| | <<<Back 1 day (to 2014/09/24) | 2014/09/25 |
| marky_ | Does anyone know if ghostscript is supposed to use return codes on linux ($?) at all? I don't seem to get any and can't seem to find anywhere that says it should. | 00:59.43 |
| norbertj | rayjj: for the logs, did you have a look at the attachment in 694750 (remember multithread-capability)? | 07:36.21 |
| | rayjj: another peculiarity: in gxclthrd.c teardown_device_and_mem_for_thread(), lines430-444. if (bfile != NULL) then fclose(cfile) !!!! and vice-versa. Is this intentional? | 07:37.55 |
| dreamer | ohai | 08:18.58 |
| | we are having a bunch of issues with a TTF and before I bug the makers of the font I'd like to know if there are some things I can figure out beforehand | 08:19.55 |
| | the font is called Lato and ghostscript gives me the following error: GPL Ghostscript 9.07: Failed to interpret TT instructions in font Lato-RegularFID126HGSet2. Continue ignoring instructions of the font. | 08:20.15 |
| kens | Well, you can run various tools on fht font. Microsoft's font validation tool is one route | 08:20.27 |
| dreamer | nothing on *NIX? | 08:21.01 |
| kens | First hting I would suggest is you use an ip to dat eversion of Ghostscript, 9.07 is not the latest | 08:21.11 |
| | Current version is 9.15 | 08:21.21 |
| chrisl_r500 | dreamer: how does the font render? Usually ignoring the TT hinting instructions is pretty benign | 08:21.57 |
| kens | On Linux you can use FontForge, but its not strictly a validation tool and may silently fix problems. | 08:22.01 |
| dreamer | we've tried on newer versions as well. this happens to be on Xubuntu 13.04, but 14.04 was tried as well (uses 9.10 .. also not newest I see) | 08:22.13 |
| kens | No0, not even close | 08:22.21 |
| | You could open a bug report, we're probably better placed to investiate such problems | 08:22.51 |
| dreamer | chrisl_r500: so, using it in libreoffice atm. looks fine on the screen. a pdf looks fine as well, but our printers die on it. unless we print with adobe-reader | 08:22.52 |
| | printing from windows seems to work as well, but our organisation mostly uses linux | 08:23.05 |
| | this error is when I export to .ps from libreoffice, then run ps2pdf on the file | 08:23.30 |
| chrisl_r500 | What you put above is not an error, it's a warning..... | 08:23.53 |
| kens | Indeed | 08:24.01 |
| | chrisl : I thought you were on holiday today...... | 08:24.23 |
| chrisl_r500 | Just making up for time I missed yesterday - finishing in six minutes...... | 08:24.52 |
| kens | :-) | 08:25.01 |
| dreamer | chrisl_r500: ah indeed, there is output I now see. looks fine, but I can't print it. | 08:25.20 |
| kens | dreamer : Then there could be several problems. What kind of printers are you using ? Are you using CUPS to print to them ? What errors do you get form the pritners ? | 08:26.03 |
| dreamer | ok, tis early here .. I'm obviously not completely awake so going at this the wrong way: we think that our print-issues have something to do with this warning, as it relates to the specific font that is giving problems. now I'd like to find out why we get this warning | 08:26.05 |
| chrisl_r500 | Well, it does sound like the font is broken - if your printer and freetype can't interpret the instructions, it's probably broken | 08:26.08 |
| dreamer | kens: cups, the printers don't even recieve the document | 08:26.24 |
| | cups doesn't give any errors | 08:26.29 |
| | it claims to have printed the docs | 08:26.37 |
| kens | Hmm, sounds like you might do better starting with debugging cups | 08:26.41 |
| dreamer | we only get it with this specific font | 08:27.01 |
| kens | Probably CUPS is using Ghostscript to either render the files, or convert them to a pritner format (eg PostScript). | 08:27.34 |
| chrisl_r500 | dreamer: You can try things like ftlint and ftdump from freetype - they might help you track down the problem | 08:27.45 |
| kens | You are going to need to start by finding out what CUPS is doing | 08:27.45 |
