| <<<Back 1 day (to 2014/05/15) | 2014/05/16 |
purb | hiii | 05:36.39 |
| how can i use WM_PAINT in viewproc.... | 05:38.40 |
| using WM_PAINT can we create annotation? | 05:39.04 |
| hiii | 05:58.28 |
| how can i use pdf_create_annot() to create a new annotation? | 06:24.42 |
Robin_Watts | purb: What sort of annotation? | 06:57.26 |
| See platform/android/jni/mupdf.c for various different ways to call it. | 06:57.43 |
purb | just want to draw a line....... | 06:59.07 |
Robin_Watts | See MuPDFCore_addInkAnnotationInternal in mupdf.c then. | 06:59.57 |
purb | I want to create an annotation in windows platform | 07:02.45 |
Robin_Watts | So ignore the android specific JNI code, and look at the fz_create_annot and subsequent code. | 07:04.18 |
purb | in windows code? | 07:08.43 |
| in window there does not exist any fz_create_annot code... | 07:09.35 |
Robin_Watts | purb: fz_create_annot exists in the central mupdf library core on all platforms. | 07:12.22 |
| There is no windows viewer code that calls it. | 07:12.31 |
| That, presumably, is what you are trying to add. | 07:12.45 |
| I am pointing you at existing code that DOES call it. | 07:13.04 |
| albeit on a different platform. | 07:13.14 |
sebras | Robin_Watts: good morning, Robin the operator? | 07:32.49 |
Robin_Watts | sebras: indeed. Do you need a hard line? | 07:33.10 |
sebras | Robin_Watts: no. I have one of my own -> -------------------------------------- | 07:33.36 |
purb | there exist fz_get_annot_type, fz_first_annot, fz_next_annot and fz_bound_annot.......... But can't see fz_create_annot function. | 07:38.20 |
tor8 | so many ops! | 09:19.43 |
| rayjj: google code has discontinued downloads, so we can't add any new ones. that's why there's no mupdf 1.4 on there. | 09:20.32 |
chrisl | tor8: I thought he might have been looking at downloads.ghostscript.com | 09:22.48 |
tor8 | chrisl: he mentioned mupdf 1.3 being marked as "Deprecated" which is a google code downloads feature- | 09:34.53 |
chrisl | Ah, okay. | 09:35.43 |
tor8 | http://mupdf.com/downloads/archive/ is where we store all the old (and current) versions | 09:36.42 |
| kens: have you got a minute for a gs PDF interpreter question? | 09:37.51 |
kens | Sure, I'm going bonkers with matrices, I coulsd use a break | 09:38.05 |
tor8 | I'm looking at CATX1385.pdf (from the PDFIA subset) highlight annotations | 09:38.28 |
| they look different in MuPDF and gs | 09:38.33 |
kens | OK let me look at the file | 09:38.38 |
tor8 | mupdf draws the highlight outlined, whereas gs has it filled | 09:38.57 |
kens | Hmm acrobat has it filled also | 09:39.12 |
tor8 | but from what I can read of the appearance stream, it looks like it should be stroked | 09:39.20 |
| evince has it filled as well | 09:39.26 |
kens | Hmmm... ok I'll have to decompress and read it, just a moment | 09:39.42 |
| THis is 'Deere's new software..' right ? | 09:40.26 |
tor8 | yeah | 09:40.32 |
kens | Well it looks stroked from the appearance stream | 09:40.55 |
| Of course Acrobat ignores that | 09:41.08 |
tor8 | does GS ignore it as well? | 09:41.15 |
| basically my question -- is there appearance synthesis going on or have I missed some part of the annotation stream? | 09:41.37 |
kens | No, if the appearance is there we should use it. If its not, then we 'might' make one up. | 09:41.39 |
| tor8 I can't see that its synthesising it, I'll stick some probes intot the interpreter and look | 09:41.56 |
| GIve me a few minutes | 09:42.02 |
tor8 | if I remove the AP in mupdf we synthesize it (but much too short) | 09:42.54 |
| no change to gs's behaviour though | 09:43.01 |
kens | Really ? THat would suggest we are synthesising it | 09:43.15 |
tor8 | the rounded corners are odd though, I wouldn't have expected those to come from synthesis | 09:43.48 |
kens | It looks like we do synthesise highlights in GS | 09:44.14 |
