| | <<<Back 1 day (to 2014/01/23) | 2014/01/24 |
| ray_laptop | mvrhel_laptop: I (finally) made my reservations. I get in Friday about 5pm, and leave Sunday night at about 8pm | 00:03.00 |
| mvrhel_laptop | ray_laptop: Sounds like you will be riding with henrys and me then | 00:03.18 |
| ray_laptop | You guys are coming in late Firday ? I thought henrys was getting in on the 6th (checking email) | 00:04.12 |
| | yep. henrys is on Thursday the 6th | 00:05.09 |
| | mvrhel_laptop: I don't have your flight info (or it has a different Subject) | 00:05.58 |
| mvrhel_laptop | oh then no | 00:06.30 |
| | I am coming in late thursday like henrys | 00:06.40 |
| | ray_laptop ^^ | 00:06.55 |
| ray_laptop | mvrhel_laptop: OK. So I'm the last one in | 00:07.16 |
| mvrhel_laptop | sounds like it | 00:07.28 |
| ray_laptop | I hope Scott saves me some BBQ :-) | 00:07.34 |
| mvrhel_laptop | don't know about marcos2 | 00:07.38 |
| | marcosw | 00:07.38 |
| ray_laptop | mvrhel_laptop: he's probably counting on driving in Sat AM :-) | 00:08.26 |
| marcosw | ray_laptop and mvrhel_laptop: I'm coming in Friday evening, so will miss the bbq. | 05:32.13 |
| mvrhel_laptop | marcosw: bummer | 05:32.35 |
| mke | hi | 09:15.59 |
| | does someone know how to print with mupdf? | 09:16.36 |
| kens | On what OS ? | 09:16.48 |
| mke | linux | 09:17.28 |
| kens | Hmm, I suspect the answer is 'you cna't' at present | 09:17.40 |
| | Ermm 'you can't' | 09:17.51 |
| | I could be mistaken, you might want to wait an hour or two until the MuPDF developers come online for a more definitive answer | 09:18.20 |
| mke | thats no problem | 09:19.21 |
| kens | henrys, Robin_Watts I haven't looked at adding Bates numbering to PDF. Robin_Watts knows *much* more about it than I do. I infer from Robin's burblings that its some kind of annotation, in which case we could add it in easily enough I guess, but he would have to add relevant pdfmarks (or we'd have to write PostScript to do it for him, more likely). Remobing it would be trivial, we just don't pass tha tannotaion through. Reading | 09:25.57 |
| | it back would be possible, but its not a 'normal' use for Ghostscript and I don't see under what circumstances he wants to. | 09:25.57 |
| | I don't see that MuPDF would need pdfmarks if it generates the Bates numbers itself. Not sure where it stands with writing PDF files, I thought Robin had that working. | 09:26.41 |
| | So... I can look at it further if required, but I really need to work on this gsprint stuff for a day or two I think. | 09:28.04 |
| | For the logs; does anyone have a real physical printer with two different size paper sources connected to a Windows system ? I need one for testing gsprint..... | 09:38.28 |
| | Robin_Watts : when you get online can you give me a ping please | 10:42.31 |
| Robin_Watts | ping :) | 10:42.39 |
| kens | ROFL | 10:42.43 |
| | How up to speed are you on this Bates numbering stuff ? | 10:42.56 |
| | It looks to me like its possible to do it without signatures, but rather pointless since you can't tell otherwise if the document has been modified. | 10:43.32 |
| Robin_Watts | kens: Only what I looked at yesterday. | 10:43.32 |
| kens | Oh, not much more than me then, did you try making a PDF with numbering ? | 10:43.47 |
| Robin_Watts | Bates numbering is not done as an annotation. | 10:43.50 |
| kens | No, I see that | 10:43.55 |
| | Its some XML and a PieceInfo | 10:44.04 |
| Robin_Watts | It's done by putting a "/Artifact <</Contents (RJWPRE000002RJWPOST)/Subtype /BatesN /Type /Pagination >>BDC" in the stream, together with a call to a form Xobject. | 10:44.28 |
| | Then an "EMC" | 10:44.36 |
| | The form XObject has some XML in as you say. | 10:44.56 |
