.47 |
kens | The poitn isn't to get high speed, its to allow you to print to a printer hwich doesn't understand PostScript (or PDF) | 12:27.49 |
Guest29625 | yes. but is there any printer which can't do this? | 12:28.26 |
| the requirements for our software presuppose an printer which understand postscript | 12:29.06 |
kens | There are any number of PostScript pritners areound, you could send the PostScript to one of those.If you insist on converting to PDF first then you need a PDF prtiner, these exist too, btu are more expensive | 12:29.08 |
| If you have a PostScript printer,why are you converting the PostScript to PDF ? You should just print the PostScript | 12:29.43 |
Guest29625 | this is what i am talking about. but we also want to have an pdf file saved on the computer. | 12:30.19 |
kens | If you don't have a PostScript printer, then you can use Ghostscript to print to it using the Windows printing system (mswinpr2 device), but you need to understand that this will be slow and use a lot of memrorfy | 12:30.29 |
| So why not send the PostScript file to the pritner *and* convert the same file to PDF ? | 12:30.51 |
| It seems like your conversion to PDF is what is causing your problem, so don't do that..... | 12:31.22 |
Guest29625 | this is my suggestion which i was tipping right now | 12:31.27 |
| but you are to fast^^ | 12:31.31 |
kens | But I make a lot of typos ;-) | 12:31.53 |
Guest29625 | no problem so long i understand it | 12:32.31 |
| so...thank you kens. now i try to explain it to our developer | 12:32.51 |
kens | OK feel free to have him come and discuss it if you like | 12:33.08 |
Guest29625 | ok thank you. i tell him this | 12:33.41 |
| bye | 12:33.51 |
henrys | mvrhel_laptop: you around? | 17:16.53 |
mvrhel_laptop | hi henrys | 17:17.40 |
| I am looking at the raspberry pi devices for the japan demo | 17:17.58 |
henrys | oh cool I was going to ask if you'd like me to take your chicago days since you have the new trip | 17:18.34 |
mvrhel_laptop | oh. I did already buy my ticket though | 17:18.52 |
henrys | oh I thought you were visiting and could spend time in chicago. | 17:19.31 |
| s/could/wanted | 17:20.07 |
mvrhel_laptop | henrys: my plan was to spend a couple days at my parents between the staff meeting and the last couple days of the show. a couple days at my parents will have me maxed out there :) | 17:20.22 |
| so I don't mind doing the show | 17:21.06 |
henrys | okay well if you want to change your flight let me know, it doesn't much matter to me either way. | 17:21.19 |
mvrhel_laptop | oh. well if I were to change my flight that would be fine | 17:21.33 |
| let me check on the cost of that | 17:21.37 |
| that would make my wife happy probably | 17:21.47 |
henrys | mvrhel_laptop: miles has lots of money ;-) | 17:22.03 |
mvrhel_laptop | it is on southwest so the cost might not be to bad | 17:22.14 |
| we had a fun time yesterday. went to this place http://www.goldmyer.org/ | 17:23.20 |
| took about 1.5 hours of driving on a rough forest service road, followed by about a 5 mile mt. bike ride | 17:23.49 |
henrys | mvrhel_laptop: wow that looks cool | 17:24.14 |
mvrhel_laptop | it is only about 30 miles from where I live as the crow flies but not so easy to get o | 17:24.33 |
| to | 17:24.34 |
| kids bellyached a bit due to the bike ride but other than that it was fun | 17:25.00 |
| we had the place to ourselves for quite awhile. right when we were leaving a pile of naked people showed up | 17:25.27 |
henrys | naked? | 17:26.12 |
mvrhel_laptop | well, they were clothed and then stripped down | 17:26.23 |
henrys | jungle ritual like stuff? | 17:26.35 |
mvrhel_laptop | it is clothing optional | 17:26.40 |
| but the place is limited to 20 visitors a day | 17:26.55 |
| just hippy northwest type of place | 17:27.38 |
henrys | I think the only difference between colorado and washington is humidity | 17:28.48 |
mvrhel_laptop | the water comes out of a little crack in the cave at about 110 and they have built a series of walls for cascading pools that end up being at different temperatures | 17:29.03 |
| yes. about CO vs WA | 17:29.13 |
henrys | anyway I'm going to book pretty soon so let me know, happy to take on a few more chicago days if it helps. | 17:34.55 |
mvrhel_laptop | henrys: ok. I will figure this out today. | 17:40.20 |