| If you look in pdf_draw.ps | 09:44.29 |
| and search for Highlight | 09:44.35 |
| You'll see we don't load the appearance stream | 09:44.43 |
| The comment says that Acrobat ignores AP and always uses transparent highlighting | 09:45.07 |
| Ah, if you turn off transparency you may see the same result -dNOTRANSPARENCY | 09:45.37 |
tor8 | kens: yep! | 09:45.57 |
| that's it | 09:46.02 |
kens | So it seems GS always synthesise highlights (yuck) presumably in order to match Acrobat | 09:46.20 |
tor8 | with -dNOTRANSPARENCY gs matches mupdf's AP rendering | 09:46.31 |
kens | That's co-incidence, we're still sunthesising, but we're using a stroke because we cna't use transparency | 09:46.55 |
tor8 | and I guess highlight-arc is the thing to synthesize the rounded edges? | 09:46.55 |
kens | I believe so, yes | 09:47.03 |
tor8 | yeah, that sounds right. if I remove the AP and run with -dNOTRANSPARENCY gs synthesizes something that matches the original AP | 09:47.41 |
kens | Its all kind of icky, but so is the fact that Acrobat ignores appearance streams altogether | 09:47.45 |
tor8 | okay, so now I can close that part of the bug :) | 09:47.57 |
kens | tor8 yeah, we *never* use the /AP for highlights it seems | 09:48.01 |
| Just like Acrobat..... | 09:48.15 |
tor8 | but maybe I should let paul know to fix the synthesis, because our own synthesis is coming out half width | 09:48.17 |
kens | Oh, well yes that sounds like a Paul thing | 09:48.31 |
| Ah, another exellent Indian bug report there I see | 09:49.17 |
| Tor, can you tell me whathttp://www.appendit.se do ? | 09:50.10 |
| Oh never mind, I found the ENglish icon | 09:50.23 |
tor8 | outsourced app development | 09:50.30 |
| for mobiles | 09:50.38 |
kens | Hmm, and it looks like they've outsourced that to India :-) | 09:50.59 |
tor8 | yeah, they even have a +91 country code phone number :) | 09:51.22 |
kens | ROFL | 09:51.29 |
| They should probably reconsider their translation service provider too: | 09:52.00 |
| "Since we reuse code and have low personnel costs, we get great price / presentation tooth" | 09:52.00 |
tor8 | I 100% expect them to have a swedish sales office and do all the work in india | 09:52.02 |
kens | Yeah, but oddly appendit.in redirected me to appendit.se | 09:52.24 |
tor8 | their head quarters are on the outskirts of stockholm | 09:52.28 |
kens | Oh right, they really are Swedish then. Well I'll wait and see if the reporter can manage to amend their bug so it makes some sense | 09:52.55 |
tor8 | "Get Touch With Us" :D | 09:53.27 |
kens | Oooh | 09:53.36 |
tor8 | http://www.appendit.se/appendit_new/RightSideContent.aspx?pageid=3 nice portfolio links they have... | 09:53.58 |
kens | Server Error in '/' application ? | 09:54.18 |
tor8 | yeah. | 09:55.19 |
kens | I'm not even sure how you got there, the right side links take me to pageid=1, 2 and 13 | 09:55.47 |
| err, 1, 12 and 13 | 09:56.07 |
tor8 | the scrolling banner thing in the bottom left of the main page under "Our Portfolio" | 09:56.32 |
kens | Oh yeah, I never thought to click that | 09:56.54 |
| Definitely broken though, maybe I shoudl send them a 'critical' bug report | 09:57.18 |
| I see their 'Products' page is 'under construction', which doesn't exactly fill me with confidence | 09:58.08 |
| "We are currently hiring full-time professionals in the areas of website design" :-) | 09:58.34 |
| OK enough hilarity back to patterns | 09:59.08 |
tor8 | they certainly look like they need them :) | 09:59.09 |
purb | hiii | 11:06.52 |
Robin_Watts | purb: Hi. If you have a question, it's best on irc to ask it rather than to wait for a response to a "hi" as you might be waiting a long time :) | 11:11.33 |
tor8 | Robin_Watts: could we program ghostbot to say that instead of replying with "hi"? | 11:12.38 |
Robin_Watts | tor8: I have pondered on that before. | 11:12.52 |