| kens | Hmm, the docs say that should eba PiceInfo dict | 10:45.00 |
| Robin_Watts | There is a PieceInfo in the Form XObject. | 10:45.19 |
| kens | Odd, I thought it should be in the page dict, I'll have to read that again | 10:45.39 |
| Robin_Watts | /PieceInfo << | 10:45.43 |
| | /ADBE_CompoundType << | 10:45.45 |
| | /DocSettings 7 0 R | 10:45.47 |
| | /LastModified (D:20140123153516Z) | 10:45.48 |
| | /Private /BatesN | 10:45.50 |
| | >> | 10:45.51 |
| | >> | 10:45.53 |
| kens | I guess I'll go make me a PDF | 10:45.59 |
| | To be honest, the signature stuff looks harder than the bates numbering | 10:46.41 |
| Robin_Watts | signatures are hard, yes. | 10:47.58 |
| kens | I don't really see the point in Bates numbering without signatures, but I'm sure someone will say I'm wrong | 10:48.27 |
| Robin_Watts | Fundamentally, AIUI, we cannot do signatures properly with gs. | 10:48.32 |
| | To do signatures you need to work on an existing pdf file. | 10:48.44 |
| kens | Well we certainly can't at the moment | 10:48.48 |
| Robin_Watts | You cannot reconstruct a file from scratch as we do in gs. | 10:49.00 |
| kens | Robin_Watts : I can post-process the PDF, just like linearisation | 10:49.05 |
| Robin_Watts | No, you can't. | 10:49.14 |
| kens | Hmm, why not ? Its an existing PDF by then | 10:49.27 |
| | Or more accurately, it can be | 10:49.48 |
| Robin_Watts | If I have a PDF and I apply a signaure to it, that signature depends on an MD5 (or something similar) of byte ranges of the entire PDF except for the signature itself. | 10:50.21 |
| kens | OK | 10:50.30 |
| Robin_Watts | Now I pass the file to you, and you want to add your signature to it. | 10:50.59 |
| kens | Well true we won'tr be able to sign an existing document | 10:51.25 |
| Robin_Watts | You have to do so without changing any of my bytes. | 10:51.28 |
| | right. | 10:51.30 |
| kens | I'm talking about signing a newly created document | 10:51.45 |
| | If the customer wants to sign existing documents we could still do it, but not with pdfwrite. | 10:52.01 |
| | And it wouldn't make any sense to do it with GS | 10:52.21 |
| Robin_Watts | right. | 10:52.31 |
| kens | Is that what they want ? Its unclear to me | 10:52.45 |
| | If so then it seems a simple separate tool is all they need | 10:52.58 |
| | for large values of 'simple' | 10:53.11 |
| kens | has been ignoring this one and hoping it would go away | 10:53.48 |
| Robin_Watts | Yes, you could potentially sign a newly created document. | 10:53.53 |
| | but I fear that the standard requirement is to sign existing documents, repeatedly. | 10:54.09 |
| | and to check signatures. | 10:54.20 |
| kens | Seems to me this is not an area GS should be involved in, we are not a general purpose PDF amnipulation tool, as I keep telling people..... | 10:54.46 |
| Robin_Watts | I can't see how it's possible to add bates numbers to a doc without breaking the signatures. | 10:54.48 |
| kens | Robin_Watts : Correct, but I could create a *new* document wiht Bates numbers and then sign it :-) | 10:55.06 |
| Robin_Watts | kens: I agree that gs should not be involved in signing. | 10:55.11 |
| | In the absence of a complete pdfwrite device in mupdf, removing bates numbers (reliably) is hard. | 10:55.59 |
| kens | Hmm. | 10:56.16 |
| | Have to take your word on that | 10:56.28 |
| | But he doesn't actually ask for that as far as I can tell. (reading back through the thread now) | 10:56.47 |
| Robin_Watts | He does. | 10:56.54 |
| | I listed 4 things, one of which was removing numbers, and he said he wanted all 4. | 10:57.14 |
| | IIRC. | 10:57.17 |
| kens | Oh, that's not what the email from him I'm reading says. | 10:57.26 |
