| | <<<Back 1 day (to 2014/11/03) | 20141104 |
| henrychen | how can i draw free text annot not in horizon? | 03:25.42 |
| kens | chrisl ping | 11:52.18 |
| chrisl | kens: pong | 12:00.31 |
| kens | chrisl how does the PCL interpreter find fonts ? | 12:01.39 |
| | More specifically, if a PCL file uses CJKV fonts, can we define a TT replacement font ? | 12:02.20 |
| chrisl | No, we can't | 12:02.31 |
| kens | Ah, so if a PCL file is missing fonts, then basically, there's no way to render it ? | 12:03.17 |
| chrisl | It depends - you won't get sane results with a missing CJKV font.... | 12:04.23 |
| kens | OK I guess that's good enough for a SO answer. | 12:05.02 |
| chrisl | kens: Basically, if a font isn't embedded in the file, then we'll try to match ("best fit") with one of the built-in fonts. To the best of my knowledge, only PCL-5/J had anything other than a western font built in, and that was only a Japanese font. And we don't do 5/J. | 12:07.11 |
| | So, at best, you might get garbled characters out | 12:07.33 |
| kens | Yes, that's pretty much what O | 12:07.52 |
| | oops, "what I've said, but in less detail" | 12:08.06 |
| chrisl | Of course, we have the usual caveat of many vendors having their own customisations/hacks in their own implementations...... | 12:09.39 |
| kens | That's OK, in this case he's using GhostPDL and can't work out how to add fonts, he's used to using Ghostscript. | 12:14.19 |
| chrisl | Ah, for GhostPDL you'd need to edit the font list in the code.... it's not a run-time option like GS | 12:15.54 |
| | I have head out - will probably struggle to get back for the beginning of the meeting..... but should be back during/near it... | 12:55.25 |
| Robin_Watts | US hours have changed now, right? We're back at 3:00pm ? | 13:03.48 |
| kens | I believe so yes | 13:05.00 |
| | But 3:30 surely ? | 13:05.06 |
| Robin_Watts | kens: could be. I get confused. | 13:39.23 |
| | bear of tiny brain etc. | 13:39.32 |
| kens | Me too, which is why I asked :-) | 13:39.35 |
| henrys | kens: adding a font to pcl will have bad side effects for font selection. The font should be downloaded, it can be saved with PJL macros to make it permanent on the file system. | 13:45.39 |
| kens | Henrys, not a problem, I just wanted to answer a SO question | 13:46.04 |
| henrys | meeting at 3:30 gmt | 13:48.50 |
| | meeting time | 15:31.02 |
| kens | is here | 15:31.41 |
| henrys | mvrhel_laptop: looks like fred has made good progress with QT. Thoughts on going with it on windows? | 15:32.26 |
| | marcosw: is this ps2write in 9.14 support email falling through the cracks? | 15:33.04 |
| Robin_Watts | henrys: Does it work for TIFKAM and win32 ? | 15:33.30 |
| | and does it look like a windows app? | 15:33.41 |
| | History suggests that it'll look "kinda like a windows app if you squint a bit and ignore your peripheral vision" | 15:34.19 |
| | and that seems like a backward step when we have a windows version that works nicely. | 15:34.38 |
| chrisl | Hmm, is "looking like a Windows app" considered a "good thing"?? | 15:34.52 |
| | ;-) | 15:34.55 |
| Robin_Watts | chrisl: On windows, yes. :-p | 15:35.03 |
| marcosw | henrys: probably, I'll look into it. | 15:35.11 |
| mvrhel_laptop | morning. | 15:35.20 |
| henrys | Robin_Watts: qt has gotten quite a bit better but also many apps are going the qt route making users more accepting | 15:35.21 |
| mvrhel_laptop | sorry I am late | 15:35.22 |
| Robin_Watts | henrys: Ah. Train the users to like the dogfood. | 15:35.42 |
| mvrhel_laptop | henrys: I think I do like the native app a bit better | 15:36.02 |
| henrys | Robin_Watts: well they've already been trained to like windows... | 15:36.07 |
| Robin_Watts | Never mind the pain the wallet. Look at the shiny! | 15:36.13 |
| | oh, sorry, no, that's apple. | 15:36.17 |
| chrisl | I don't think QT have a "proper" Win8 port just now, anyway | 15:36.55 |
| mvrhel_laptop | also, at 100MB install with all the DLLs it seems a bit bloated | 15:37.02 |
| rayjj | 100Mb !!!! | 15:37.13 |
| mvrhel_laptop | hehe | 15:37.18 |
| henrys | mvrhel_laptop:oh I didn't notice that. | 15:37.40 |
| mvrhel_laptop | At least that was the size of the stuff that I unzipped | 15:38.01 |