kens | ghostbot is not infallible, it misses things like hiii | 11:13.03 |
tor8 | and also reply "can I ask a question?" or "I have a question" with our usual replies. | 11:13.08 |
| (or, if it has some sort of timer, anything with "question" from a newly logged in person?) | 11:14.05 |
kens | we could put all that in the intro, but then nobody reads that.... | 11:14.41 |
| we need Google's context aware NLP processor | 11:17.21 |
Robin_Watts | hello | 11:21.56 |
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. | 11:21.56 |
kens | well that works | 11:22.14 |
purb | I can't find fz_create_annot function | 11:28.22 |
tor8 | prub: it's pdf_create_annot | 11:29.35 |
| purb: to create annotations you need to work with a pdf_document instead of the fz_document | 11:30.00 |
| purb: use pdfdoc = pdf_specifics(fzdoc) to convert | 11:30.35 |
| it will return NULL if the fz_document is not a PDF file | 11:30.48 |
Robin_Watts | as I said, this can all be seen in MuPDFCore_addInkAnnotationInternal within platform/android/jni/mupdf.c | 11:31.54 |
tor8 | Robin_Watts: I think there's a case to be made for writing a "mutool annot" to show/add/delete/edit annotations | 11:33.17 |
purb | ok.... at first I will study the mupdf.c | 11:33.18 |
Robin_Watts | tor8: There is a case for adding 4 more hours to the day too :) | 11:33.44 |
purb | where the how/add/delete/edit annotations parts? | 11:34.01 |
tor8 | Robin_Watts: or maybe that should just be exposed by some scripting language bindings | 11:34.03 |
Robin_Watts | but yes, if you can think of a nice command line interface, that would be a nice tool. | 11:34.11 |
kens | pdfmark :-) | 11:34.56 |
Robin_Watts | purb: There are 2 uses of pdf_create_annot in that file. The first is for adding highlight/underline/strikeout annotations. | 11:35.21 |
| The second is for adding ink annotations. | 11:35.26 |
tor8 | Robin_Watts: I was going to say, having lua bindings to mupdf would make scripting stuff like that nice | 11:35.35 |
| but now we have mujs ... but ick, javascript! | 11:35.41 |
Robin_Watts | purb: pdf_delete_annot is used in just one place in that file. | 11:36.55 |
kens | lunches | 11:39.04 |
purb | tor8: is there any case made for writing a "mutool annot" to show/add/delete/edit annotations? | 11:40.37 |
tor8 | purb: "making a case for" means "I wish we had" :) | 11:43.04 |
| Robin_Watts: DUMP_GROUP_BLENDS has suffered bitrot ... error: compression error -2 | 12:14.56 |
| anyway, I've simplified my nested tiling pattern woes down into a plain 'mask' problem | 12:18.05 |
| I suspect I've hit a snag with the complicated state handling for transparency groups :( | 12:18.32 |
Robin_Watts | So it seems WebOS has now moved to LG. | 13:10.49 |
| Can we sue them too? :) | 13:10.56 |
kens | Hi Michael | 14:22.05 |
rayjj | morning, all | 14:32.47 |
| mvrhel_laptop: thanks for the comment on cust 532's transparency issues. I have the first almost fixed (09_0920.pdf) and will go on to the 2524 next. | 14:33.38 |
| mvrhel_laptop: I won't bother you unless I get wrapped up in it, but they are such simple files, they are easy to work with (compared to ATS files) | 14:34.32 |
| kens: I see you are having the usual fun with a not very informed or informative free user (missing images) | 14:35.35 |
kens | rayjj yes pretty much | 14:35.47 |
rayjj | finally the heat wave is supposed to ease off today :-) | 14:37.45 |
tor8 | Robin_Watts: two fixes on tor/master, would be nice to have them in before the weekend | 14:37.52 |
rayjj | tor8: and Robin_Watts: is mupdf 1.4 released anywhere ? | 14:38.13 |
henrys | rayjj: I hear police are going around ticketing folks for watering. | 14:38.44 |
tor8 | rayjj: yes, but not on the google code project site | 14:38.45 |
| rayjj: they disabled downloads, so I've hidden the downloads page on google code with a page saying where to go | 14:39.50 |
