| <<<Back 1 day (to 2015/04/23) | 20150424 |
godd_fella | Hi, guys! | 03:36.05 |
| Wanna leave some gratitude for mupdf developers. | 03:37.35 |
| Men, you are the best! | 03:37.44 |
| One question... | 03:38.23 |
| You've implemented JPG, PNG, XPS, CBZ and even EPUB support in mupdf. | 03:40.01 |
| But how about DJVU, would it be sometime supported by mupdf? That would be great. | 03:41.02 |
chrisl | godd_fella: djvu will almost certainly not get into mupdf as long as the patents that it involves are current | 06:51.09 |
godd_fella | sad... | 06:51.59 |
chrisl | It's the primary reason djvu support (and use) isn't more wide spread | 06:52.43 |
godd_fella | But here http://djvu.sourceforge.net/lti-licensing.html stated: | 06:59.52 |
| With this source code release, developers around the world have permanent access to the DjVu specification, as well as critical parts of the technology. Developers can build on this foundation to create DjVu viewers, create translators to other file formats, integrate DjVu with other open source tools, and use new technological advances to improve image quality or the compression ratio. | 07:00.03 |
| abc | 07:00.19 |
chrisl | Doesn't change the fact that it's patent encumbered. Don't forget, mupdf is also available under a commercial license | 07:02.13 |
kens | Making the source code open souirce doesn't get round patents | 07:02.15 |
godd_fella | "mupdf is also available under a commercial license" so that's the reason? | 07:03.13 |
chrisl | godd_fella: no, but it feeds into the decision making process | 07:04.17 |
kens | No, because its patent encumbered. The same reason we don't (for example) have JPEG2000 output in Ghostscript | 07:04.28 |
godd_fella | Okay | 07:05.29 |
| So, if I will ever be an entrepreneur, to spread MuPDF in my company's environment would require commercial licensing? | 07:08.23 |
kens | That depends on how you use it. | 07:08.36 |
godd_fella | Watching documents, etc. | 07:09.02 |
kens | You should read the Aferro General Public LIcence. If you are ciompliant with that then you don't need a licence. | 07:09.32 |
| I hsould say, you don't need a *commercial* licence | 07:10.14 |
godd_fella | I suppose that commercial licencing is necessary if I will emmit my product which uses or based on MuPDF, right? | 07:11.42 |
kens | Broadly speaking, if all you do is use a AGPL piece of software, you don;t develop any applications which depend on it and you don't seel it, then you don't nbeed a commercial licence. If you create a product which uses MuPDF then you are more likely to need a licence. | 07:12.14 |
| It is possible to use AGPL software in a commrrcial product without having a commercial licence, there are people using GS like that, but those of us who make our living that way don't like it, and you can forget support if you try that :-) | 07:13.09 |
chrisl | godd_fella: the conditions are the same as any (A)GPL software. The GPL and APGL are (especially for legal documents) quite easy to read and easy to understand, so are worth reading if you plan to push almost any open source software in a business setting | 07:14.34 |
kens | The main difference with AGPL as opposed to GPL is Software-as-a-service, where clients provide you with 'data', you run applications on your own hardware, and return 'something' to the client. In tat case under the AGPL you would still need a commercial licence (or have to open source the entire process) | 07:14.55 |
godd_fella | Thanks for clearing the things up for me. | 07:15.06 |
| And if I would like to make a donation to support MuPDF developent, where do I go? | 07:16.14 |
kens | Currently we don't work that way :-) | 07:16.28 |
godd_fella | damn :-D | 07:16.40 |
