| <<<Back 1 day (to 2014/07/24) | 2014/07/25 |
kens | Morning Robin_Watts | 10:03.34 |
Robin_Watts | kens: Morning. | 11:01.59 |
| Were you after me for something? Or was it just that my internet dropped? | 11:02.14 |
kens | Both, I was wondering if someone had responded to the queries baout the android build and strtof | 11:02.33 |
| MuPDF of course | 11:02.50 |
| THere's a bug report, an SO query and 2 people here yesterday asking | 11:03.02 |
| Sadly I know zero about building for Android..... | 11:03.19 |
Robin_Watts | I didn't see any of that. | 11:10.50 |
kens | Hmm, let me look at the irc logs, one second | 11:11.02 |
| OK right at the start of the log yesterday, from callamits and then later (10:49) from govind_coll referencing bug #695381 | 11:12.02 |
| I can't now find the SO one, maybe I made that up | 11:12.47 |
Robin_Watts | Ta. will look when I get a mo. | 11:13.16 |
kens | No rush, I know you're off tomorrow, so feel free to pass it to someone else, I don't know who would be appropriate | 11:13.41 |
| Off to grab some lunch, bbl | 11:16.06 |
jogux | the android one is presumably a problem with building the dll but I can't imagine what :-S it'd probably be worth sending the guy a pre-built library to prove that. (or suggesting he extracts one from the apk on mupdf.com) | 11:32.04 |
Robin_Watts | tor8: ping | 13:20.10 |
| I think android simply doesn't provide strtof. | 13:20.46 |
kens | I was wondering if that was the case, I found a web page which claimed it was defined but not implemented | 13:21.31 |
Robin_Watts | oh, they weren't in older versions of the ndk, but they are in r5 and later. | 13:21.35 |
kens | That was going to be my next comment...... | 13:21.49 |
Robin_Watts | Given that the ndk is at v9 or higher now, I have little sympathy with him. | 13:22.00 |
| r10 in fact. | 13:22.21 |
kens | That I didn't know, so that sounds reasonable to me. | 13:22.22 |
| I'd suggets sticking that in the bug report and closing it | 13:22.55 |
| Wow, serious rain suddenly.... | 13:23.11 |
jogux | I'm relatively amazed it would actually build with r4 to be honest. | 13:23.17 |
kens | Oh and thunder,I guess that's why my network keeps dying | 13:23.25 |
jogux | although I guess mupdf has a less 'involved' compilation process than SOT :) | 13:23.36 |
Robin_Watts | mupdf is deliberately kept very simple in it's needs. | 13:24.59 |
jogux | to play devil's advocate: http://www.mupdf.com/docs/how-to-build-mupdf-for-android appears to imply any NDK will do :-) | 13:28.45 |
Robin_Watts | Yeah, well, blame tor8 for that. He said the version in platform/android/Readme.txt was too scary :) | 13:30.05 |
jogux | :-) | 13:30.39 |
| the README at the top of mupdf should perhaps include references on how to build for android / iOS / ... | 13:30.58 |
| ie. say "go look at platform/android/README.txt" | 13:31.38 |
Robin_Watts | TBF, if you google for the error message it's pretty obvious what the problem is. | 13:32.01 |
kens | THBH I thought the error message itself was reasonably clear | 13:32.25 |
| I just thought it weird that strtof wouldn't be present | 13:32.37 |
tor8 | Robin_Watts: well, it is! the people coming up with these questions obviously haven't read your instructions and don't have enough clue to follow mine :( | 13:35.17 |
kens | There's a *lot* of SO questions from obviously clueless people....... | 13:35.55 |
tor8 | SO needs a bigger clue stick to beat these people with... | 13:36.55 |
jogux | robin_watts: right, I suck then, but I googled it and failed to figure it out from the first few results I looked at :-) | 13:43.39 |
| I'd argue that tor8's instructions are missing the fact that ndk r6+ is required, though I'm fairly certain it wouldn't have helped if they had :-) | 13:44.34 |