| | maybe some of it is not needed | 15:38.07 |
| rayjj | for the complete SDK or just the Qt app ? | 15:38.11 |
| henrys | mvrhel_laptop: fine to stick with the original plan of native on windows and qt elsewhere I was just hoping we could avoid the extra maintenance. | 15:38.32 |
| mvrhel_laptop | henrys: I think it will not be that much more maintenance | 15:38.55 |
| henrys | kens: curious how firstpage/lastpage is going? had all the trapping of a hellish project but I hope I'm wrong. | 15:39.33 |
| mvrhel_laptop | rayjj: that size is due to all the dlls | 15:39.35 |
| tor8 | all Qt apps ship really big bloated with the DLLs | 15:40.13 |
| kens | henrys its hard, but getting there. Assuming I can make this all work it should also help with some other stuff, some of it on my plate too | 15:40.13 |
| tor8 | it's not as visible on Linux, where the bloat is usually pre-installed | 15:40.28 |
| henrys | mvrhel_laptop: so presumably if other qt apps were installed those would be shared or probably not? | 15:40.29 |
| Robin_Watts | henrys: Yes, because windows apps have such a great history of sharing dlls. | 15:40.50 |
| mvrhel_laptop | good question. it depends upon how the installer was set up | 15:41.02 |
| | in this case there was no installer | 15:41.09 |
| | which windows did not like | 15:41.12 |
| henrys | I fooled with it and didn't mind the interface much. But I'm fine with the current course. | 15:42.17 |
| rayjj | mvrhel_laptop: so unless we have a web based installer app that checks which DLL's already exist and only downloads those that aren't present, we are stuck with a bloated install package ? | 15:42.27 |
| | mvrhel_laptop: or can we 'prune' the DLL's down to the ones that we actually need ? | 15:42.47 |
| mvrhel_laptop | probably, but I suspect he already did that. his first package for me was missing some dlls | 15:43.14 |
| | or a dll I should say | 15:43.21 |
| | anyway. I vote that we stick with the native windows solution and use qt for mac and linux | 15:43.52 |
| | I will work with fred in getting things similar. we are really almost there. he just needs to add a few more capabilities | 15:44.22 |
| chrisl | mvrhel_laptop: note that you might have to remove a few capabilities..... did you have "save as XPS"? | 15:45.12 |
| mvrhel_laptop | yes I did | 15:45.43 |
| | we both should be able to do that | 15:46.01 |
| | using the xpswrite device | 15:46.18 |
| chrisl | So that should go since xpswrite doesn't produce XPS suitable for general consumption | 15:46.18 |
| henrys | well that brings up the next question: can we now do local cluster pushes for xpswrite changes marcosw ? | 15:46.19 |
| mvrhel_laptop | chrisl. oh yes but I am hoping that will change soon | 15:46.40 |
| | we need that for the printing | 15:46.43 |
| | I realize transparency wont work but if we can get text and images working | 15:47.08 |
| chrisl | mvrhel_laptop: even if we get it good enough for printing, it still won't work well for document exchange - no text | 15:47.10 |
| mvrhel_laptop | I think henrys has a solution for text | 15:47.22 |
| kens | may disappear.... | 15:47.28 |
| mvrhel_laptop | hehe | 15:47.32 |
| chrisl | But it won't be text | 15:47.33 |
| mvrhel_laptop | true. it will be paths | 15:47.42 |
| henrys | chrisl: I'm going to do something with text. | 15:47.43 |
| kens | carpet layer needs to move stuff by the front door, where the braodband router is | 15:47.50 |
| chrisl | So copy and paste, search etc will still not work in the xpswrite output | 15:48.07 |
| mvrhel_laptop | that is true | 15:48.17 |
| henrys | chrisl: once again I'm going to do something for text. | 15:48.21 |
| rayjj | chrisl: we can just blame it on Microsoft XPS ;-) | 15:48.37 |
| chrisl | henrys: I thought your plan was to use a patch cache? | 15:48.44 |
| | s/patch/path | 15:48.55 |
| Robin_Watts | Can we not put invisible text in to allow for cut/paste/search ? | 15:48.59 |
| mvrhel_laptop | that is an idea | 15:49.07 |
| henrys | chrisl: that is the right plan if we are going to restrict to printing. | 15:49.09 |
| chrisl | henrys: that's my point - not text | 15:49.20 |