| | If you have a PostScript printer, you cna try printing the PostScript you produced by sending it directly to the printer. | 08:28.50 |
| dreamer | my colleague has tried using a different driver for our printers and then does get paper out of the printer with the message 'error: invalidfont offending command: awidthshow' | 08:29.32 |
| | so, meh | 08:29.37 |
| | chrisl_r500: I'll look in to those. thnx | 08:29.43 |
| kens | Well, that tells you the font is broken | 08:29.45 |
| | Its a bit difficutl to discuss this in such vague terms. When we talk about this we need to know what Page Description Language the printer accepts (ESC/P, PostScript, PDF, PCL, XPS etc) what PDL you are sending to the printer, how the file was produced, and how you are sending it. | 08:31.02 |
| | 'a different driver' doesn't tell us much | 08:31.22 |
| dreamer | kens: yeah sorry. I don't know all these specific things atm. just wanted to know what this warning was | 08:31.39 |
| kens | THe warning is just that, a warning. But interpreters differ, and sometimes we ignore things that other interpreters might choose not to. | 08:32.11 |
| | FWIW it does sound like your font has a problem | 08:32.21 |
| dreamer | uhuh, looking at ftlint and ftdump now | 08:32.32 |
| chrisl_r500 | dreamer: Fonts have a concept called "hinting" which is intended to give "hints" to renders to improve display, mainly at low resolutions or very small text sizes. | 08:33.03 |
| dreamer | chrisl_r500: how do I use ftlint? | 08:33.19 |
| chrisl_r500 | dreamer: read the manual? It's not something I use | 08:33.46 |
| dreamer | No manual entry for ftlint | 08:33.59 |
| chrisl_r500 | ftlint: Usage: ftlint ppem fontname[.ttf|.ttc] [fontname2..] | 08:34.41 |
| dreamer | yes, I have no idea what 'ppem' is | 08:35.22 |
| chrisl_r500 | An integer - it's the size it will use to test | 08:35.57 |
| kens | Grr, network hiccuped.... | 08:36.51 |
| dreamer | chrisl_r500: tried a bunch. all say 'Lato-Regular.ttf: OK.' | 08:37.25 |
| kens | dreamer : If you make a copy of the font available to me I can run it through the MS font validation tool and send you a report | 08:37.32 |
| chrisl_r500 | As I was saying, hinting: truetype does hinting using a bytecode program which the renderer can choose to interpret or not. The warning you get from Ghostscript says we've tried to execute the bytecode program, and it has caused an error. So we'll carry on without it | 08:37.54 |
| kens | (assuming I cna find my copy of the tool) | 08:38.01 |
| dreamer | kens: http://mrtoasted.com/~dreamer/Lato-Regular.ttf | 08:38.50 |
| kens | OK give me a minute, I found the tool | 08:39.03 |
| dreamer | cool | 08:39.08 |
| chrisl | Hrm, okay, simply moving the file names into heap memory doesn't solve the Windows performance problem - I'll need to pick up again tomorrow..... | 08:39.15 |
| | Bye! | 08:39.18 |
| kens | bb chrisl | 08:39.23 |
| | THe font has a few warnings, a few errors, at least one is essentially catastrophic, one of the gl;yphs has intersecting contours | 08:41.32 |
| | Oh and the GASP table is incorrect | 08:41.44 |
| dreamer | hmmm | 08:41.49 |
| | I don't know what these things mean, but it sounds bad | 08:41.58 |
| kens | *lots* of glyphs with the same problems | 08:42.09 |
| | THe HEAD tbale has some problems too (unitsperem is invalid) | 08:42.26 |
| | Now, how do I save this report..... | 08:42.38 |
| dreamer | what is the tool you used? | 08:43.04 |
| | maybe it works under wine? | 08:43.09 |
| kens | MS FOnt Validator | 08:43.13 |
| | Probably it works on wine yes | 08:43.20 |
| | Oh, the rasterisation tests might not, they may depend on the MS TT font rasteriser | 08:43.36 |
| dreamer | https://appdb.winehq.org/objectManager.php?sClass=application&iId=9587 | 08:44.31 |
| | I'll try | 08:44.34 |
| kens | Well if they say it works.... | 08:45.35 |
| dreamer | 'bronze' status | 08:45.43 |
| kens | Probably good enough ofr your purposes, you can always tunr off the rasterisation tests | 08:46.00 |
| | Ah, here's the puppy | 08:46.18 |
