| | <<<Back 1 day (to 2016/02/14) | 20160215 |
| Asuran | hey kens | 07:50.18 |
| kens | Morning | 07:50.36 |
| | YOu had a question about undefinedfilename, right ? | 07:50.48 |
| Asuran | i got problems with creating a PDF/A with color profile in it, it wont be embedded, it seems the PDFA_def.ps is not correct or idk? really idk :P | 07:50.54 |
| | im unsure, i had such a problem but "solved" it | 07:51.13 |
| kens | OK | 07:51.17 |
| | Well I know its possible to make a PDF/A, I was doing so on Saturday | 07:51.36 |
| Asuran | yea it is | 07:51.42 |
| | it works | 07:51.43 |
| | but the thing i bother with is about embedding color profile | 07:51.55 |
| kens | *which* colour profile | 07:52.05 |
| Asuran | i could do it without but i want professional result if it is possible | 07:52.08 |
| | which i assume | 07:52.11 |
| kens | As I recall you *mus* embed an OutputICCProfile to be valid | 07:52.37 |
| Asuran | ah i saw this maybe in some other spec file but im not capable of ps scripting | 07:53.28 |
| | i could paste it if you like | 07:53.53 |
| kens | You shouldn't need to write any PostScript, other than what's already in place in pdfa_def.ps | 07:53.59 |
| | Can you put it on a pastebin please, it'll get throttled here | 07:54.11 |
| Asuran | http://pastebin.com/hVRrvVXG | 07:54.29 |
| | yep i was about to hehe | 07:54.44 |
| kens | Well that includes the code we put in pdfa_def.ps | 07:55.11 |
| | Its expecting a profile filename defined in PDFAICCProfile | 07:56.11 |
| | Which isn't the command line switch we use | 07:56.19 |
| | Is there a reason to use this instead of pdfa_def.ps ? It looks to me like it does the same job i a more complicated way | 07:56.58 |
| Asuran | no i dont use it | 07:58.39 |
| | i just looked in it to see if thats the problem | 07:58.50 |
| kens | TO be honest, you haven't explained the problem to me, or at least I don't understand your problem | 07:59.19 |
| Asuran | so the program maybe temp create this profile file? | 07:59.30 |
| kens | which program ? | 07:59.41 |
| Asuran | okay, i got pdfs done with this program (and other) and they show up in Adobe Reader DC that the color profile seems to be included | 08:00.05 |
| kens | And what is the program name ? | 08:00.18 |
| Asuran | when i do my own pdf/a with ghostscript commandline i just get information is pdf/a-2 compilant | 08:00.22 |
| | pdf24 | 08:00.24 |
| | and eDocPrinterPro | 08:00.29 |
| | both produces pdfs files which seem correctly include colorprofile | 08:00.45 |
| kens | Right so its another one using Ghostscript to do the work | 08:00.49 |
| Asuran | exactly | 08:00.56 |
| | but it seems to do it better then myself | 08:01.06 |
| kens | The Postscript you posted does hte same job as our pdf_def.ps file, which isn't surprising if they are using GS | 08:01.20 |
| Asuran | yea its a modified one of yours | 08:01.40 |
| | idk, why, but yea | 08:01.46 |
| kens | If the PDF is PDF/A compliant, what's the problem ? | 08:01.49 |
| Asuran | ill screenshot it | 08:02.09 |
| | https://i.imgsafe.org/a6c318a.jpg <- this is what i produce with ghostscript | 08:05.16 |
| kens | OK...... | 08:05.42 |
| Asuran | sec | 08:06.18 |
| | its hanging | 08:06.20 |
| | stupid website, poor programmed | 08:06.26 |
| | nvm... | 08:06.36 |
| | https://i.imgsafe.org/0611946.jpg and this with other printer which seems use ghostscript | 08:06.49 |
| kens | OK well there's really not a lot I can say from the screenshot. | 08:07.10 |
| Asuran | its in the left side | 08:07.19 |
| | of them | 08:07.21 |