tor8 | ndk r6 is several years out of date by today | 13:44.55 |
| jogux: but I can certainly add a sentence to that effect | 13:45.09 |
jogux | oh, indeed, no argument there :-) | 13:45.12 |
| kens: the more worrying tend is the upvoted SO answers from obviously clueless people :-( | 13:46.11 |
kens | :-) | 13:46.19 |
tor8 | jogux: thankfully, one upvoted answer will pay for ten downvotes in their ranking game. | 13:46.59 |
jogux | nods | 13:47.30 |
mattchz | In general, I think I was told that it was best to use the latest and greatest SDK/NDK. | 14:04.06 |
jogux | mattchz: excluding the broken Android L one? :-) | 14:10.37 |
mattchz | well yeah. Actually, you probably want to use the latest NDK, and whichever SDK the source is already configured to use. | 14:11.16 |
| although iâd hope Google have fixed the L sdk by now ;) | 14:11.51 |
cats | tor8: I sent you an email about this but I'll ask it here for completeness - what is the preferred way to extract field data from a fillable pdf using the mupdf api? | 14:41.20 |
tor8 | cats: I don't have time for a full answer right now (I need to go AFK soon) | 14:47.14 |
| the form field interfaces are still a work in progress, and setting up a convenient API for accessing this sort of data is on my TODO list | 14:47.39 |
| for now, you'll probably need to access them at the pdf_obj level | 14:47.53 |
cats | tor8: ok, cool. If you have the time, could you quickly show me how I'd bridge the gap between a pdf_document and pdf_obj? | 14:48.47 |
Robin_Watts | cats: Do you understand the structure of a PDF document? and what pdf objects mean within that context? | 14:49.50 |
cats | Robin_Watts: tbh, I'm not too clear on PDF document structure, nor on the subject of objects within one | 14:51.14 |
Robin_Watts | OK, well, it's not going to be a quick question then. | 14:51.33 |
| PDF documents are basically a list of objects. | 14:51.43 |
| together with some tags at the beginning/end of the file so you can recognise it as being a PDF file, and then get a list of the positions of the objects within the file. | 14:52.24 |
cats | ok, gotcha so far | 14:53.15 |
Robin_Watts | You can call 'pdf_trailer(doc)' to get a pointer to the trailer object. | 14:53.16 |
| Everything can then be traversed by looking up objects from the trailer object. | 14:53.42 |
| Basically objects in a PDF file are a directed graph. Mostly acyclic, but not always. | 14:54.11 |
| So from the trailer you can lookup 'Root' to get the root object, then from there 'Pages' will give you the top of the page tree etc. | 14:54.50 |
| somewhere hung off either Root or the Page tree entry will be the list of Fields. | 14:55.11 |
| each of which will be an object. | 14:55.19 |
cats | ah ok | 14:55.39 |
Robin_Watts | Read the sections of the pdf reference manual to do with file structure, objects and fields, and you'll be in a good place to start from. | 14:55.46 |
cats | awesome, thanks so much | 14:56.03 |
Robin_Watts | np. | 14:58.32 |
rayjj | cats: note that the Fields dictionary is accessed from the Root via the AcroForm dictionary (the Interactive Form Dictionary) See Section 8.6 of the PDF Ref Manual | 16:04.44 |
| AFAIK, there aren't any accessed via the Page objects | 16:05.15 |
cats | rayjj: perfect - thanks | 16:08.19 |
rayjj | darn, this might be tedious. The HEAD code handles the 2526 FTS OK even with my provisional small pattern clist band changes, but cust 532's code gets it wrong. Time to set some gs_debug flags and compare the output :-( | 16:08.54 |
| cats: Note, that I know PDF structure pretty well, but NOT mupdf (my primary area is ghostscript) | 16:09.40 |
| oh, better. I forgot another difference to cust 532's config. They default to a smaller MaxPatternBitmap, so it wasn't using pattern-clist. So it's my changes that don't work, but at least I can debug on our code | 16:43.10 |