| dreamer | it installs, but doesn't run. meh | 08:46.27 |
| | puppy? | 08:46.43 |
| kens | During the rasterisation tests, there was an exception 'CVT number out of range' CVT = Control Value Table, and is vital for hinting | 08:46.50 |
| | So, basically, the font is broken | 08:47.04 |
| dreamer | uhg | 08:47.10 |
| | I'll send an email to the developers | 08:47.17 |
| kens | GS gets away with it by ignoring the broken hinting | 08:47.22 |
| | Other interpreters may or may not do so | 08:47.32 |
| dreamer | and talk to the colleagues that thought it was a good idea to use this font in the new house-style .. | 08:47.48 |
| kens | THe trouble is that if you ignore this kind of error the output may be incorrect, many users would prefer an error to incorrect output, especially if it involves printing thousands of glossy brochures and not finding out there's a problem till the customer complains..... | 08:48.28 |
| dreamer | we have seen prints where the kerning was off | 08:49.01 |
| kens | Bear with me a couple of minutes, I'm just trying to get a report out | 08:49.06 |
| dreamer | so that could be related to the broken hinting | 08:49.07 |
| | cool, that would be nice | 08:49.16 |
| kens | Possibly, 'hinting' in TrueType is a bit more than in other font technologies and font foundries have been known to abuse it | 08:49.36 |
| | Aha I have a PDF file, can I email it to you ? (obfuscate your address or go private chat so its not picked up by scammers) | 08:51.19 |
| dreamer | hah | 08:51.35 |
| | you can dcc it ;) | 08:51.40 |
| kens | Umm, maybe. | 08:51.46 |
| dreamer | I'll pm my mail | 08:51.51 |
| kens | OK that's easier | 08:51.57 |
| | (for me) | 08:52.02 |
| dreamer | hmm, just downloaded the font again from their site. and all the filesizes are different .. maybe they secretly did a bug-fix .. | 08:52.28 |
| kens | THat's possible of course, amybe someone else complained | 08:53.16 |
| dreamer | hah, and no more warning in ghostscript .. this is eerie :P | 08:54.04 |
| | let me send you the same font | 08:54.09 |
| kens | OK go ahead, the original report is on its way | 08:54.19 |
| dreamer | mrtoasted.com/~dreamer/Lato-Regular2.ttf | 08:54.44 |
| | cool | 08:54.49 |
| kens | Hmm, well it gets much further, its actually rasterising the glyphs this time | 08:56.16 |
| kens | should have limited the sizes before starting | 08:56.53 |
| dreamer | heh | 08:57.36 |
| kens | Its up to 15 point now | 08:57.52 |
| | I don't recall how far it goes | 08:58.05 |
| | I'd have to say the fact that it actually rasterises indicates that its definitely a changed font, and probably been fixed. You may be able to use this version without problems | 08:58.55 |
| | Sigh 25 point and counting | 08:59.49 |
| | I'd cancel it but I'm not sure that ouwld stop the whole validation | 09:00.07 |
| | I cancelled it it was going to test 4-72 at 1 point increments, I don't have all day :-) | 09:02.07 |
| | There are some warnings, but they are OK, some Unicode error, but that's not important (a range with no glyphs in it) | 09:03.15 |
| | Still a GASP error (wrong version number, and some invalid flags) | 09:03.38 |
| | Contour intersections are fixed | 09:03.52 |
| | Yes, its generally much better | 09:04.13 |
| | dreamer : second report on its way, the font looks OK to me this time | 09:05.44 |
| dreamer | sweet. thnx | 09:06.37 |
| | what a suprise :') | 09:06.43 |
| kens | I guess osmeone already told them their font was broken | 09:06.54 |
| dreamer | kens: uhuh. must've happened in the last few months. good thing I bothered to look haha | 09:08.24 |
| kens | :-) | 09:08.29 |
| dreamer | I'll see if it's worth it to send your new report as well, if there are more errors in there | 09:08.39 |
| kens | If you'd complained they'd probably just have told you it was fixed, so no big deal I guess | 09:08.48 |
| | There are errors and warnings, but they are pretty much harmless, I've seen fonts from MS bundled in Windows with more warnings than that. Its a pretty picky tool | 09:09.29 |