| | I suspect (given the customer) he just said 'yes' | 10:57.37 |
| Robin_Watts | See my mail of 16/7/13 at 18:00 | 10:58.15 |
| kens | In his mail on 15th Jan he says they need to sign documents, verigfy a signed document, add Bates numebring and read Bates numbers | 10:58.22 |
| Robin_Watts | and his reply 5 hours later. | 10:58.41 |
| kens | Robin_Watts : yes, I recall the mail, but I suspect he says 'I'll take everything you offer me' without thinking about what they need | 10:59.11 |
| Robin_Watts | I made the effort to ask explicitly. | 10:59.12 |
| | Regardless of his need, to offer a complete solution (coming anywhere close to what Acrobat can do) we need all 4. | 10:59.44 |
| kens | I get the impression Acrobat only removes Bates numbers incidentally | 11:00.10 |
| | If he doesn't need it, it could get him satisfied before you are ready to offer a complete solution | 11:00.33 |
| Robin_Watts | Docuement -> Advanced Processing -> Bates Numbering -> remove | 11:00.44 |
| | kens: Well, again, I suspect not. | 11:01.03 |
| kens | OK, I hadn't seen that, I was reading the Bates doc | 11:01.07 |
| Robin_Watts | The purpose of bates numbering, AIUI, is that a PDF file can be submitted as a legal document. | 11:01.32 |
| kens | OK in point 4 you say 'editing' rather than specifically deleting. Might be worth checking (I assume editing is easier then deletion) | 11:01.46 |
| | Robin_Watts : yes, which is why I can't see the point in Bates numbering it wqithout signing it | 11:02.09 |
| Robin_Watts | and the court/lawyers add 'Bates numbers' to it to identify it for the records and ensure it won't be changed. | 11:02.12 |
| | Hence it's important (IMHO) that the numbers are added to an existing document (with all the indexing etc that may be there) rather than being added to a 'reconstructed' document. | 11:02.48 |
| | i.e. the metadata (outlines, thumbnails, etc) are important too. | 11:03.19 |
| kens | Hmm, I'm not certain I agree. Its entirely possible to scan lots of docs and make a PDF from them, and want Bates numbering added immediately | 11:03.22 |
| | But nevertheless, its still sounds like its not somethign we wNT TO DO WITH gs | 11:03.39 |
| Robin_Watts | I have an idea for a way to do this in mupdf. | 11:04.18 |
| kens | Fair enough, the only way to do it in GS would be PostScript programming and I don;'t want to get involved in that. | 11:04.44 |
| | (for existing dsocuments that is) | 11:04.54 |
| Robin_Watts | Currently we parse pdf operators out of the stream, and call "pdf_run_XX" for each operator XX. | 11:05.41 |
| | I want to turn that into an interface, so XX is a number and we call pdf_run[XX] within an array of function pointers. | 11:06.21 |
| | Then I can write a new set of pdf_run[XX] operators that pass through to the original ones, possibly eliding content. | 11:07.14 |
| kens | Well, it sounds plausible, I don't know enough about MuPDF to have an opinion. | 11:07.43 |
| Robin_Watts | That way I can operate on operator streams without having to break them down to graphical objects. | 11:08.28 |
| | I'd done at least part of the job before. | 11:08.42 |
| | I suspect i can get something up and running within a day or two. | 11:08.53 |
| kens | We 'could' do that iwht the PDF interpreter in GS, but I shudder to even consider it | 11:09.05 |
| Robin_Watts | programming in postscript is just nuts. | 11:09.21 |
| kens | Ist not that, its our PDF interpreter which is nuts..... | 11:09.42 |
| Robin_Watts | Who had the mooscript stuff dumped on them at the last staff meeting? | 11:09.43 |
| kens | finger on nose | 11:09.51 |
| | Possibly chrisl | 11:09.59 |