kens | All our work is funded by our commercial activities at the moment | 07:16.43 |
chrisl | godd_fella: and for completeness, our commercial arm also offer support contracts, so again, in a business setting, you might not need a commercial license, but many companies stipulate that user/technical support must be available | 07:16.57 |
kens | So kind words are reward enough :-D | 07:16.59 |
godd_fella | I just have no words of how awesome MuPDF is... After so long times of suffering in a Adobe Reader cave, it's like breath of fresh air. | 07:19.01 |
| Thanks for everyone of you involved! | 07:19.13 |
| I have some feature requests, can I post them here? | 07:19.56 |
kens | The people who do the real work will read the logs, thans for the kind words. You can post feature requests here, but it might be better to do so in a bug report, as that way they won't get lost. | 07:20.23 |
chrisl | godd_fella: it would be preferable if you put them in our bugzilla tracker: http://bugs.ghostscript.com/ | 07:20.33 |
godd_fella | OK, I'll dublicate them there. | 07:21.02 |
chrisl | The main mupdf developers tend to be around here from a little later in the day - two/three hours from now | 07:21.39 |
godd_fella | How do you print PDF on Windows? Going throug mudraw export and printing images degrades quality of output. So it would be nice to have something like printing dialog like in Adobe Reader. | 07:23.34 |
chrisl | godd_fella: we have a more fully featured viewer in development right now, it might be worth trying that: http://www.gsview.com/ | 07:24.51 |
kens | Good grief, apparently RackSpace have opened a data centre just down the road from me, I had no idea..... | 07:24.55 |
godd_fella | chrisl: Have tried it, but got ugly print output :-) | 07:26.37 |
chrisl | Hmm, you shouldn't..... perhaps with mudraw you need to try a higher resolution? | 07:27.31 |
kens | I thought Micahel would pick up the resolution form the device spec | 07:27.55 |
| But its undoubtedly not well tested | 07:28.07 |
chrisl | Since it's using XPS for printing, the resolution *shouldn't* matter.... | 07:28.41 |
| Unless there's transparency..... etc | 07:29.00 |
| But I was meaning using mudraw to dump image files to print | 07:29.27 |
kens | Don't forget there are caveats with the XPS output and it *defintely* isn't well tested :-) | 07:29.43 |
chrisl | Very true | 07:29.53 |
godd_fella | Yep, right now I'm using mudraw's 600 dpi PNGs for printing, but it's a pain. | 07:30.19 |
chrisl | What resolution is your printer? | 07:30.35 |
kens | Better yet, what *is* your printer ? | 07:30.47 |
godd_fella | Xerox_5325 | 07:31.30 |
kens | Hmm *big* printer | 07:31.56 |
| OK so its 600 dpi | 07:32.22 |
godd_fella | I'm at work, right now) | 07:32.36 |
kens | You can almost certainly get decent output by usign lower resolution bitmaps | 07:32.44 |
godd_fella | Yes, decent. | 07:32.59 |
| But I would like perfect :-) | 07:33.16 |
chrisl | You could try running mudraw without anti-aliasing..... | 07:34.09 |
kens | Since the printer will have to halftone/dither you are essentially throwing up to 75% of the data away. Its not going to be 'perfect' starting fomr a continuous tone bitmap and printing to a halftone | 07:34.12 |
godd_fella | And increasing resolution isn't an option, 'cause printer got much time receiving this megabyte load... | 07:34.13 |
kens | Increasing resolution definitely won't help, that just increases the amount of data to be discarded | 07:34.34 |
| It accepts TIFF didrectly, so you could render to TIFF and send that. | 07:35.09 |
| It also (apparently) accepts PDF so you should be abl;e to send the original PDF to it. | 07:35.47 |
godd_fella | So there is no way for MuPDF to support direct PDF printing? | 07:36.10 |