| | Which is why I use it of course :-) | 09:09.41 |
| dreamer | hah | 09:09.49 |
| | too bad it doesn't run with wine | 09:10.01 |
| kens | Yeah that's a shame | 09:10.08 |
| | I'd guess its the rasterisation that breaks it | 09:10.31 |
| dreamer | well thnx a lot for helping with this :) | 09:13.15 |
| kens | NP | 09:13.19 |
| dreamer | I see you develop MuPDF. was using that recently on my netbook :) | 09:13.37 |
| kens | Not me personally, but the MuPDF developers are here too :-) | 09:13.53 |
| dreamer | well ok, good busy anyway | 09:14.05 |
| kens | well, one of them is at the moment | 09:14.08 |
| | BB | 09:14.11 |
| | Robin_Watts : your email thread, they could insist on PDF/A files, that includes all fonts | 09:15.20 |
| | But I'd guess your comment is correct anyway :-D | 09:15.55 |
| amyn | I have an application developed in Visual Studio 2010 (C#/C++). In that application, I want to add the functionality of converting PDF to Images. Currently what my application does is run the following command using Process Class in C#. | 11:04.07 |
| | "gs9.14\\bin\\gswin32c -Igs9.14\\lib;fonts -dSAFER -dNumRenderingThreads=2 -dBATCH -dUseCropBox -dNOPAUSE -sPDFPassword=\"\" -sDEVICE=png16m -r200 -sOutputFile=\"output\" >> \"output\\conv.tmp\" 2>> \"output\\conv.err\" " | 11:04.12 |
| | this is time consuming as i have to start the Process again and again for different files. so now i am planning to embed ghostscript with my application | 11:05.11 |
| kens | You are aware of the AGPL restrictions ? Is your application going to be open source ? | 11:05.37 |
| amyn | i only us the above command so can any one tell me on how to go further with embedding GS with my application? | 11:05.57 |
| kens | You need to use the GS API | 11:06.11 |
| | Are you planning on using the DLL or a static library ? | 11:06.32 |
| amyn | Regarding the license, the application is for a comany. they will be using it for their internal computers only. it will not be sold to anyone else. it will be used internally only | 11:06.44 |
| | what would you suggest? what would be less time consuming and easy to understand and implement? | 11:07.06 |
| kens | Either way you should read api.htm, and look at the existing Ghostscript code on Windows, which already uses the Ghostscript DLL | 11:07.20 |
| amyn | i prefer using C# but i can use C++ too | 11:07.21 |
| kens | I have no clue how to interface C# to a DLL | 11:07.49 |
| | With C++ you cna use the import library and you then will need to decalre the interface functions as 'C' I htink. | 11:08.17 |
| amyn | oh ok i see. by any chance do u have a simple code 1 or 2 lines which is similar to running a basic command on gswin32c | 11:09.14 |
| kens | There is no such sample code, the interface is not so simple. | 11:09.39 |
| | You need to at the least execute gsapi_new_instance and gs_api_init_with_args tehn you will need one or more of gs_run_string_begin, gs_api_run_string_continue, gs_api_run_string_end or gs_api_run_string_with_length or gs_api_run_string or gs_api_run_file | 11:11.04 |
| | On exit you will need gsapi_delete_instance and gsapi_exit | 11:11.29 |
| | You will probably want gsapi_set_stdio so you can see the back channel (stdout and stderr) | 11:11.56 |
| | You may want gsapi_set_poll so that you application remains responsive (or you may choose to run Ghostscript in a separate thread) | 11:12.46 |
| amyn | ok before discussing further, do you think make this change would increase speed in any way? considering now i wont have to create a Process(which actually just runs the command prompt) again and again | 11:12.55 |
| kens | Its definitely faster to run in job server mode | 11:13.07 |
| | THat is, without continually tearing down the interpreter and restarting it | 11:13.20 |
| | THere is an overhead of several seconds on starting up GS | 11:13.32 |
| | But that's all you gain | 11:13.55 |
| amyn | so basically if my application is not restarting, this new technique would increase the speed (even if its by a small factor)? | 11:14.38 |
