| <<<Back 1 day (to 2014/08/20) | 2014/08/21 |
ari | hi, i "open" a pdf file with ghostscript. The input.pdf has formular fields which can be filled with acrobat reader. the open.pdf contains no formular fileds. i use this command: gswin32c.exe -q -dNOPAUSE -dBATCH -sDEVICE=pdfwrite -sOutputFile=e:\open.pdf -c.setpdfwrite -f "e:\input.pdf" maybe i have to add an additionel parameter. can anybody help please. thank you | 08:31.28 |
kens | pdfwrite doesn't preserve many kinds of annotations (in this case form fields) | 08:32.05 |
| Also, some kinds of forms can only be created by Adobe applications. Forms which can be filled in with Reader fall into this category. There is no way for pdfwrite to create PDF files of this type (this is proprietary to Adobe). If you use pdfwrite with this kind of input, the resulting PDF file will not be a fillable form in Reader. | 08:34.05 |
ari | thank you, now i can stop searching ;-) | 08:36.16 |
kens | No problem, I'm sorry its not the answer you wanted... | 08:37.04 |
ari | bye | 08:37.13 |
Robin_Watts | kens: hmm. I wonder what ari was trying to achieve there... | 08:51.47 |
kens | Robin_Watts : I have no idea :-) Possibly a smaller PDF file..... | 08:52.01 |
Robin_Watts | tor8: Morning. | 11:39.02 |
| Patch for review on robin/master to update the test-device. | 11:39.16 |
| zeniko also pointed us at a possible mupdf infringement last night, that I haven't looked at. | 11:39.43 |
| One thing that I was slightly surprised at when doing the test-device thing... | 11:40.12 |
| I was hoping that when we detected a color page we could stop the device using the cookie and thus avoid having to run through the rest of the page etc. | 11:40.54 |
| but the cookie isn't available to the device. | 11:41.06 |
| something to ponder on for future, maybe. | 11:41.44 |
tor8 | Robin_Watts: we could stop processing by throwing an exception... | 11:51.14 |
Robin_Watts | tor8: A special class of exception that is swallowed so the caller doesn't see it? | 11:51.40 |
tor8 | I feel that this is a very narrow use case, not worth complicating the interpreter or interface for | 11:51.42 |
| Robin_Watts: the interpreter/null-device already knows to swallow errors | 11:52.14 |
Robin_Watts | tor8: it does? | 11:52.32 |
tor8 | now, the easiest way to speed up processing is to set the ignore images/shadings flags once we've detected color | 11:52.45 |
| yes, the null device catches errors and sets the cookie and error count, doesn't it? | 11:53.04 |
| but I was thinking the test device would eventually grow to detect all manners of things | 11:53.21 |
Robin_Watts | It does. | 11:53.22 |
tor8 | and as such, shorting out early might not always be desirable | 11:53.38 |
Robin_Watts | fair enough. | 11:53.54 |
| In other news, I had company K skype me last night, asking if we were planning to support epub. | 11:54.18 |
tor8 | telling the interpreter to stop decompressing images via the flags is probably going to have the only noticeable impact on performance | 11:54.20 |
| company K, are they the ones that got bought by company I? | 11:54.46 |
Robin_Watts | yes. | 11:54.51 |
tor8 | right. | 11:54.54 |
Robin_Watts | Sumatra has epub support. | 11:55.05 |
tor8 | well, I have a rudimentary start of an epub viewer | 11:55.06 |
| it parses the CSS and XHTML, but currently only processes it enough to do some heuristics and clean up the formatting | 11:55.35 |
Robin_Watts | I wonder if zeniko would be interested (and if we would be interested in) moving sumatras epub viewer into mupdf. | 11:55.43 |
| tor8: oh, right, didn't know you'd got even that far. | 11:55.52 |
tor8 | I use it to clean up crap formatting in ebooks, but extending that to do actual layout and rendering shouldn't be too much work | 11:56.06 |
| I suspect the sumatra epub viewer is all C++ | 11:56.23 |