| thanjs | 17:40.24 |
| rats the raspberry is on backordered | 17:40.37 |
| marcosw: do you have one of these? | 17:40.58 |
marcosw | mvrhel_laptop: a raspberry pi? Yes. | 17:41.33 |
mvrhel_laptop | is it the model b and already up and running gs? | 17:41.52 |
| oh here is one | 17:42.07 |
marcosw | yes to both. | 17:42.21 |
mvrhel_laptop | was it pretty easy to get up and running. i am thinking that this device would be good for the japan demo | 17:42.56 |
marcosw | it was trivial to get up and running. you download raspbian (raspberry debian) and put it on a SD card and that's it. | 17:44.22 |
mvrhel_laptop | ok. I can handle that | 17:44.41 |
chrisl_away | mvrhel_laptop: you could try e-bay, over here there are quite a few people selling them from stock on there | 17:45.56 |
marcosw | looks like they've made it easier (not sure how it could be), the new software is called NOOBS :-) | 17:46.33 |
mvrhel_laptop | the thing will be to figure out how best to showcase it ripping through files and rendering. ray and I had talked about showing it creating bitmap files at printer resolution for speed demo | 17:47.55 |
| and maybe showing rendered pages out the hdmi port for rendering demo | 17:48.26 |
ray_laptop | henrys: did you end up with the gumstix that I had put in a box ? | 18:06.28 |
henrys | ray_laptop:yes I have it. | 18:06.54 |
| I was going to toss it though - I assume it's dated | 18:08.07 |
dogisfat | How do small platforms like the raspberry pi handle ripping with such little memory? | 18:14.02 |
henrys | mvrhel_laptop: I though the sales guy was setting up a demo? | 18:20.35 |
mvrhel_laptop | henrys: miles, asked me to put something together | 18:22.50 |
| I think I found everything I need at mcm | 18:23.18 |
ray_laptop | dogisfat: 512M is not that small | 18:25.59 |
| One of our printer companies uses us to render PDF, WITH transparency and only gives us 14M (but that _is_ monochrome) | 18:26.51 |
dogisfat | How do you handle bitmap images with such small memory places | 18:27.49 |
ray_laptop | dogisfat: with the clist stored in memory, we render in bands. If the memory gets low, we compress the clist | 18:27.59 |
| dogisfat: the source data bitmap is processed from the stream incrementally and never expanded to full size | 18:28.41 |
| dogisfat: (unlike mupdf) | 18:28.57 |
| dogisfat: And we can limit the image resolution stored in the clist to lower resolution than full printer res, since when halftoning is used, the images at full res aren't needed | 18:30.25 |
dogisfat | That makes sense | 18:31.17 |
mvrhel_laptop | ok. that is all ordered. now on to the P1 bug | 18:31.32 |
ray_laptop | mvrhel_laptop: great. Does it have HDMI out ? | 18:32.19 |
mvrhel_laptop | ray_laptop: yes | 18:32.28 |
ray_laptop | cool. | 18:32.33 |
henrys | ray_laptop: do you want me to bring the gumstix to chicago? | 18:35.12 |
ray_laptop | henrys: Sure, if you were just going to throw it out. | 18:41.51 |
| It's only 600MHz and 256M iirc | 18:42.10 |
| but it has a CF card "disk" | 18:42.46 |
henrys | will do | 18:43.03 |
ray_laptop | hi, robin_watts_mac | 18:46.29 |
| how's it going ? | 18:46.37 |
robin_watts_mac | going fine so far - except for losing my glasses on a dune. | 18:46.56 |
mvrhel_laptop | any interesting animals? | 18:47.21 |
robin_watts_mac | mvrhel_laptop: Zebra, Oryx, Springbok, Yellow mongoose, bat eared fox. | 18:48.01 |
mvrhel_laptop | cool | 18:48.19 |
robin_watts_mac | but we're not really in the animal bit of the holiday yet. | 18:48.34 |
| scenery/landscape so far. | 18:48.48 |
ray_laptop | mvrhel_laptop: for the demo on the ZBOX, I suggest mentioning multi-threaded rendering since both of the ones they suggested are dual core | 18:49.23 |
| robin_watts_mac: sounds fascinating | 18:49.39 |
paulgardiner | robin_watts_mac: Coo, sounds like you aren't going to be disappointed then. | 18:50.31 |
mvrhel_laptop | ray_laptop: ok. good point. | 18:50.41 |
robin_watts_mac | paulgardiner: Indeed not. We've just met with people doing the same tour but in the opposite direction, and they said they saw loads of animals. | 18:51.15 |
| paulgardiner: You so should have come along. | 18:51.26 |
mvrhel_laptop | just bring an eyeglass strap | 18:51.41 |
paulgardiner | robin_watts_mac: Hmmm, just what I'm thinking | 18:52.01 |