| kens | Yes. Its quicker by several seconds to not close down and restart | 11:15.02 |
| amyn | so all i will need is the gsdll32.dll file which i can import in a C++ project and then with the help of api.htm, i can achieve my task | 11:16.14 |
| kens | Well you have to implement some code to do this, but technically yes | 11:16.45 |
| | As I said, the existing gswin32/gswin32c executables already use the DLL and the API | 11:17.06 |
| amyn | just an estimate, if i just want to convert a pdf to images(only the above command i want to use), how much code will I have to write? and if i give enough time and dedication, is it easy to understand and implement? | 11:17.46 |
| | yes but rather then running those executables each time i want to convert, i want to make it a part of the application and move further | 11:18.29 |
| kens | It doesn't matter what you want to do, the amount of code you require to interface to Ghostscript is roughly the same. I can't answer whether *you* will find it easy to implement. I suspect that you may find it tricky as the existing code is all written in plain old C. It will depend how well you can understnad the existing code. | 11:19.17 |
| | The point about the existing code using the DLL is that you can rebuild the executable yourelf, in debug mode. Reading the code and stepping through the exacutable should give you a good insight into how the existing code works. You will need to replicate at least some of the same functionality. | 11:20.14 |
| amyn | well if i just create an empty project in C++, import the dll and write the code example in api.htm (4.1), it should work straight away? | 11:21.02 |
| kens | I would imagine so, it's a start at least | 11:21.33 |
| | Note that there are several different examples in there | 11:21.51 |
| amyn | so looks like to me this would be tricky but not time consuming. the example is written nicely and somewhat self explanatory | 11:22.45 |
| kens | Note that you should be able to use the gsdll32.lib import library to ease the task of linking to the DLL. I htink we probably should consider adding a header file with teh declarations too. I wonder if they are in gsapi.h | 11:24.03 |
| | Looks like iapi.h is the relevant header file | 11:25.56 |
| | It declares all the DLL interface routines | 11:26.08 |
| kens | lunches | 11:31.22 |
| meluha | hi | 11:35.37 |
| 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. | 11:35.37 |
| meluha | Is there any efforts going on to make use of GPU in Ghostscript processing? | 11:35.47 |
| amyn | can I use this http://www.java2s.com/Tutorial/CSharp/0520__Windows/DllImportloaddlllibrary.htm and use C# with gsdll32.dll? | 11:56.38 |
| tor8 | meluha: no. there is an experimental branch to make use of the GPU in MuPDF processing though. | 11:57.30 |
| kens | amyn : I have no idea | 12:04.49 |
| | I know nothing about C# | 12:05.01 |
| amyn | can u tell me how do u include .dll in C++? | 12:07.15 |
| kens | I don't, I use C. However the Windows API call LoadLibrarayEx would be my first choice. Again I have no idea how/if you cna do that in C# | 12:07.58 |
| tor8 | nice bash exploit :( | 12:09.10 |
| meluha | tor8: where can I look for that branch? | 12:09.12 |
| kens | tor8 yeah, that's a real big booboo there :-( | 12:09.28 |
| | Makes Heartbleed look minor | 12:09.37 |
| tor8 | meluha: http://git.ghostscript.com/?p=user/tor/mupdf.git;a=shortlog;h=refs/heads/opengl2 | 12:09.40 |
| meluha | tor8: thank you | 12:10.33 |
| amyn | +kens: thanks. i will look into it futher | 12:12.30 |
| kens | Oh, I was about to ask if he'd looked at Ghostscript.NET...... | 12:14.10 |
| | LOL turns out he asked on Stack Overflow so I was able to answer it there :-) | 12:19.07 |
| Robin_Watts | kens: I pointed him at Ghostscript.NET yesterday. | 12:21.00 |
| kens | Aha, then he already knows, thanks Robin_Watts | 12:21.11 |
| tor8 | marcosw1: (for the logs) could you please add MuJS to the list of products in Bugzilla? | 12:21.58 |