| but if you still have an old bookmark, I expect you'll still see the old downloads list | 14:40.07 |
| rayjj: mupdf.com/downloads/ | 14:40.18 |
rayjj | henrys: that's why we water in the wee hours. It's also more effective since there's less evaporation | 14:45.41 |
henrys | rayjj: and the police are sleeping ;-) | 14:46.21 |
rayjj | tor8: I see them at mupdf.com/download but the links on the main page take you to the google code page https://code.google.com/p/mupdf/downloads/list?q=OpSys%3DWindows which says it has been moved to a new location, but that link goes to https://code.google.com/p/mupdf/wiki/Downloads?tm=2 | 14:53.39 |
tor8 | rayjj: that's odd ... maybe you're seeing a cached main page? | 14:54.28 |
rayjj | tor8: then that page says: Go to the official MuPDF page instead and that link takes you back to the main page (a roundabout) | 14:54.30 |
tor8 | rayjj: yeah, that sounds like a roundabout! but the main page *should* be pointing to mupdf.com/downloads, not the google code page, for quite some time now | 14:55.12 |
rayjj | tor8: nope -- I just tried it with chrome (which I rarely use) and it wraps around too | 14:55.35 |
tor8 | rayjj: ohhh!!! I see what I forgot. | 14:56.09 |
rayjj | tor8: but I got an updated version the hard way (building it myself) | 14:56.22 |
tor8 | the link in the side bar points to the right place, but not the ones on the page >.< | 14:56.23 |
| rayjj: okay, I've fixed the main page links now too. | 14:58.11 |
rayjj | tor8: np. Thanks. It's nice having somebody do it that doesn't do it often to do it a bit differently. We should have Scott try it :-) | 14:58.17 |
tor8 | rayjj: yeah, it's easy to be blind to this sort of thing. | 14:58.42 |
rayjj | tor8: I notice that download and downloads both go to the same place -- linked, I presume ? (I just wanted to make sure they aren't two different directories) | 14:59.44 |
tor8 | yeah, /download/ is just a symlink | 15:02.38 |
rayjj | kens: is it safe to rely on BT / ET being present ? I check for a bare SMask at the BT, push a trans group, and end the group in the ET, but if ET is not reliable ... | 15:02.45 |
kens | how do you mean rely on thieir presence ? | 15:03.29 |
| OH, ET is mandatory | 15:03.38 |
| even Acrobat gets *really* Upset if its missing | 15:03.51 |
rayjj | kens: will there always be an ET for the BT (and does text ALWAYS have BT .. ET ) ? | 15:04.07 |
tor8 | rayjj: I've seen bad PDF generators that don't wrap text in BT/ET (and even more put illegal operators inside the BT/ET pair) | 15:05.03 |
kens | illeghal ops inside BT/ET can happen | 15:05.30 |
| I've never seen BT without an ET | 15:05.45 |
| I suppose I could try making one.... | 15:06.01 |
tor8 | yeah, I hope that if there is a BT you can safely assume there will be an ET | 15:06.14 |
rayjj | kens: I see a comment in pdf_ops that "The text state can be set even outside BT/ET." | 15:06.18 |
kens | rayjj yes, but that's not quite the same | 15:06.32 |
tor8 | but we don't make any assumptions about the presence of BT/ET in mupdf except to treat arrays differently inside a BT/ET pair (as an optimization for running TJ) | 15:06.55 |
rayjj | kens: right, and that won't bother me since I am only concerned about when the text is actually painted, since the trans group needs to be in place by then | 15:07.23 |
kens | that's what I thought | 15:07.42 |
rayjj | kens: thanks | 15:09.17 |
mvrhel_laptop | morning | 15:10.59 |
kens | mvrhel_laptop : I installed gsview on my laptop and got the same problem. I'll be bringing it to the staff meeting so you cna have a play of you like | 15:12.21 |
mvrhel_laptop | kens: ok. very strange | 15:12.39 |
henrys | kens: are you the only one seeing this? | 15:24.47 |
kens | the installation problem / I htink Ray said he had hte same problem | 15:25.01 |