kens | To print you have to hook into the underlyign OS printing system. That is (of course) OS-depenendent and MuPDF does not currently do so, at least on WIndows. | 07:36.49 |
| chrisl I thought you were out today | 07:38.23 |
chrisl | kens: I'm just faffing around until my ride turns up - which, it turns out, won't be until 10...... | 07:41.28 |
kens | Ah | 07:41.34 |
chrisl | I'm going to collect the new car :-) | 07:41.51 |
kens | Cool | 07:41.59 |
| Have fun | 07:42.03 |
chrisl | Thanks. Shame it's Friday - visions of wafting around the South Downs to put a few miles on it rather wilt in the face of Friday traffic...... | 07:43.00 |
kens | Mmm, yes, not the best day probably | 07:43.17 |
chrisl | Still, I'm sure I can find a suitably indirect route home! | 07:43.57 |
kens | :-D | 07:44.03 |
| Grrr kicked off gs-devel for 'excessive bounces' again | 07:49.37 |
chrisl | Yep, me too - I think the errors are incorrect. | 07:50.04 |
sebras | doesn't mailmain keep a whitelist? | 07:50.19 |
kens | I'm sure they must be, I don't believe I'm bouncing mails | 07:50.20 |
kens | has no clue about mailman | 07:50.42 |
chrisl | I'm not sure about a whitelist. I did poke around in the logs quite a while back, and I couldn't find any hint of bounces | 07:51.21 |
| Plus, given that kens and I were both kicked off the list this morning, that's too big a coincidence - I doubt we'd both be bouncing mails at exactly the same rate | 07:52.22 |
kens | Yeah its more like a timeout or something | 07:52.36 |
chrisl | Ah, actually..... A mail went through gs-devel earlier which is *clearly* spam (html message with nonsense in it, and a .wav attachment). I bet gmail "bounced" that..... | 08:04.33 |
kens | Hmm, that's probably it | 08:08.27 |
| But 'one' mail ?...... | 08:08.35 |
chrisl | Well, probably more than one, but I only looked at the last few days. | 08:09.13 |
| And the mail in question doesn't appear in the Spam "label" in gmail, so <shrug> | 08:09.40 |
kens | Well if gmail bounced it I guess it won't appear in the spam box | 08:09.56 |
| Grr gmail is spam filtering gs-regression again | 08:10.22 |
chrisl | gmail is so sh*t about these things - bemoning the loss of spamcop :-( | 08:11.13 |
kens | I know :-( | 08:11.22 |
| I keep on marking the gs-regression mails as 'not spam' and a few weeks later, it starts deciding theyt are spam again | 08:11.43 |
chrisl | Yes, I find that as well. And now it seems to be silently bouncing (what it judges to be) spam mails, without any chance for the user to intervene..... | 08:13.53 |
kens | Which is even worse.... | 08:14.08 |
chrisl | I think another coffee...... | 08:15.26 |
kens | Haven't had my first one yet :-( | 08:15.40 |
chrisl | Good grief! How can you function......? | 08:16.50 |
kens | Badly ? :-D | 08:16.57 |
godd_fella | Have a nice day, guys! | 08:17.51 |
kens | bye bye | 08:17.58 |
Robin_Watts | kens, chrisl: gs-devel kicked me too this morning. | 08:29.48 |
kens | So its probably the same reason all round, I suppose that's something | 08:30.05 |
chrisl | Something we should bitch at google about, then? | 08:31.59 |
kens | I doubt we'd get any joy there | 08:32.14 |
chrisl | I'd sort of hoped as a paid customer, they'd at least *pretend* to listen...... | 08:32.52 |
Robin_Watts | Your email is in a queue... | 08:33.18 |
kens | Google wants you to talk to The Machine, no matter who you are. Oh, maybe paid advertisers get better treatment | 08:33.27 |
chrisl | Off now! Bye.... | 08:59.25 |
kens | have fun | 08:59.29 |
chrisl | :-) | 08:59.35 |
Robin_Watts | http://www.bzarg.com/p/how-to-pronounce-hexadecimal/ | 14:06.43 |
kens | bibbity-bobbity, bibbity-bobbity, bibbity-bobbity-doo | 14:08.25 |
| And Isee someone feels the same way in the comments | 14:09.58 |