| Robin_Watts | He's on here asking a lot of questions, wasting a lot of our time, and is determined not to consider a commercial license for the bank he is working for, so I've lost interest in trying to help him. | 12:22.01 |
| kens | Robin_Watts : yes, pretty much how I feel. | 12:22.16 |
| amyn | +kens: i am back as soon as i saw your reply on stackoverflow | 12:29.18 |
| kens | Well, I was simply going to suggest that you try Ghostscript.NET | 12:29.47 |
| | Which I believe is a C# implementation around the Ghostscript DLL. | 12:30.08 |
| amyn | i have already gone over that but am not sure if that fulfills those requirements that I am looking for | 12:30.52 |
| kens | Well if it doesn't, I very much doubt you are going to find anything which does | 12:31.15 |
| amyn | on the Codeplex page, there are examples but it would be better to have a list of functions that it supports | 12:31.26 |
| kens | You oculd ask the developer | 12:31.39 |
| amyn | ok i am asking the developer. but as far as i can see in the code, i cannot find the option to convert pdf to tiff image using ccitg4 compression | 12:43.44 |
| jhabjan | amyn: what exactly are you trying to achieve? | 12:48.21 |
| amyn | i want to convert a pdf to image which should be in tiff format. and to reduce the size of the output images i want to use ccitg4 compression | 12:49.06 |
| jhabjan | amyn; Ghostscript.NET allows you to pass parameters exactly like you would do it via command line | 12:49.07 |
| amyn | to achieve this now, i run the following command on gswin32c | 12:49.33 |
| | gs9.14\\bin\\gswin32c -Igs9.14\\lib;fonts -dSAFER -dNumRenderingThreads=2 -dBATCH -dUseCropBox -dNOPAUSE -sPDFPassword=\"\" -sDEVICE=tiffg4 -r300 -sOutputFile=\"output\" | 12:49.43 |
| | how would i do the above in Ghostscript.Net? | 12:49.54 |
| kens | tor8 I just tried 2 intersecting annotations in Acrobat. THey don't get fainter, and the intersection is the combination of the 2. That is, I highlighted one in cyan and one in yellow, the intersection is red | 12:50.18 |
| | Of course, because Acrobat writea na Appearance, this works when rendered in other tools | 12:50.41 |
| jhabjan | amyn: take a look at the: https://github.com/jhabjan/Ghostscript.NET/blob/master/Ghostscript.NET.Samples/Samples/ProcessorSample.cs | 12:50.47 |
| | amyn: simply list your parameters like the ones in the example I provided | 12:51.55 |
| amyn | i think this is what i was looking for. can u tell me if i use ghostscript.net instead of using process class in c# and running the process for each pdf file, would there be difference in speed? | 12:52.33 |
| kens | tor8 NB that both GS and MuPDF redner the Acrobat annotations correctly, even though the appearance streams are both transparent, which sort of, maybe, does indicate a problem in MuPDF..... | 12:53.15 |
| amyn | jhabjan: also, if i create a new empty project in vs2010, what references would i have to add and how in order to be able to run the example code that u provided above? | 12:54.26 |
| kens | Oh interesting, I added a 3rd highlight, in green, and now the intersection has gone 100% black.... | 12:54.35 |
| jhabjan | amyn: Ghostscript.NET is simply a wrapper for the native ghostscript library, I'm not sure for the speed difference between starting a process and dll call | 12:57.27 |
| kens | jhabjan, he wants to start Ghostscript up, then run jobs one at a time, without restarting Ghostscript. That is, job server mode. | 12:58.09 |
| jhabjan | amyn: you can use nuget to install Ghostscript.NET in Visual Studio Project | 12:58.11 |
| amyn | in does your code simply run the desired command on command prompt and returns the output or is there something else? | 12:58.17 |
| jhabjan | amyn: Ghostscript.NET uses API functions exposed by the native Ghostscript library. There is no command lines calls there. | 13:00.44 |
| kens | tor8 So if I remove the appearance streams, Ghostscript still draws the same result as Acrobat. TO me that does suggest a problem with the MuPDF appearance stream generation or rendering | 13:00.59 |