| | on lists the color profile | 08:07.26 |
| kens | Yes but it doesn't help much | 08:07.27 |
| Asuran | one not | 08:07.28 |
| | yea but i cant tell you more | 08:07.35 |
| kens | But I have no idea what its talkign about there | 08:07.39 |
| | SO, what you need to do is open the PDF in a text editor and look for OutputConditionIdentifier | 08:08.01 |
| Asuran | ausgabe bedingung = output condition | 08:08.04 |
| | which references to color profile | 08:08.11 |
| | okay | 08:08.24 |
| | sRGB_v4_ICC_preference | 08:09.02 |
| | its standing there | 08:09.06 |
| kens | THen you have a PDF file with a ICC profile | 08:09.18 |
| Asuran | oh cool | 08:09.27 |
| kens | But...... | 08:09.27 |
| Asuran | yes? | 08:09.30 |
| kens | You haven't emdedded the profile (I suspect), just referenced the name | 08:09.45 |
| | Whcih is allowable for a 'standard' profile | 08:09.56 |
| Asuran | maybe, i use this command line: http://pastebin.com/MPnLvXWW | 08:10.15 |
| kens | To be honest, ths is not a sensible way to proceed. | 08:10.19 |
| Asuran | hmm | 08:10.38 |
| | i dont understand this | 08:10.42 |
| kens | Drop NumRenderingThreads, it has no effect when not rendering. Don#t specify a resolution, doubly don't specify one that's the same as the default. I would no reccomend using -DPDFSETTINGS, ever. | 08:11.24 |
| | Don't use -dRenderIntent or OverrideICC | 08:11.42 |
| | You've set ProcessCOlorModel twice, once with an invalid entry | 08:12.03 |
| | You've set ColorConversionStrategy incorrectly | 08:12.17 |
| | You don'#t need to specify the Default ICC profiles | 08:12.31 |
| | You don't weant to set OutputICCProfile | 08:12.48 |
| | And you don't want DeviceGrayToICC | 08:13.03 |
| Asuran | fixed | 08:14.34 |
| kens | I would suggest: | 08:14.52 |
| | gswin64c -dBATCH -dNOPAUSE -sDEVICE=pdfwrite -dPDFA=2 -dPDFACompatibilityPolicy=1 -sColorConversionStrategy=RGB -sOutputFile="%~n1.pdf" PDFA_def.ps %1 | 08:14.52 |
| Asuran | hmm | 08:16.43 |
| | nope | 08:16.44 |
| | still not embedding | 08:16.47 |
| kens | Put the file on dropbox | 08:16.54 |
| | What I was in the middle of saying above is that its not sensible to *talk* about the files, its only sensible for me to look at them. | 08:17.30 |
| Asuran | http://pastebin.com/vD6W99VJ its the PDFA_def.ps | 08:18.37 |
| kens | No, I meant put the outptu PDF file somewhere where I can look at it. I cannot tell what the content of the file is by looking at Acrobat screenshots | 08:19.04 |
| Asuran | ah ok | 08:19.09 |
| | https://www.docdroid.net/file/download/nhYgJ4W/ghostscript-freenode-web-irc.pdf | 08:21.03 |
| kens | Well that looks fine to me | 08:22.47 |
| Asuran | that sounds good | 08:22.57 |
| | is it really embedded the file? | 08:23.04 |
| kens | THe OutputConditionIdentifier specifies the human readable name of the ICC profile, and the profile is embedded | 08:23.06 |
| Asuran | and the reference correct? | 08:23.08 |
| | nice | 08:23.11 |
| | thanks kens | 08:23.14 |
| | so i can assume its working fine? | 08:23.19 |
| kens | It looks fine to me, you should run it through a validator, its impossible to validate manually in reasonable time, and I'd have to decompress it to do so as well. | 08:23.52 |
| Asuran | hmmm i know only a online validator from pdf-tools.com | 08:24.11 |
| | which didn't gave me an error | 08:24.19 |
| kens | Acrobat doesn't like it claiming that the ICC profile is the wrong type | 08:24.35 |
| Asuran | ah! | 08:24.48 |
| kens | I suspect ytou have not used a printer profile | 08:24.55 |