Robin_Watts | tor8: Interesting. We should mention this to henrys. | 11:56.24 |
tor8 | and from what I've seen when I tried to use it, it doesn't cope very well with CSS styling | 11:56.41 |
Robin_Watts | and it's quite possible that it leverages HTML to actually do the display as it's windows only. | 11:56.44 |
tor8 | it ignores text-indent and margins, which makes it IMO quite unusable :( | 11:57.03 |
| I suspect the biggest stumbling block for company K with epub is going to be DRM | 11:57.33 |
Robin_Watts | tor8: I suspect they will roll their own. | 11:59.02 |
tor8 | Robin_Watts: right. if they're not interested in reading general epubs then I expect it'll be quite a bit easier | 12:00.43 |
| and we can focus on supporting the formatting that they specifically need | 12:00.57 |
Robin_Watts | tor8: Want me to send an email between you and company K where we can discuss their needs? | 12:04.05 |
tor8 | Robin_Watts: we might want to get henry in before we commit to developing yet another tech... | 12:04.34 |
Robin_Watts | tor8: Wasn't suggesting we commit to it, just we find out what they want, but yes. | 12:05.05 |
tor8 | even probing for what they want sends a signal... | 12:05.21 |
| now, I wouldn't mind taking this project on at all! in fact I'd love to. but scheduling's a bitch. | 12:05.43 |
Robin_Watts | tor8: email sent to miles/support summarising this. Please follow up if you feel I've misrepresented the situation in any way. | 12:19.48 |
tor8 | Robin_Watts: looks good. | 12:23.58 |
| Robin_Watts: wouldn't it be easier to store the grayscale difference in a float? | 13:29.20 |
| s/easier/more appropriate/ | 13:29.26 |
| Robin_Watts: and I'm not convinced we need the test_image_data flag, we should just be doing that by default | 13:32.05 |
| so I'd rather see fz_new_test_device(ctx, &colorness) and let the user test the max value coming out | 13:32.35 |
| then it'll be easier to add more test parameters to the device without upsetting the interface | 13:33.06 |
| also, you've rewound the mujs thirdparty submodule with the commit | 13:33.14 |
| I'm not happy about the amount of boilerplate-style code is needed for the shading | 13:35.03 |
| but if we don't care about raw performance, just calling fz_convert_color is a much easier interface than using the cached color converter with callbacks | 13:35.48 |
| which we should do, if we intend this to be a device we point users to to look at how easy it is to write a custom device | 13:36.12 |
henrys | kens:didnât you say the pcl->pdf problem from potential customer Hans was fixed? | 13:43.26 |
kens | henrys I believed it was, did you try it and find out its not ? | 13:43.46 |
henrys | kens: no I was being lazy and asking you. | 13:44.00 |
| ;-) | 13:44.05 |
kens | :-D | 13:44.06 |
| I'd check it before responding to them, but I seem to recall its fixed | 13:44.25 |
henrys | yea Iâll look | 13:44.34 |
rayjj | it looks like Acrobat and mupdf are doing TilingType 2 even when 1 is in the pattern. mudraw -r73 fts_31_3124.pdf has 'gaps' every 7+ cells horizontally. Adobe has a double dot (as they both do vertically) | 13:49.55 |
tor8 | rayjj: I don't think we differentiate between the TilingTypes in mupdf | 13:50.43 |
Robin_Watts | tor8: possibly, we should store the greyscale difference in a float. | 13:50.56 |
| for the external API at least. | 13:51.04 |
| I would be reluctant to use FP internally though. | 13:51.28 |
tor8 | Robin_Watts: yes. internally in the image test we could use ints and then convert to float at the end. | 13:51.41 |
rayjj | apparently gs _does_ respect TilingType. It matches Adobe if I change the file to TilingType 2 (except Adobe fills in with dots and gs leaves a gap (bot vertically and horiz.) similar to mupdf | 13:54.00 |
Robin_Watts | tor8: So what floating point range do you consider reasonable? | 13:56.46 |