| amyn | jhabjan: then i guess Ghostscript.NET would be a little faster since there is a overhead when we execute gswin32c on command line each time | 13:01.43 |
| jhabjan | amyn: could be | 13:02.59 |
| amyn | great. thanks a lot :) | 13:03.17 |
| jhabjan | +kens: yep, either that or run multiple instances at a same time... | 13:04.21 |
| kens | what ? Me ? | 13:04.45 |
| tor8 | kens: yeah. it could be either the choice of blending mode, or lack of merging the annotations, when mupdf generates the highlight annotations. | 13:05.06 |
| kens | tor8 I have to say I'm not certain, but the appearance streams generated by Acrobat look fine. I know how the appearance streams are generated by Ghostscript when they are missing, and those seem to render just fine too. Both cases use transparent objects. So my guess is something 'odd' about the way that paul is doing the annotations. Of course, I can't try it myself..... | 13:06.33 |
| | For certain Ghostscript doesn't mege the annotations when generating the appearances, it just draws each annotation as it encounters it, so the second one will be impacted byt hte result of drawing the first. | 13:07.21 |
| tor8 | kens: yet one more reason to get that minimal x11/win32 viewer to do annotation editing... | 13:07.24 |
| kens | I guess I'd appreciate that :-) | 13:07.46 |
| paulgardiner | The blending mode used in highlight annotations is wrong. We have a user looking at that around the time I stopped working in MuPDF. | 13:13.21 |
| | I thought he produced a patch | 13:13.35 |
| kens | Ah, well that might explain the problem | 13:13.41 |
| | GS and Acrobat use a Multiply blending mode | 13:14.04 |
| kens | thought paul was out today too...... | 13:14.33 |
| paulgardiner | Last minute change of plan | 13:14.47 |
| mvrhel_laptop | good morning | 15:20.16 |
| Robin_Watts | Morning mvrhel_laptop | 15:20.23 |
| mvrhel_laptop | Robin_Watts: what day do you fly out? | 15:21.04 |
| Robin_Watts | Next wednesday. | 15:22.02 |
| henrys | mvrhel_laptop: all the code looks good I'd like to have marcos set up regression testing though. | 15:22.04 |
| mvrhel_laptop | henrys: yes. that would be a good idea. | 15:22.15 |
| Robin_Watts | mvrhel_laptop: are you going early to see your folks? | 15:22.25 |
| henrys | mvrhel_laptop: ie I can see a lot more potential for problems vs. the rectangles. | 15:22.28 |
| mvrhel_laptop | I would like to see the pdf fts tests pushed through the xpswrite device | 15:22.38 |
| | is there an easy way for me to use the clusters to do that and look at the result? | 15:23.04 |
| Robin_Watts | mvrhel_laptop: Not an easy way. | 15:23.23 |
| mvrhel_laptop | ok. I will set up something here this weekend and have a look | 15:23.34 |
| Robin_Watts | You *could* hack xpswrite to be called pdfwrite, and pdfwrite to be called something else, and then push that. | 15:23.47 |
| henrys | mvrhel_laptop: let's ask marcos to do it. | 15:23.49 |
| | mvrhel_laptop: if we are going to get serious about this device it needs to be supported by the regression tools | 15:24.21 |
| mvrhel_laptop | henrys: yes I agree | 15:24.31 |
| Robin_Watts | Then the standard testing of pdfwrite would test xpswrite instead and it'd report differences from that. | 15:24.34 |
| henrys | mvrhel_laptop: the set up is identical to pdfwrite. | 15:24.43 |
| Robin_Watts | but setting it up for longer term testing seems better. | 15:24.46 |
| | I can probably tweak the cluster so that it always tests xpswrite if you want. | 15:25.20 |
| henrys | except gxps will be used to make the rasters not gs | 15:25.23 |
| Robin_Watts | oh, hmm, yeah. | 15:25.37 |
| mvrhel_laptop | yes. that is how we would want it | 15:25.41 |
| henrys | the only concern I have is if we have files that will take forever, so first go should be an overnight that doesn't drag down the cluster | 15:26.27 |
| mvrhel_laptop | henrys: I agree with that. I was thinking that for now, lets just do the pdf fts tests | 15:26.49 |