| Robin_Watts | I wouldn't mind having a quick look at it if it has no other victim in its sights. | 11:10.23 |
| kens | I'm sure chrislk won't say no even if it is his | 11:10.35 |
| | I cna't see anything in workflowy | 11:12.24 |
| | THough I did just discover I'm supposed to look at PDF/VT (oops, guess I'd better do that) | 11:12.52 |
| | OK I found it under 'someday projects' but there's nobody assigned to it | 11:14.05 |
| | SO I guess if you want it Robin, its yours :-) | 11:14.25 |
| | Anyway, thanks for the discussion. I'm going to duck out of Bates numbering again, the log here should explain why using GS isn't a good option and won't satisy the customer request | 11:14.52 |
| Robin_Watts | ok. | 11:17.06 |
| mke | guess i implemented the printing myselfÂ… its ugly, but it works. | 11:34.38 |
| chrisl | kens, Robin_Watts: sorry phone. I have to look at the mooscript build, get rid of the manual final link part | 11:39.25 |
| ray_laptop | morning, all | 15:32.50 |
| kens | Hi ray | 15:33.16 |
| Robin_Watts | morning ray | 15:47.09 |
| henrys | hi ray_laptop | 15:53.15 |
| paulgardiner | Hi Ray Henry | 15:55.19 |
| henrys | ray_laptop:did you see the health net law suit stuff? | 16:06.47 |
| ray_laptop | henrys: no | 16:24.39 |
| henrys | apparently member's identities were stolen and there is a class action again them, they sent me a postcard asking if I wanted updated by email. I expect they'll send my SS and other personal info by email next ;-) | 16:26.17 |
| | paulgardiner: did anybody check if marcos office files were broken in smart office display or if it was a pdf generation problems? | 16:29.05 |
| Robin_Watts | henrys: Some of the differences were layout, which you have to imagine is going to be at least partially down to mismatching fonts. | 16:30.54 |
| paulgardiner | No. We can generate page bitmaps for each test, rather than PDF. That wont change the pagination, but may show up some of the missing stuff. | 16:30.58 |
| | In fact, I can remember noticing, during other tests, noticing the pdf missing some aspect visible in the bitmaps. | 16:32.09 |
| | Robin_Watts: yeah, I thought it might have been font mismatch | 16:34.42 |
| henrys | paulgardiner: the plan is to buy a test suite from QL, actually targeted to test printers, which tests all (?) basic functionality of office. We should set up marcos so he can discriminate between what is at pdf generation bug, bitmap, display etc. | 16:38.17 |
| Robin_Watts | henrys: We can do him an exe that he can call that will either dump bitmaps or pdf as he wants. | 16:38.52 |
| | A windows exe like that is easy. | 16:38.59 |
| henrys | Robin_Watts: perfect | 16:39.06 |
| Robin_Watts | a linux exe should be possible too. | 16:39.10 |
| | but I don't think either paul or I have actually tried building on linux yet. | 16:40.06 |
| paulgardiner | Or a Brew exe, or a PalmOS exe, or a Wince exe. :-) | 16:40.07 |
| Robin_Watts | paulgardiner: I'd be reluctant to promise what I hadn't tested. If it needs python 2.4 on linux too, that might require installing the correct python version. | 16:40.55 |
| henrys | I'm betting a windows exe will work under wine on mac os x and marcosw will like that okay | 16:41.10 |
| paulgardiner | Yeah Windows is most sensible to rely on | 16:41.19 |
| henrys | Robin_Watts: it needs python to build or run? | 16:41.51 |
| Robin_Watts | build | 16:41.56 |
| | the FBS is written in python. | 16:42.03 |
| ray_laptop | old python | 16:42.09 |
| henrys | Robin_Watts: I think most distros support both pythons there is way too much stuff that depends on 2. | 16:46.46 |
| Robin_Watts | Right, but there is python 2.7 | 16:47.02 |