rayjj | mvrhel_laptop: I get Mupdf DLL not found | 15:39.42 |
| there are two mupdf dll's in the gsview folder | 15:40.03 |
mvrhel_laptop | ok. you did not do the right installer then | 15:40.19 |
kens | rayjj you should try the newer installer | 15:40.20 |
mvrhel_laptop | I guess I should start putting a sub/sub version on these | 15:40.52 |
| again you might want to wait a bit. fixing a few more things today, then I will create a new one | 15:42.16 |
rayjj | kens: downloading whatever is on mvrhel's gsview now | 15:43.40 |
kens | Will be interested to hear your experience with that one | 15:43.55 |
rayjj | mvrhel_laptop: and I'll let you know if it's any different | 15:43.58 |
mvrhel_laptop | if it is doing the install shield style installer vs. the basic windows one then you have the wrong one | 15:45.11 |
| install shield had some "smarts" to it where it was scanning for dependencies in the code and determining what DLLs to insert. Unfortunately, you can't override that unless you buy the $$ version. | 15:46.19 |
| hence, I moved to the older installer | 15:46.55 |
rayjj | mvrhel_laptop: OK. It's installing 4.5 .NET (I thought that was standard on Win 7) | 15:47.10 |
mvrhel_laptop | win 8 | 15:47.30 |
rayjj | mvrhel_laptop: I see. It must be pretty big (taking FOREVER) | 15:47.54 |
mvrhel_laptop | or you have a little straw... | 15:48.04 |
kens | Its only 13Mb | 15:48.27 |
| chrisl already downloaded it, but the .NET download is taking him a little longer..... | 15:48.45 |
rayjj | mvrhel_laptop: well I probably just don't have adequate negative pressure on the straw to counterattack the suction from MS :-) | 15:48.58 |
mvrhel_laptop | :) | 15:49.06 |
kens | misread it *is* the .NET that's taking ray a long time | 15:49.34 |
mvrhel_laptop | so, I have a version that is coming that includes drag and drop, stores some state information in the registry (so your previous file will open), allows AA to be disabled, comes up as an option when you to OpenWith for all the file types that it handles, and fixes scrolling issues and a pile of minor bugs that I found. | 15:55.17 |
| I need to fix a couple more things first though | 15:55.51 |
kens | It all sounds good, pity it won't work for me :( | 15:57.34 |
| good night all have a good weekend | 16:19.03 |
Robin_Watts | night kens | 16:22.59 |
mvrhel_laptop | bbiaw | 17:22.29 |
rayjj | mvrhel_laptop: I tried the Install64 I downloaded earlier today, and the installer looks different, but I still get MuPF DLL not found | 19:06.43 |
mvrhel_laptop | rayjj: that is odd. ok. how many files do you see in the installed directory? | 19:17.10 |
rayjj | mvrhel_laptop: sorry I was composing an email ... | 19:44.38 |
| 14366720 May 9 09:19 gsdll64.dll* | 19:45.29 |
| 1167360 May 12 12:42 gsview.exe* | 19:45.30 |
| 8800256 May 12 10:54 mupdfnet64.dll* | 19:45.32 |
| mvrhel_laptop: I'm runniing Win 64 Home Premium SP 1 | 19:49.22 |
mvrhel_laptop | ray_laptop: that all looks correct. I suspect there is something in my 64 bit testing that is causing some issue. This is likely the same issue that I am seeing with ken's machine. At the staff meeting maybe I can figure it out. I also should do some testing here on my other machines | 20:23.29 |
Guest10400 | hi need some help on gs or pcl6 to print barcode from pcl files | 22:39.27 |
| using ghostpdl 9.14 version | 22:40.19 |
JakeSays | hey i'm trying to determine the bounding rectangle of images in a pdf, and its just not making sense. can someone take a look at this code and see if it looks right? http://www.privatepaste.com/ce90244986 | 23:04.55 |
| for example, for an image of size 30, 100 i get the bounding box of {x0=608.399963 y0=508.800049 x1=609.479980 y1=520.80049} | 23:07.32 |
| which is saying the image is ~1 point wide | 23:08.27 |
| Forward 1 day (to 2014/05/17)>>> | |