| | there are going to be a pile of issues | 15:26.55 |
| | I suspect | 15:27.00 |
| | I have some ideas about how we can deal with transparency. I am not going to spend much time on this but I may play with it a bit | 15:27.37 |
| henrys | mvrhel_laptop: I'll send mail to marcosw1 and reference this irc discussion and we'll see how we get on. | 15:27.39 |
| mvrhel_laptop | ok | 15:27.44 |
| | henrys: I am flying out Saturday to spend a couple days with my folks before the staff meeting, so my hours will likely be a bit shifted | 15:28.46 |
| henrys | mvrhel_laptop: is it your read MS is going to stick it out with XPS? It isn't being adapted elsewhere and even they aren't doing anything with it? | 15:29.07 |
| | mvrhel_laptop: oh I fly saturday too. | 15:29.42 |
| mvrhel_laptop | henrys: as I understand it, they like this solution for the print pipeline. but with them who knows. there is 0 work/support going on in the area | 15:29.58 |
| Robin_Watts | mvrhel_laptop: Presumably you'll be back in Chicago Wednesday evening? | 15:35.16 |
| | We land at 13:35, so we'll be around for food in the evening if you want to meet up. | 15:35.51 |
| mvrhel_laptop | I think any dreams they had of xps being like pdf are long gone. I was bummed to see that they added a pdf rendering in windows 8.1 and we somehow missed that potential chance | 15:38.24 |
| | Robin_Watts: yes. I will be back Wed afternoon | 15:38.42 |
| | dinner would be great | 15:38.50 |
| henrys | Robin_Watts, mvrhel_laptop oh you guys aren't going to rush over and enjoy the last day of the show ;-) | 15:40.25 |
| mvrhel_laptop | oh. is the show before the meeting? | 15:40.44 |
| | henrys: does the show end wed? | 15:41.04 |
| henrys | mvrhel_laptop: yup | 15:41.13 |
| Robin_Watts | By the time we clear customs/immigration, get to hotel, shower, and get into downtown I think we'd be too late. | 15:41.43 |
| mvrhel_laptop | I guess I could swing by there. I actually will be returning car to midway and then riding the EL downtown and out to Ohare | 15:41.48 |
| Robin_Watts | (Phew, think I got away with that one) | 15:41.51 |
| kens | thoguth the show was after the meeting, ah well | 15:42.12 |
| mvrhel_laptop | but this all depends upon traffic etc | 15:42.20 |
| henrys | last thing you want to do is drag your jet lagged selves through that nightmare ... | 15:42.56 |
| kens | I don't think we could make it before 4 pm anyway, what with all the ocmplications, as Robin says | 15:43.17 |
| mvrhel_laptop | that is when everyone is already packing it up | 15:43.35 |
| | and ready to get out of there | 15:43.41 |
| kens | Exactly :-) | 15:43.45 |
| | Finishes at 3pm ont eh 1st anyway | 15:44.33 |
| henrys | I've gone through the brochure - just don't see anything interesting going on this year. Seems to get worse each year. | 15:44.42 |
| kens | I really don't think we could land at 13:35 and get to the hall by 15:00 | 15:44.56 |
| Robin_Watts | Stand 4: BigPrintCorp. Visit our show theatre with Hookers and Blackjack! | 15:45.28 |
| kens | ROFL Miles will be happy | 15:45.46 |
| | Though I think he prefers poker ? | 15:45.53 |
| mvrhel_laptop | oh 3pm. I doubt I will make it by then | 15:46.00 |
| henrys | If anybody sees anything they'd like me to "recon" let me know. | 15:47.26 |
| jogux | mvrhel_laptop : I have some new SOT bugs for you if you have time at some point :-) | 16:19.55 |
| mvrhel_laptop | jogux; ok. let me come to skype | 16:20.19 |
| kens | goodnight all | 16:21.38 |
| rayjj | norbertj: (for the logs). Thanks for spotting the bfile/cfile confusion in gxclthrd.c. I just made that change. | 16:49.25 |
| mvrhel_laptop | henrys: doing a cluster push just to make sure there are no major goof ups with the xpswrite changes then I will push. | 22:00.02 |
| henrys | mvrhel_laptop: why push before marcosw1 has it testing? | 22:16.57 |
| mvrhel_laptop | henrys: ok. fair enough | 22:17.13 |
| | will keep cracking on SOT... | 22:18.29 |
| | Forward 1 day (to 2014/09/26)>>> | |
IRC Logs