ray_laptop | that ZBOX is quite fast 1.6GHz dual core DDR3 1333 with a 64GB SSD -- should really scream compared to a raspberry | 18:53.18 |
| mvrhel_laptop: I just noticed that you weren't cc'ed on the email from them. I just forwarded it to you. Takane-san should be notified that you are going, and make sure that you get cc'ed | 18:56.06 |
mvrhel_laptop | ray_laptop: yes I seem to be out of the loop on some of this. | 18:56.56 |
robin_watts_mac | henrys: You asked the other day: "marcosw1:I wonder if a live test directory can be exported for network mount somewhere?" | 18:59.33 |
| You mean, you want the test files available as a network mount somewhere? | 18:59.53 |
henrys | are you on a cell phone hot spot or do they have wireless there? | 18:59.59 |
robin_watts_mac | wireless in the guest house here. | 19:00.09 |
kens | Hello robin_watts_mac | 19:00.18 |
robin_watts_mac | 75 quid for 20 meg if I used the cellphone. Bargain. | 19:00.31 |
| I did wonder about having a web interface behind the password wall to get individual test files out. | 19:01.19 |
henrys | robin_watts_mac: yea why not? most of my testing is bmpcmp ... | 19:01.23 |
robin_watts_mac | and then I could hook the bmpcmp results up with web links to get the files. | 19:01.42 |
| I need to figure out the right commands to drive svn to get files out, but it should be a nice solution. | 19:02.10 |
henrys | what was the reason we haven't moved tests to git? | 19:02.39 |
robin_watts_mac | henrys: With git you can't do partial checkouts | 19:02.56 |
henrys | submodules? | 19:04.06 |
robin_watts_mac | henrys: ick. The problem is with git you can't easily say "give me all the tests_private/pdf/forms" files | 19:04.42 |
| unless we set all the directories up as submodules. | 19:04.53 |
| which would lead to madness, I fear. | 19:05.02 |
mvrhel_laptop | bbiab | 19:06.42 |
SpNg | I'm working on reading color values from a .eps by invoking a postscript program with ghostscript. It works great on most eps files, but on some I get Error: /typecheck in --setmatrix-- what is the best way to troubleshoot this? | 19:09.08 |
henrys | robin_watts_mac: from reading sounds like there are diamonds in that desert Has Helen been keeping here eyes open for one? | 19:10.26 |
robin_watts_mac | henrys: In the "diamond regions" they don't let you off the road. | 19:10.56 |
kens | SpNg you need a good working knowledge of PostScrip t to debug it. | 19:11.09 |
robin_watts_mac | cos 1/2 the diamonds belong to the government, and the rest to de beers. | 19:11.13 |
SpNg | kens: I'm using the technique you suggested for overriding fill, stroke, ufill, etc. procedures to flush the color values, then calling the original procedure. Could the error mean a corrupt eps file? | 19:13.21 |
kens | SpNg possibly, its hard to tell. Does hte EPS render if you render it to the display device ? | 19:13.45 |
SpNg | kens: it does, I can open it with Illustrator | 19:15.05 |
kens | THen its not corrupt | 19:15.12 |
| Actually, Illustrator is not a good test | 19:15.26 |
SpNg | kens: when I run eps2eps on it, I get a similar error, Error: /typecheck in --moveto-- | 19:15.36 |
kens | Open it in Ghostscript without your special stuff | 19:15.39 |
| Illustrator probably isn't interpreting the PostScript | 19:16.06 |
SpNg | kens: I'm running Ghostscript on a linux vm with no GUI. This same script works fine with the majority of EPS I run through it | 19:18.16 |
kens | Yeah but that doesn't guarantee anything. | 19:18.32 |
| If you can't use the disply, try a file device. TIFF or PNG or JPEG | 19:18.47 |
SpNg | kens: ok. one sec | 19:19.03 |
| kens: just output to a png fine. command: gs -dEPSCrop -dSAFER -dBATCH -dNOPAUSE -sDEVICE=pngalpha -dGraphicsAlphaBits=4 -dEPSCrop -sOutputFile=tf-rgb.png tf-rgb.eps | 19:21.11 |
kens | THen its not a corrupt EPS, its your modifications | 19:21.25 |
SpNg | kens: are there procedures that can't be intercepted? | 19:23.31 |
kens | PostScript is a programming language, attempting to alter the behavious of a program can cause trouble | 19:24.10 |
SpNg | I'm intercepting: stroke, ustroke, rectstroke, fill, rectfill, ufill, eofill, ueofill and then calling the original for each one | 19:24.37 |
kens | There can be strange side-effects, which can only be deduced after the fact. | 19:24.56 |