| 0 to ... ? | 13:56.54 |
tor8 | I would say just have the squared sum, but using the original float color values in 0..1 range | 13:57.20 |
Robin_Watts | ok. | 13:57.33 |
tor8 | I don't understand how you arrive at the 43350 number though | 13:58.25 |
| oh, difference from grayscale | 13:58.37 |
| not total color difference | 13:58.44 |
Robin_Watts | yeah. 43350 was the largest I could come up with. I'm open to other suggestions. | 13:59.18 |
tor8 | no, that's reasonable, just use the same math but unscaled floats for the interface | 13:59.38 |
henrys | tor8, Robin_Watts: did you see the message from zeniko about nva and report it to Miles or should I? They claim to be 100 programmers from Hanoi. I donât know if Miles can shut them down but he can make the road a little bumpy. | 14:26.24 |
Robin_Watts | henrys: I saw the message, I mentioned it to tor this morning. I haven't investigated it further, or mentioned it to Miles | 14:27.11 |
henrys | tor8: do you want to verify it or can I just report zenikoâs finding to miles? Or obviously you can talk to miles. | 14:28.30 |
| windows users: any uninstall recommendations? something that removes all the related files? | 14:31.41 |
kens | Create a restore point before you install..... | 14:32.14 |
| Other than that, only the application uninstaller knows what went where | 14:32.29 |
henrys | kens: thatâs what I feared | 14:32.56 |
Robin_Watts | henrys: What did you install? | 14:33.03 |
henrys | lots of stuff - dropbox goole drive etc. why canât I remove it properly. I use the stupid thing in the control panel and it leaves all kind of droppings. What a mess. | 14:34.33 |
Robin_Watts | yes, windows is shocking in this regard. | 14:35.52 |
kens | Poor uninstaller. To be honest, the whole thing has been a mess since Microsoft introduced the Registry and started recommending against putting all the application in a singel directory/tree. In the days of .ini files and everything in one place, it was easy | 14:35.53 |
Robin_Watts | tor8: Updated patch online. | 14:37.00 |
| Separate for simplicity of reviewing. I will squash together. | 14:38.45 |
| for the commit. | 14:38.52 |
henrys | kens: yes restore points look like the way to go. I also did something to corrupt my user profile which might have been easily fixed with an restore point. | 14:40.22 |
kens | Oh, corrupted user profiles are effectively unfixable, I've found :-( | 14:40.49 |
| Have you considered re-installing Windows ? | 14:41.03 |
henrys | kens: I backed up my home directory created a new account and copied everything except the NT* files back, does hell await me? | 14:42.04 |
kens | has no iddea.... | 14:42.26 |
jogux | henrys: just get a mac ;) | 14:42.50 |
kens | He's got one :-) | 14:43.05 |
| And Mac OS is even more opaque when it goes wrong. | 14:43.31 |
| Though MS seems to be trying hard to catch up | 14:44.02 |
jogux | :) | 14:44.15 |
henrys | kens: from my experience, the frequency of problems seems much higher on MS but perhaps Iâm just less familiar with it and more likely to step on the mine. | 14:45.31 |
kens | I rarely have roblems, but when I do, they tend to be impossible to fix. | 14:45.59 |
tor8 | Robin_Watts: looking better. but: use floats with the f suffix (0.5f, etc) or the C compiler will generate double precision math | 14:51.48 |
| Robin_Watts: I'm about to head out for dinner. I want to look at the patch in more detail later. | 14:52.34 |
Robin_Watts | tor8: OK. | 14:54.18 |
| Oh, have any of you been watching "Silicon Valley" ? You really should. | 14:55.35 |
kens | Haven't noticed that, what channel's it on ? | 14:56.37 |
Robin_Watts | kens: Sky Atlantic over here :) | 14:57.04 |
kens | Can't get it then | 14:57.18 |
Robin_Watts | It's very well observed. | 14:57.24 |