| | and on windows at least, the FBS doesn't work under 2.7 | 16:47.11 |
| henrys | Robin_Watts: oh | 16:47.19 |
| Robin_Watts | I have no idea what the problem is, not could my spies tell me offhand. | 16:47.39 |
| | It may be that it's a windows only thing, but until I've actually tested a build under linux, I'd not want to promise it. | 16:48.16 |
| henrys | chrisl: for http://bugs.ghostscript.com/show_bug.cgi?id=691213 - urw wants a link to the fonts to which the bug refers. | 16:49.29 |
| chrisl | henrys: I'd expect URW to access to their own fonts..... | 16:51.03 |
| henrys | chrisl: I don't think they are very good with release numbers | 16:51.36 |
| chrisl | henrys: well, I don't think we have the 136 fonts set available for download anywhere at the moment | 16:52.15 |
| henrys | chrisl: we probably should put them up somewhere but we have to be careful to emphasize their AFPL'ness | 16:54.38 |
| chrisl | henrys: I wasn't aware they were AFPL, I thought they were commercial only | 16:55.11 |
| henrys | chrisl: yikes I didn't think that. You may be right. | 16:55.51 |
| ray_laptop | I am pretty sure they are commercial only | 16:55.55 |
| chrisl | We should probably clear that up, then | 16:56.09 |
| henrys | chrisl: it does seem odd to me they would release the pcl 80 AFPL and not the rest but I could be mistaken | 16:59.29 |
| paulgardiner | Well, thanks to Henry's nice idea, Robin's brave attack on the FBS and a little fiddling from me, the Android app now opens Office documents. | 17:02.09 |
| henrys | paulgardiner: wow that was fast | 17:03.18 |
| | paulgardiner: let's get it to mvrhel_laptop so he can be ready for a demo in Japan! | 17:03.57 |
| paulgardiner | Okay, I'll upload the APK somewhere and post a link. It's had only 5min testing of course. | 17:04.42 |
| Robin_Watts | paulgardiner: Did you solve the mystery problem then? | 17:06.32 |
| | Or is this using the debug non-rom build? | 17:06.46 |
| paulgardiner | No afraid not. I get the app to squirt a build of tar onto the app and then pipe an archive of GhostDocs to it. | 17:07.39 |
| henrys | chrisl: I asked miles to review the paperwork. | 17:08.14 |
| Robin_Watts | A while ago, I asked a question on stack overflow. I have some answers now. | 17:08.23 |
| | paulgardiner: This might be useful to the android app: http://stackoverflow.com/questions/9215631/multiple-file-extension-mimetype-intent-filters-with-one-activity/20664445?noredirect=1#20664445 | 17:08.33 |
| henrys | chrisl: I guess if it is commercial we could put it up with our commercial release with a password. | 17:08.48 |
| mvrhel_laptop | paulgardiner: nice! | 17:08.56 |
| chrisl | henrys: thanks. I can put the fonts in a password protected download directory on casper | 17:08.58 |
| Robin_Watts | In particular if we want to update the apk to respond to clicks on office docs. | 17:09.13 |
| henrys | chrisl: oh like in public_html, yes | 17:09.42 |
| chrisl | Yeh | 17:09.47 |
| henrys | chrisl: why don't we just do that now and I'll send it along to them. | 17:10.15 |
| Robin_Watts | paulgardiner: Nice work, btw. | 17:10.27 |
| chrisl | henrys: I'll do it before I finish tonight | 17:10.31 |
| henrys | chrisl: thanks | 17:10.42 |
| paulgardiner | apk is here: http://intranet.glidos.net/~paul/MuPDFOffice20140124.apk | 17:15.00 |
| Robin_Watts | paulgardiner: Random thought... the exe in that is unprotected, right? | 17:18.28 |
| | So anyone could copy it out of the apk and then reuse it? | 17:18.45 |
| paulgardiner | Yes, possibly readable off the device. | 17:19.16 |
| Robin_Watts | possibly we should do some dumb DRM. (eor it with "CopyrightArtifex" or something) | 17:19.45 |