SpNg | so with this situation, I get a call to my stroke intercept, get the output, then process fails and Last OS error: No such file or directory shows up | 19:25.57 |
kens | well that just means things went wrong I suspect | 19:26.24 |
SpNg | haha, your answers are so matter of fact. I'm trying to figure out how to ask a question that can help me move forward here. ;-) | 19:27.24 |
kens | Honest answer; probably you can't | 19:27.38 |
| debugging PostScript requires an excellent working knowledge of the language | 19:28.02 |
| You can probe thee PostScript by inserting things like "(I am here now) == flush" to see what part of the program causes the problem | 19:28.48 |
SpNg | yeah that is what I'm doing. I have added a snippet to each of the intercepts. I'm following your technique here: http://stackoverflow.com/questions/16680747/how-can-i-extract-color-values-from-an-eps-file | 19:29.26 |
kens | I meant you can insert the probes in the original EPS file | 19:30.20 |
| TO find out how its calling the code which causes the error. | 19:30.36 |
| Undoubtedly it will be using a named routine, which may well call other named routines | 19:30.50 |
SpNg | oh I got you, that makes sense. | 19:31.05 |
kens | My usual technique is to 'unroll' those in place back to basic PostScript operations | 19:31.14 |
| And finally to see which operation causes the problem, which will usually give me a clue as to *why* its a problem. | 19:31.39 |
SpNg | start dissecting the eps. sounds nice and tedious ;-) | 19:31.52 |
kens | Essentially you need to reverse engineer the EPS file | 19:31.57 |
| Yes, its tedious, there is no really good way to do this I'm afraid, which is why I say it really needs a good working knowledge of the language | 19:32.31 |
SpNg | ok, that is fair. I think I can move around the language enough to at least try to find where the problem exists. | 19:33.23 |
| Thanks for taking the time to point me in the right direction. I will circle back on the next installment, but this gives me a direction | 19:33.52 |
kens | No problem, feel free to ask again if you have more questions | 19:34.05 |
| OK off again, night all | 19:45.02 |
ray_laptop_ | since the customer is hot for bug 694502, I did a little playing with it. mupdf can display it as well as Acrrobat, and mutool clean -d decompresses the image (still binary, but no filter) and gs then displays it OK. | 20:12.49 |
| the GS CCITT decoder gets an ioerror after 184 lines, but the mupdf filter keeps going past this. The truncated lines of text begin about line 375. | 20:14.08 |
| (all that's in the comments), but in case any mupdf filter person (tor?) can recall any "fixes" to the FaxDecode filter, it might help us figure out wazzup with gs having the problem. | 20:15.24 |
robin_watts_mac | ray_laptop_: zeniko did some fixes. | 20:22.05 |
ray_laptop_ | OTOH, gs copes with 2131_-_broken_faxd_image.pdf better than mupdf (and Acrobat for that matter) displaying more of the data before giving up with "Insufficient data for image" | 20:22.14 |
robin_watts_mac | but I'm not sure the gs and mupdf fax decoders share a common ancestry. | 20:22.30 |
robin_watts_mac | must go to bed. night! | 20:22.37 |
ray_laptop_ | robin_watts_mac: well, I searched the logs, and the the 1d negative and 1d overflow files run OK on gs | 20:22.54 |
| g'nite, robin_watts_mac | 20:23.02 |
| robin_watts_mac: IIRC, to wrote the mupdf fax decoder from scratch | 20:23.25 |
robin_watts_mac | ray_laptop_: yeah. | 20:23.33 |
sebras | robin_watts_mac: heh, you have the cool camera equipment, but I only see Helen's pictures. :) | 20:26.53 |
| ray_laptop_: yeah, the fax decoder is developed separately from gs. | 20:27.29 |
mvrhel_laptop | great. the softmask is an image and that image has a softmask, | 20:47.09 |
| we seem to be going through the correct pushes and pops though | 20:48.01 |
| oh interesting. getting rid of the images softmask we still dont render properly | 20:53.08 |
| something funny going on here | 20:53.18 |
| hmm there is a unique case here in that the softmask group has a gray icc color space associated with it | 21:00.09 |
| something odd is going on with the image for the softmask | 21:39.53 |
| this is looking less like a transparency issue | 21:40.24 |
| Forward 1 day (to 2013/08/03)>>> | |