kens | night all. | 16:59.47 |
mvrhel_laptop | brb | 17:54.45 |
henrys | turning off henrysx6 and macpro have air conditioning problems at home. | 18:16.13 |
e98 | any mupdf devs listening? | 20:12.13 |
rayjj | mvrhel_laptop: I have a newbie XAML question for you if you have some time. Please let me know (or call if I am online and unresponsive) | 22:28.01 |
| (obviously gsview related) | 22:35.24 |
| mvrhel_laptop: please check logs if you missed above ^^^ | 22:39.21 |
mvrhel_laptop | hi rayjj | 22:40.04 |
| what is it that you want to do? | 22:40.08 |
| changes to the xaml code can be tricky due to all the data bindings that exist | 22:40.26 |
rayjj | mvrhel_laptop: trying something that *should* be simple -- add a hotkey for PageDown PageUp | 22:40.58 |
mvrhel_laptop | I thought I already had that | 22:46.38 |
rayjj | mvrhel_laptop: after my change I am getting an Exception during Initialize_Components. My changes are at http://pastebin.com/SPZbBTgW | 22:47.28 |
mvrhel_laptop | but I guess not. I | 22:47.29 |
rayjj | mvrhel_laptop: the "Additional information" doesn't seem very useful "Provide value on 'System.Windows.Baml2006.TypeConverterMarkupExtension' threw an exception." Line number '132' and line position '10'. | 22:50.16 |
mvrhel_laptop | hmm. let me open up the project and take a look | 22:50.27 |
rayjj | mvrhel_laptop: if you are busy, that's fine. This is a "back burner" for me when I get fed up with cust 532 issues | 22:51.16 |
mvrhel_laptop | I need this as a back burner too. SOT has so many things that need to be done just in the chart area it is a bit overwhelming. | 22:51.55 |
rayjj | mvrhel_laptop: well, you could always work on Zoltan's "unique" (to put it kindly) issues ;-) | 22:53.21 |
mvrhel_laptop | I was really hoping someone would grab that one. | 22:53.52 |
| but I suspect it won't go away | 22:54.07 |
rayjj | seriously, though, I recommend staying away from him as much as possible. If I get inclined to work on his issues and need help, I'll holler. | 22:54.42 |
mvrhel_laptop | rayjj: thanks! I owe you one | 22:54.59 |
| actually I owe you several for all that you have helped me with over the years | 22:55.12 |
rayjj | mvrhel_laptop: np. Likewise | 22:55.27 |
| mvrhel_laptop: if you want to ignore my newbie attempt at adding the hotkey ans show me how you would do it (the easier way, no doubt) that's fine | 22:56.55 |
mvrhel_laptop | rayjj: I will look this over a bit. iirc the command bindings are "special" in some sense | 22:57.28 |
| we may need to do this in the way that I did the up and down arrow keys | 22:57.58 |
rayjj | mvrhel_laptop: from all the frustrated questioners on stackoverflow, I got that impression | 22:58.03 |
| mvrhel_laptop: oh, right. I didn't think of that. you use that to "scroll" | 22:58.40 |
mvrhel_laptop | yes | 22:58.45 |
| sigh. I am frustrated how quickly I forget what I wrote | 22:59.55 |
rayjj | mvrhel_laptop: I see the OnKeyDownHandler let me try hacking that. | 23:00.50 |
mvrhel_laptop | yes, that is the place to do it | 23:01.59 |
| then you can do a call to OnForwardPageClick or OnBackPageClick | 23:02.35 |
rayjj | mvrhel_laptop: OK. Thanks. I'll let you know. Have a good weekend | 23:02.42 |
mvrhel_laptop | rayjj: thanks, you too. have fun with c# | 23:02.54 |
rayjj | mvrhel_laptop: yeah, right :-/ | 23:03.10 |
mvrhel_laptop | ;) | 23:03.14 |
rayjj | am I having fun yet ? | 23:03.36 |
mvrhel_laptop | hehe | 23:03.39 |
rayjj | hi, marcosw: Sorry for the premature comment on the JPEG2000 bug. I'll blame it on a brain fart | 23:05.28 |
| I better go see about helping with dinner... | 23:06.40 |
| Forward 1 day (to 2014/07/26)>>> | |