| Asuran | idk know it really, but i will try one from pdf24 | 08:25.10 |
| kens | THe profile needs to have a type of 'prtr' I thnk | 08:25.25 |
| | Or 'mntr' I can't recall which | 08:25.36 |
| Asuran | hmm didn't changed anything | 08:26.50 |
| kens | OK You are using an invalid ICC profile. | 08:29.23 |
| | You need one whch has a type of either 'mntr' for viewing on a mopmnitor or 'prtr' fro printing | 08:29.41 |
| Asuran | where to get one? | 08:29.55 |
| kens | It also needs to have a colour space of Gray, RGB or CMYK | 08:29.57 |
| | Lab is not a vliad space | 08:30.07 |
| Asuran | Lab? | 08:30.25 |
| kens | I can't tell you where to get a profile | 08:30.26 |
| | Lab is a CIE colour space | 08:30.46 |
| | More properly written as L*a*b | 08:30.59 |
| Asuran | nope the file is okay | 08:31.36 |
| | i used the one from pdf24 now | 08:31.42 |
| | same result | 08:31.48 |
| kens | What result ? | 08:32.09 |
| | If you mean the Acrobat display, I do not care what it says | 08:32.19 |
| Asuran | as i posted in the screenshot, adobe reader doesnt seem to recognize an embedded profile | 08:32.27 |
| kens | I'm using a preflight validator | 08:32.30 |
| | That does not mean there is no profile | 08:32.42 |
| Asuran | idk, because i cant find information about it | 08:32.54 |
| kens | I don't know what is in the PDF file which you haven't posted | 08:32.58 |
| | The PDF file you did post is 'nearly' valid except that you are using an invalid profiole for the OutputIntent | 08:33.20 |
| Asuran | sec | 08:33.49 |
| | https://www.docdroid.net/file/download/s1fGqMD/get-list-of-passed-arguments-in-windows-batch-script-.pdf | 08:35.35 |
| kens | well that one uses a mntr RGB profile for the OutputIntent | 08:37.15 |
| | By the way it would simplify matters enormously if you would use a *much* simpler PDF file | 08:37.39 |
| | That last one decompresses to 10MB | 08:38.05 |
| Asuran | ok | 08:39.13 |
| kens | The second fikle includes 2 ICC profiles | 08:39.49 |
| Asuran | https://www.docdroid.net/file/download/PMvtYTw/omegle.pdf simpler one | 08:40.10 |
| | i got the profile from: http://www.color.org/srgbprofiles.xalter | 08:40.42 |
| kens | Right. The second file has written all the page content in an ICC colour space and its a different space to the OutptuIntent | 08:40.44 |
| | I imagine that is what Acrobat is telling you. | 08:40.55 |
| | Its also a completely pointless exercies | 08:41.02 |
| | Unless you are working in a tightly controlled colour calibrated workflow | 08:41.17 |
| | Which you clearly are not. | 08:41.26 |
| | So in short you don't need to worry about it IMO | 08:41.38 |
| | Ah in fact the ICC profile and the OutputIntent profile are the same, which makes it even more pointless :-D | 08:42.25 |
| Asuran | as i said i just wanted professional output and it makes sense embed a profile which can be used | 08:43.04 |
| | but yea im not working on color calibrated monitor | 08:43.25 |
| kens | If you want a professional output, then you need to embed thw *right* profile | 08:43.30 |
| | If you don't know what that means then its time to stop worrying about it being 'professional' | 08:43.46 |
| | Just embed a profile which works | 08:43.52 |
| | THat is what tghe pdf24 code is doign | 08:44.03 |
| Asuran | so it does make it better? because i use that profiles from it and it still dont work | 08:45.35 |
| kens | It makes it better *IF* you know exactly what you are doing, where your colour is managed from beginning to end. | 08:46.18 |
| | I imagine ths is not the case for you (it practically never is) | 08:46.31 |