| paulgardiner | Maybe I should move the apk onto casper | 17:19.47 |
| Robin_Watts | and possibly we should put the apk on casper, yes :) | 17:19.54 |
| | you typed faster than me :) | 17:20.00 |
| paulgardiner | It's now not downloadable. I'll move it | 17:21.15 |
| | It's now on casper in my home directory | 17:26.11 |
| Robin_Watts | paulgardiner: Does it respond to clicks on office docs? | 17:27.43 |
| | or do you need to start it, and then pick a doc from the list ? | 17:28.21 |
| paulgardiner | Should do. I added all the gubbins. I've tested only from the picker though | 17:29.01 |
| Robin_Watts | paulgardiner: My point is that I don't think the gubbins worked as was there already. | 17:29.23 |
| | in particular xps files couldn't be picked from fileexplorer last time I checked. | 17:29.34 |
| | which is what my stackoverflow question (above) was about. | 17:29.44 |
| paulgardiner | Works from ES File Explorer | 17:29.54 |
| Robin_Watts | oh, cool. Ignore me then :) | 17:30.13 |
| paulgardiner | I think we got it working, although I'm not sure how | 17:30.13 |
| | One problem we had was some file explorers not generating correct intents, so we had to tell them how to do it. :-) | 17:35.08 |
| Robin_Watts | paulgardiner: Fair enough. | 17:38.40 |
| | Ah, do files work from the browser? | 17:39.06 |
| | rereading my stackoverflow question, that may have been the problem. | 17:39.17 |
| paulgardiner | I think it works from the browser sometimes but not others. The gubbins works and gets the app recognised as a potential viewer, but I think there are still some content uris we don't handle. | 17:40.41 |
| | We may be able to fix that with your openBuffer stuff. | 17:41.01 |
| | or was it openStream? | 17:41.25 |
| | Or did that fix the problem already? I can't remember | 17:41.43 |
| Robin_Watts | or just by tweaking the xml as in the answer to the stackoverflow question. | 17:41.51 |
| paulgardiner | Ah get it now. I reread it. That's worth doing in any case because it's neater. | 17:45.24 |
| | Not sure it'll affect the use of content uris, but you never know. | 17:45.57 |
| ray_laptop | ah ha. I've been digging into a problem with the using the clist to render pdf14 (pdf14 clist accumulation) and I just happened to run ppmraw and it turns out the difference is there between clist and non-clist mode as well. | 18:10.12 |
| | I guess I'll open a bug on it. | 18:10.25 |
| | page mode matches Adobe. Clist mode is wrong. | 18:18.13 |
| henrys | Robin_Watts: have a look at your email if you've not looked recently | 18:20.20 |
| | Robin_Watts: ah nvm | 18:20.44 |
| chrisl | henrys: I've put an archive with the fonts, in a password protected directory on casper: the details are in a private comment in the bug | 18:21.50 |
| henrys | thanks chrisl | 18:22.16 |
| mvrhel_laptop | bbiab | 18:30.19 |
| ray_laptop | hurray. At least one difference is a progression Bug693480.pdf (to ps2write). | 18:36.35 |
| Robin_Watts | heads out. I'll be on the laptop for the next couple of days. | 18:54.23 |
| | henrys: yes. interesting mail. :) | 18:57.36 |
| bencc | can I do something like this http://stackoverflow.com/a/4484293/189815 | 19:08.46 |
| | to add video annotation? | 19:08.52 |
| palfvin | Trying to rotate PDF output with gs for non-90 degree amount, output = png. Should there be any issues with that? | 22:14.43 |
| | Using gs -dSAFER -dBATCH -dNOPAUSE -r200 -sDEVICE=pngmono -dAutoRotatePages=/None -sOutputFile=output.png -c 10 rotate -f input.pdf | 22:16.19 |
| | outputs fine, but without any rotation | 22:17.08 |
| robin_watts_mac | palfin: (For the logs) gs only rotates to orthogonals. | 23:20.52 |
| | Forward 1 day (to 2014/01/25)>>> | |
IRC Logs