| rayjj | Robin_Watts: the problem is getting the Unicode for the text -- we don't always have that | 15:49.25 |
| Robin_Watts | rayjj: Right. In the cases where we do. | 15:49.44 |
| henrys | mvrhel_laptop and fred have thrown "save as" on the table so I think we need text. | 15:49.47 |
| mvrhel_laptop | ok. well I feel we should do the best we can and the few nuts who want xps output will have to deal with it | 15:49.52 |
| henrys | real text if we do that. | 15:49.55 |
| rayjj | and that's why kens ran away -- dealing with the encoding to Unicode is a pain | 15:50.13 |
| Robin_Watts | People can't expect to be able to convert a file that can't be searched etc to one that can. | 15:50.14 |
| henrys | I did look at it and I don't think it'll be hard to get some common cases. Type 1 and TT and punt for the rest. | 15:50.43 |
| rayjj | Robin_Watts: unfortunately, since Adobe wrote the PSCRIPT5.DLL they know what the encoding tricks used are and are able to sus it our somehow | 15:51.15 |
| | henrys: CIDFonts are *very* common | 15:51.36 |
| chrisl | especially in PDFs these days | 15:51.51 |
| | But they are predominantly CIDFonts with TT outlines, so..... | 15:52.20 |
| henrys | rayjj: I'll look at it more this week if I get stuck I'll let folks know. | 15:53.17 |
| rayjj | chrisl: yes, but getting the correct Unicode from it (since the TT outlines are often subsets) is the issue | 15:53.19 |
| chrisl | rayjj: true, but we have the same restrictions as with pdfwrite and txtwrite | 15:53.54 |
| rayjj | and lots of apps can't seem to be bothered with writing ToUnicode | 15:53.57 |
| henrys | I don't think the bar is that high for xps consumers. They'll probably be surprised if it works at all ;-) | 15:54.57 |
| | anything else for this meeting? | 15:55.46 |
| kens | Not from me | 15:56.07 |
| Robin_Watts | Can bugzilla be made so that GhostDocs bugs are 'open' to people that are CCd on them? | 15:56.24 |
| | (maybe that should be the other meeting) | 15:56.35 |
| henrys | has anyone converted a pdf with text to xps using microsoft tools? what do they do? | 15:56.48 |
| kens | What MS Tools are those ? | 15:57.10 |
| chrisl | XPS printer? | 15:57.24 |
| henrys | kens: like printing to the xps writer or whatever they call it. | 15:57.34 |
| kens | I think I've done that before, but it depends almost totally on what the *printgin* application does | 15:58.08 |
| | If it sends a bitmap, then your XPS file will contain a bitmap | 15:58.21 |
| mvrhel_laptop | yes I have done that quite a bit | 15:58.27 |
| chrisl | Well, print from Acrobat would be the benchmark I'd think | 15:58.52 |
| mvrhel_laptop | cant say I tried any copy and paste or search of text afterwards | 15:58.55 |
| | I will try that today | 15:59.02 |
| kens | I can do it now, I have that set up already, just a moment | 15:59.15 |
| mvrhel_laptop | henrys: so I am off to boston tomorrow for the color imaging conference, back friday night and then off saturday morning to south korea until wed. | 15:59.47 |
| | I am probably going to spend most of today getting ready for that stuff | 16:00.02 |
| | and take care of a couple things around here before I go | 16:00.20 |
| | we had a wacky day yesterday as some fool set off a small chemical bomb at my daughters school which shut it down | 16:00.50 |
| henrys | mvrhel_laptop: okay have a good trip. | 16:00.59 |
| kens | Well the XPS file fomr that combination has fonts in it | 16:01.13 |
| henrys | cff's? | 16:01.54 |
| | can you email the xps kens? | 16:02.25 |
| kens | I didn't look that closely yet, I'm trying to find a viewer that will let me copy/paste text | 16:02.30 |
| | henrys, yes | 16:02.36 |
| | to just you or anyone else want it ? | 16:02.44 |
| henrys | kens: why not just put it on casper in your home directory? | 16:03.16 |
| kens | Copy/paste in the MS XPS viewer results in garbage | 16:03.16 |
| henrys | kens: I can just see it creating a bogus encoding and adding pieces of the font. | 16:03.54 |
| kens | The fonts all looked like subsets, they have weird names | 16:04.15 |
| henrys | if ms doesn't do I wonder about that utility of us doing it. | 16:04.34 |
| chrisl | I guess it just pulls them directly from the PDF..... | 16:04.48 |