sebras | kens: new no new age advanced ambient motor music MACHINE! | 14:17.03 |
kens | :-D | 14:17.41 |
sebras | kens: somehow I dropped out after you mentioned that google wants you to talk to "The Machine", but I noticed that I missed a few messages today. | 14:21.02 |
kens | I've been kicked off a couple of times too | 14:21.21 |
| Wow, C stack overflow, excellent error | 14:24.02 |
| I wonder what I broke :-) | 14:24.11 |
Laurenceb__ | hi | 14:27.41 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 14:27.41 |
Laurenceb__ | gs -q -dNOPAUSE -dBATCH -dPDFSETTINGS=/prepress -sDEVICE=pdfwrite -sOutputFile=output.pdf input.pdf | 14:27.58 |
| im using this to build fonts into a pdf file | 14:28.22 |
| but i worry there is an issue | 14:28.28 |
| i see a lot of YTBXNU+f-0-0 TrueType yes yes yes 2343 0 | 14:28.35 |
kens | And ? | 14:28.42 |
Laurenceb__ | where previously there was Helvetica in the origionsl file | 14:28.51 |
| is it poissible ghostscript is included the font multiple times? | 14:29.10 |
kens | There may have been a request for Helvetica, but no font | 14:29.12 |
Laurenceb__ | right | 14:29.19 |
| Helvetica is in embedded graphics | 14:29.29 |
kens | You are seeing multiple subsets. | 14:29.32 |
Laurenceb__ | ill pastebin | 14:29.39 |
kens | THats what the XXXXX+ prefix means | 14:29.44 |
| Clearly you have a TrueType substitute defined for Helvetica | 14:30.00 |
Laurenceb__ | http://pastie.org/10111748 | 14:30.10 |
| my file size has grown by 3.5M | 14:30.22 |
| im trying to reduce it | 14:30.27 |
kens | Not much I can say without seeing the original file. Note that GS+pdfwrite is not intended as a font embedding tool | 14:31.02 |
Laurenceb__ | yeah | 14:31.21 |
| http://pastie.org/10111753 | 14:31.22 |
| origional file | 14:31.26 |
kens | When you run a file through GS+pdfwrite the goal is that teh visual appearance of the PDF should match the visual appearance of the original input | 14:31.31 |
| Those 2 pastes look identical to me | 14:31.53 |
Laurenceb__ | first is smaller | 14:32.07 |
| yeah looks like the font isnt in there multiple times | 14:32.20 |
| grrr | 14:32.21 |
| something else is causing file size growth | 14:32.33 |
kens | Like I said, I'd have to see the original file | 14:33.19 |
Laurenceb__ | its confidential lol | 14:33.33 |
kens | Well, not really a lot I can say, there could be *lots* of reasons | 14:33.50 |
| There are multipe subsets of Helvetica in the new file, and Helvetica is not present in the original. My first suggestoin would be to sort out whatever is causing a substitution with a TrueType foint instead of a type 1. TrueType font embeddding and subsetting isn't as good as type 1. | 14:36.34 |
Laurenceb__ | is there any way to work out how much of the sizesize is due to fonts? | 14:37.31 |
kens | Acrobat used to be able to provide such information | 14:37.46 |
| I htnk it got dropped in the more recent versions. | 14:38.02 |
| That's Acrobat Pro, not Reader | 14:38.14 |
| Yeah its in Acrobt Pro 9, not in X | 14:39.37 |
Laurenceb__ | i see | 14:47.29 |
| any idea if there is a way to simply add Helvetica? | 14:47.45 |
| maybe make a new pdf container, add the previous pdf and include the font? | 14:48.01 |
kens | I cannot think of any easy and reliable way to do this with Ghostscript and pdfwrite its never been a design goal for that device | 14:50.10 |
Laurenceb__ | yeah | 14:50.17 |
| is there a sane way? | 14:50.26 |
kens | Well I would have thought what you were doing woudl work if you din't have Helvetica substituted with a TrueType font. | 14:50.53 |
| You could try setting -dSubsetFonts=false, which might help, maybe. | 14:51.10 |
Laurenceb__ | ok | 14:51.17 |
mvrhel_laptop | bbiab | 14:57.42 |
Laurenceb__ | nah its even larger | 15:00.09 |