| tor8: updated patch with f suffixes online. | 15:00.11 |
jogux | robin_watts : hm, I intended to and never did :) | 15:08.31 |
Robin_Watts | henrys: Did we get an answer from Miles about getting copies of Office 2013? | 15:56.04 |
henrys | Robin_Watts: last I heard joann was working on it. | 15:56.30 |
| I got 4 | 15:56.44 |
Robin_Watts | henrys: Ok. It should just be a question of giving them a credit card, and they give you some numbers. | 15:57.23 |
henrys | sorry for the delay I would have done it myself but didnât want to expense it. | 15:57.31 |
Robin_Watts | np. I'm in a similar position :) | 15:57.45 |
henrys | Robin_Watts: I imagine weâll be set up when she gets in today, but I never hear from her until the afternoon. I donât know her exact work hours. | 15:59.10 |
Robin_Watts | ah, of course. | 15:59.25 |
| I forget that Joann actually has to go to an office, and that not all USers spring out of bed as early as you lot. | 15:59.49 |
henrys | Robin_Watts: she is only part-time also. | 16:00.28 |
Robin_Watts | I sprang out of bed early this morning so I could be up in time for the solar installers. | 16:00.31 |
| And an hour later I phoned the company to say "where are they?" and they said "we opened the panels this morning, and they were portrait". | 16:01.03 |
henrys | Robin_Watts: I thought you had that all setup, are you adding more panels? | 16:01.11 |
Robin_Watts | and I said "I wanted portrait, not landscape". | 16:01.18 |
| and they said "oh, right. Well, can we come tuesday?" | 16:01.27 |
| This is for hot water, rather than PV. | 16:01.38 |
| I can't add more PV panels, as I'm at the max allowable under the govt scheme. | 16:02.41 |
henrys | Robin_Watts: Iâve never seen the sun in the UK Iâm amazed it works ;-) | 16:02.50 |
Robin_Watts | We get 66% of the sun of spain, apparently. | 16:03.26 |
jogux | henrys: they work even when it's cloudy ;) | 16:03.49 |
Robin_Watts | jogux: ish :) | 16:04.20 |
kens | Our panels heated our tank to 52 degrees (centigrade) today, and its not that sunny today | 16:04.23 |
Robin_Watts | kens: Do you have it set so that the tank gets topped up by normal heating methods each day ? | 16:05.14 |
kens | Robin_Watts : yes, but the hot water is switched off, we just turn it on as needed | 16:05.32 |
Robin_Watts | (I gather you're supposed to, to avoid legionairres) | 16:05.37 |
| (bloody french, getting everywhere...) | 16:05.58 |
henrys | kens, Robin_Watts, jogux : the payback times are too high, the UK govât will suck it up. I donât think it would work without large subsidies. | 16:10.42 |
kens | Our was fitted by previous occupants.... | 16:11.06 |
henrys | http://spectrum.ieee.org/green-tech/solar/how-free-is-solar-energy | 16:11.15 |
jogux | henrys: the subsidies have been reducing | 16:11.23 |
kens | henrys that seems to be PV, we don't have PV (well Robin does) | 16:11.52 |
Robin_Watts | This is costing me 4400 quid, and I get 2100 back from the govt over the next 7 years. | 16:11.59 |
jogux | I think there's some key underlying thing, like we'll get fined by the EU if we (as a country) don't achieve certain renewable targets. | 16:12.35 |
Robin_Watts | my heating oil costs 1000-1500 quid a year currently, of which 35-40% is likely to be for hot water. | 16:12.48 |
kens | We reckon we save 300-500 per year, though obviously it depends how much hot water you use | 16:12.48 |
| Having a teenage daughter pushes that up.... | 16:13.03 |
Robin_Watts | figure on this system saving half of that. So thats 250quid a year or so. | 16:14.25 |
| figure that heating oil has been going at at 8% a year... it should pay for itself with 10 years. | 16:14.48 |
kens | That sounds simimlar to us, given there are 3 of us | 16:14.49 |
henrys | Robin_Watts: oh right you tax the shit out of gas. Point is I think if the government stepped out of the whole thing and it was a free marketplace hard to imagine many solar panelsâ¦. | 16:15.41 |