| henrys | and is there even a viewer that would work properly if the encoding was correct and all the info was there? | 16:05.11 |
| kens | ok test.xps on casper in my home directory now. Its a 9 page PDF with text in it,originally | 16:05.50 |
| chrisl | XPS will just rely on the encoding in the font - I don't think there's provision for anything like a ToUnicode map | 16:05.51 |
| Robin_Watts | henrys: MuPDF, maybe :) | 16:05.59 |
| kens | FWIW Acrobat coudl search the text in the original document | 16:06.51 |
| tor8 | chrisl: XPS text can have explicit and separate unicode and glyph index lists | 16:06.59 |
| kens | also copy/paste from the original PDF works | 16:07.17 |
| tor8 | and fall back to the encoding in the font | 16:07.27 |
| chrisl | tor8: Oh, yes, it's done inline isn't it? Rather than a separate mapping like PDF | 16:07.42 |
| henrys | well if no tools expect fonts to be useful there is no point in us doing text other than to possibly save space with caching. | 16:08.12 |
| chrisl | And rendering speed..... | 16:08.30 |
| henrys | chrisl: right | 16:10.51 |
| | marcosw: is yosemite working well for you? Can I upgrade macpro? | 17:35.58 |
| Robin_Watts | There is a thing about the upgrade itself. It will take a very long time if you have much in /usr/local/bin | 17:47.53 |
| latho | Hello. Is somebody here, who can answer me a question about mupdf? | 18:25.15 |
| sebras | latho: there are, please do ask the questions and be prepared to wait though. :-/ | 18:35.27 |
| Robin_Watts | sebras: Too slow. | 18:35.38 |
| sebras | latho: better than logging out 3mins later. | 18:35.45 |
| | Robin_Watts: I know. | 18:35.47 |
| sebras | likes to talk to ghosts... | 18:35.59 |
| avih | tor8: should the arguments object be useable as an argument to another functions i.e. function x(){return processArgs(arguments);} ? | 18:58.35 |
| | e.g. * | 18:59.18 |
| henrys | Robin_Watts: thanks was worried about breaking the cluster mainly, but I bet marcosw has been using it. | 19:08.25 |
| tor8 | avih: not directly if you mean to call another function with the same arguments you recieved just forwarded | 21:07.40 |
| avih | tor8: so maybe throw a type error on such case? | 21:08.20 |
| tor8 | avih: what you wrote just passes the arguments array as an array | 21:08.39 |
| | maybe you're looking for return processArgs.apply(null, arguments) | 21:08.57 |
| avih | tor8: hpw so? when does it get converted? | 21:09.00 |
| tor8 | 'arguments' is an array object that contains the arguments that were passed | 21:09.25 |
| avih | tor8: the spec says it's an "array-like" object which can be convrted to an array using slice. | 21:10.09 |
| | i.e. not necessarily an array. so you say that in mujs it's a real array? | 21:10.37 |
| tor8 | it is an unclassed object that behaves like an array but does not have [[Prototype]] set to Array.prototype | 21:11.08 |
| avih | so should it be possible to pass it as an argument like function x does in my example? | 21:11.42 |
| tor8 | yes. | 21:11.49 |
| avih | and processArgs should slice it if it wants a real array? | 21:12.10 |
| tor8 | why shouldn't it? | 21:12.12 |
| | yes. you'll need to convert to an Array object if you want a real array | 21:12.43 |
| | again, it's the stupid JS spec, not much I can do to help there. | 21:12.56 |
| avih | iirtc i had a crash with usage similar to that. no specific str yet but i'm checking first what should and shouldn't be possible | 21:12.56 |
| | iirc* | 21:13.08 |
| | k, thx for the info. (not touching the js project these days... got some real work...) | 21:14.19 |
| tor8 | so, argsAsArray = Array.prototype.slice.call(arguments) | 21:14.55 |
| avih | indeed. that's what i've been using at the calee | 21:15.14 |
| | tor8: | 21:18.32 |
| | > function pa(args){args = Array.prototype.slice(args); print (args);} function x(){pa(arguments)}; | 21:18.44 |
| | > x(1,2,3) | 21:18.44 |
| | > | 21:18.44 |
| | (empty output) | 21:19.01 |
| | ah. call. | 21:19.18 |
| | works with the call. if i get an STR i'll let you know. thx. | 21:20.24 |
| tor8 | avih: yeah, otherwise you slice the slice Array.prototype | 21:45.55 |
| | Forward 1 day (to 2014/11/05)>>> | |
IRC Logs