| shrug | 15:00.13 |
| ill trade non embedded fonts for file size | 15:00.24 |
kens | Like I said, its probably caused by using a TT substitute | 15:00.30 |
kens2 | Hmm Robin_Watts ping | 15:32.02 |
Robin_Watts | pong | 15:32.11 |
kens2 | BUg #695954 does appear to be a real problem | 15:32.23 |
| I created a direcotry with an accented character d:\Tést and I can't us it on the command line | 15:32.46 |
| wchar_to_utf8 does produce 2 bytes for eh eacute | 15:33.12 |
Robin_Watts | as it should. | 15:33.19 |
kens2 | But it seems the OS can't read it | 15:33.20 |
| I'm not really sure why | 15:33.33 |
Robin_Watts | Breakpoint gp_open_file | 15:33.42 |
kens2 | OK 1 sec | 15:33.47 |
| Didn't get there | 15:34.06 |
Robin_Watts | gs_open_file_utf8 ? | 15:34.18 |
kens2 | ah could be | 15:34.24 |
Robin_Watts | (I'm pulling these names out of my memory, so...( | 15:34.29 |
kens2 | Doesn't seem like VS likes wither of those | 15:35.39 |
| eigher* | 15:35.43 |
| Good grief *either* | 15:35.51 |
Robin_Watts | gp_fopen | 15:36.55 |
kens2 | Ah, at least I can set a breask there :-) | 15:37.14 |
Robin_Watts | Hmm. | 15:37.52 |
kens2 | Didn't get there with that filename though | 15:38.11 |
Robin_Watts | kens2: Assign me the bug. I'll look at it as soon as I get out from under my current panic project. | 15:41.11 |
kens2 | I'll poke it a little more, I got to where the file gets opened | 15:41.26 |
| Seems we open it using the libfile syntax | 15:42.34 |
| And abort because file_status says it doesn't exist | 15:42.51 |
Robin_Watts | kens2: OK. If you look in gp_mswin.c you'll see gp_fopen | 15:43.19 |
kens2 | Uh OK one moment | 15:43.31 |
Robin_Watts | We assume the fname coming in is in utf8. | 15:43.41 |
| we therefore convert it to wchars from that, and feed it to _wfopen. | 15:43.58 |
kens2 | Oh yeah, I have a breakpoint in there already, we never got there | 15:44.03 |
| The file-_status call failed before we get that far | 15:44.19 |
Robin_Watts | I'm guessing that the libfile thing is missing that logic. | 15:44.23 |
kens2 | I'd guess os yes | 15:44.30 |
Robin_Watts | so if you can walk through and find out where it's going wrong, it might be an easy fix. | 15:44.49 |
kens2 | Yeah I'm doing that now | 15:45.00 |
| THough there are (of course!) at least 3 different routiens for ths.... | 15:45.18 |
| Yeah we're just doing a 'stat' in fname, which is the wchar form | 15:45.55 |
| Err I mean UTF8 of course | 15:46.59 |
Robin_Watts | kens2: Where is the stat call (if you know offhand) | 15:48.10 |
kens2 | gsiodev.c, os_status, at around line 229 | 15:48.30 |
Robin_Watts | right. I see that. | 15:48.43 |
| I think we should call gp_stat, probably, and define that as appropriate for the platform. | 15:49.01 |
kens2 | But I'm not sure where is the best place to alter ths. We get here from lib_file_open_search_with_no_combine... Yeah that was what I was thinking, more or less | 15:49.20 |
| It obviously needs to become platform dependent | 15:49.37 |
Robin_Watts | kens2: yeah. | 15:49.49 |
kens2 | Let me write htis into the bug | 15:49.51 |
Robin_Watts | Thanks. | 15:49.56 |
ManDay | Hi, I'd like to parse only a subset of pages. I found in google that gs -dFirstPage=... -dLastPage=... would allow me to do that. What kind of class these identifiers fall into and why is it not listed in the manpage for gs? Is there a manpage which covers that? | 15:54.31 |
kens2 | We don't do man pages. Look at our documentation for details on how to use GS. They aren't 'classes' they are command line parameters | 15:55.27 |
ManDay | kens2: well gs does have a manpage and it does document certain -s and -d special names. so i suspected "-dLastPage" falls into a different "class" which is not document there but elsewhere | 16:00.58 |