Robin_Watts | 5% VAT on heating oil, I think. | 16:16.02 |
jogux | henrys: yep. | 16:16.10 |
Robin_Watts | The price of PV installations has plummetted since I had them done. | 16:20.06 |
| which is kind of an indication that the subsidies worked to kickstart the market. | 16:20.25 |
kens | You certainly see a lot more new builds with PV round here. I'm less convinced its useful in Glasgow.... | 16:21.06 |
jogux | or that the price went up when there was government money sloshing about :-) | 16:21.15 |
| kens : government targets again - new builds have to hit certain efficiency targets | 16:22.03 |
henrys | you do have pretty good wind resource right? | 16:22.07 |
Robin_Watts | Wind is a farce. | 16:22.20 |
kens | yes, but wind is less useful in many ways | 16:22.20 |
jogux | henrys: according to Donald Trump ;-) | 16:22.23 |
kens | its especially less useful on domestic installations | 16:22.35 |
Robin_Watts | onshore wind, certainly | 16:22.38 |
kens | OK time to go, night all | 16:27.24 |
Robin_Watts | henrys: That web page is out of date now. Most installations are now mono-crystalline I think. | 16:28.51 |
| It's more efficient (20% more maybe? I can't remember) | 16:29.06 |
henrys | Robin_Watts: Iâm sort of surprised the govt didnât get more behind wind. Here in some areas wind competes with cheap US gas and is much cheaper than solar except in very hot places. | 16:37.49 |
Robin_Watts | henrys: We have huge wind subsidies here. | 16:42.19 |
| But (onshore) wind farms are a really bad idea in the UK. | 16:42.41 |
| It's a different trade off to the states; you have land to burn. | 16:43.26 |
| wind farms are generally unpopular with the public. | 16:43.58 |
henrys | Robin_Watts: true. | 16:44.18 |
rayjj | anyone available, please have a glance at my commits on http://git.ghostscript.com/?p=user/ray/ghostpdl.git;a=summary | 18:12.19 |
rayjj | probably will wait for kens since everyone else is on SOT | 18:12.52 |
Robin_Watts | the last 2 ? | 18:23.26 |
| the tag one looks fine. | 18:25.45 |
| trying to get my head around the last one. | 18:25.58 |
| You're keeping an extra pointer to the icc_struct, so that it doesn't get rc_decremented away? | 18:27.53 |
| No, that'd be a nasty hack. | 18:29.16 |
| I'd better leave that one to someone that understands it. | 18:30.29 |
rayjj | Wow! confirmed that forcing gs to use TilingType 2 matches Acrobat. A real easy to see one is comparefiles/Bug688728a.pdf | 20:23.25 |
| there are many more progressions. Now the question is, do we have gs 'respect' TilingType, possibly with a parameter, or only do it for PDF input, or always do TilingType 2 ? | 20:24.48 |
| check this page to see it in bmpcmp http://www.ghostscript.com/~regression/ray/compare3.html | 20:26.58 |
| files such as comparefiles/Bug690637.pdf suggests that we shouldn't do it -- gaps are visible even at 300 dpi, but with Acrobat I see lines coming and going depending on zoom | 20:34.35 |
| Bug692517.pdf on http://www.ghostscript.com/~regression/ray/compare7.html looks like a progression, however. Some white stitch lines are "fixed" | 20:40.48 |
| tests_private/comparefiles/PP0001G0.pdf at 72 dpi is interesting -- most of the patterns become "transparent" (on the people's clothing toward the bottom left). Acrobat looks pretty ugly at 144% zoom | 20:55.02 |
Urchlay | can anyone name a linux application that uses libmupdf, besides mupdf itself and zathura-pdf-mupdf? packaging mupdf for slackbuilds.org, want to make sure stuff that tries to link with it is able to | 23:37.00 |
| (in the slackbuilds repo, nothing uses libmupdf except zathura-pdf-mupdf, which for long and boring reasons isn't suitable for testing my libmupdf.a) | 23:38.06 |
| Forward 1 day (to 2014/08/22)>>> | |