kens2 | THe GS documentation isn't really in the man pages, its in the HTML which we supply with Ghostscript. The command line switches are extensively documented. | 16:01.43 |
| So basically, don't use man pages, use our documentation | 16:02.13 |
Robin_Watts | ManDay: We do not supply a man page. It's possible your distribution has added manpages, in which case you should report the bug to them. | 16:02.17 |
kens2 | Robin_Watts : Actually, I thnk we *do* supply a (very simple) set of man pages | 16:02.33 |
Robin_Watts | kens2: Crap. Delete it now :) | 16:02.43 |
kens2 | I could be mistaken but I thnk chrisl said that at some time in the past. My opinion concurs, we should either make it a full set of docs or kill it. As far as making it into full docs, finger on nose right now | 16:03.35 |
ManDay | I see, /usr/share/doc/ghostscript-gpl-9.10-r2/html/Use.htm has it | 16:07.08 |
kens2 | If that's where you installed Ghostscript, then yes | 16:07.26 |
| Robin_Watts : the directory ghostpdl/gs/man seems to have the sources for the man pages. No ideas how to use that stuff though | 16:08.17 |
| And *I've* never updated any of it, so whatever's in there for pdfwrite is likely wildly out of date | 16:08.42 |
| Seems the last update was for 9.15 in September 2014 | 16:09.27 |
| Personally I think the man page should simply say 'for Ghostscript documentaton see....' | 16:10.24 |
Robin_Watts | kens2: I am inclined to agree. | 16:11.17 |
kens2 | Anyway, I'm off, have a good weekend all | 16:11.52 |
Robin_Watts | I could understand the argument for it having a basic set of information, but it should say "This documentation is incomplete and not definitive. For proper documentation see...." | 16:11.55 |
| kens2: Have a good one. | 16:11.59 |
mvrhel_laptop | Robin_Watts: how do we want to do help doc for gsview | 16:12.21 |
Robin_Watts | mvrhel_laptop: Write a PDF document with instructions, bundle it, load that as the startup doc. | 16:13.15 |
mvrhel_laptop | that is a good idea | 16:13.27 |
Robin_Watts | That would work best if you could share the doc between your version and freds, but ... :) | 16:15.59 |
ManDay | well thanks | 16:16.40 |
henrys | we supply a set of man and we've accepted a german set from a contributor. I hate to say that's all bogus. | 16:23.40 |
| we should probably be looking at pandoc or something similar to generate the man pages from the html. | 16:25.35 |
fredross-perry | we could share it by stashing it in the mupdf tree, which we both use. | 16:58.58 |
Robin_Watts | fredross-perry: Better to have a gsview_doc.git, probably as a submodule that both of you can include. | 17:28.06 |
fredross-perry | thatâs fine too | 17:28.19 |
Robin_Watts | henrys: I suspect that to use pandoc or something, we'd need to completely rewrite the html to be the right structure. | 17:35.39 |
| henrys: Probably just amending the man pages with a note saying: "This documentation is just an introduction. For more in depth/complete documentation see ..." | 17:36.21 |
henrys | Robin_Watts: yeah I'm sort of curious what the default output looks like. I might just try it and see. | 17:36.39 |
Robin_Watts | henrys: not sure how pandoc would cope with links between pages. | 17:37.01 |
henrys | Robin_Watts: I'd rather remove the man pages that say something like that, but yes we could do it. The problem is it's not an introduction, a lot of it is wrong. | 17:37.53 |
| s/that/than | 17:38.09 |
Robin_Watts | henrys: Oh, if it's wrong that's bad... | 17:38.24 |
henrys | to the agenda... | 17:41.28 |
mvrhel_laptop | I finally think I see what the issue is with this altona file on page 7 | 18:20.32 |
| at least part of it | 18:43.18 |
| Forward 1 day (to 2015/04/25)>>> | |