| | For example, where are your input files coming from, are you making them yourself ? | 08:46.46 |
| Asuran | i use eDocPrinterPro to post it as .ps | 08:47.05 |
| | and then i use ghostscript for doing the pdf | 08:47.18 |
| kens | So not colour managed in any way at all | 08:47.22 |
| | You do realise that your Windows system is perfectly capabl;e of producing a PostScript file without the 'eDocPrinterPro' software ? | 08:47.51 |
| Asuran | not really | 08:48.03 |
| kens | Well, it is. | 08:48.19 |
| | Create a PostScript printer, tell it to use the FILE: port and every time you print to it it will ask you where to save the PostScript file | 08:48.45 |
| | Without knowing more about the eDocPrinterPro software I would guess its doing the same | 08:49.03 |
| | eDocPrinterPro seems to be intended to create PDF files, so using it to create PostScript and then converting to PDF seems a little excessive | 08:49.49 |
| | Good grief, they even have a download button on their web site for Ghostscript | 08:51.19 |
| Asuran | yea | 08:51.36 |
| | and auto download on install | 08:51.43 |
| kens | I'm not sure that's legit if they don't include a link to the sources | 08:52.02 |
| | Anyway | 08:52.22 |
| Asuran | well they want 80 ⬠for their product | 08:52.35 |
| kens | Yeah and its GS doing the heavy lifting, so that's a cheek | 08:52.52 |
| | ANyway | 08:53.08 |
| | The point is that your PDF/A file will be valid if you just use an acceptable profile | 08:53.26 |
| | Do not worry about the content being in an ICC space | 08:53.39 |
| Asuran | sRGB_v4_ICC_preference_displayclass.icc was really wrong? | 08:53.47 |
| kens | Yes | 08:53.51 |
| Asuran | hmm then idk | 08:53.55 |
| kens | Use the one from pdf24 | 08:54.05 |
| Asuran | i did | 08:54.08 |
| | same | 08:54.10 |
| kens | That is actually an HP profile | 08:54.15 |
| | I know its the same, but you are aiming for something you don't need | 08:54.25 |
| | Yo udo not need Acrobat to say the content is in an ICC space | 08:54.35 |
| Asuran | yea makes sense it will use clients profile so yea maybe youre totally right | 08:54.47 |
| kens | SO forget that little thing on the left it is not implortant | 08:54.51 |
| Asuran | i have to i guess | 08:55.25 |
| | another thing, you said the /prepress option was not good idea to use? why? | 08:55.43 |
| kens | Because it alters stuff. | 08:55.54 |
| Asuran | ah. | 08:55.58 |
| kens | People think 'prepress = best quality' that is not true | 08:56.07 |
| Asuran | that could result in not pdf/a-2 compilant? | 08:56.13 |
| kens | No | 08:56.19 |
| Asuran | hmm | 08:56.27 |
| kens | But it depends on how you construct the command line | 08:56.30 |
| | It will, for example, set the colour space to CMYK | 08:57.14 |
| Asuran | okay | 08:57.15 |
| | oh this could be trouble | 08:57.26 |
| kens | Indeed | 08:57.31 |
| Asuran | when i use params? | 08:57.31 |
| kens | It depends ot he order of the parameters | 08:57.41 |
| Asuran | i knew it | 08:57.51 |
| kens | Get it right and its OK< get it wrong and its not | 08:57.51 |
| Asuran | i swear i knew it that order is important | 08:57.58 |
| kens | The PDFSETTIONGS switch changes a *LOT* of controls | 08:58.14 |
| Asuran | but i thought the programm parses it first anyways and then do the right anrangement | 08:58.14 |
| | and quality wont improve at all? | 08:58.30 |
| kens | If you then change them back afterwards, ts OK, but if you run PDFSETTIONGS last, its not | 08:58.34 |
| | No. You can't improve on the original input | 08:58.45 |
| Asuran | okay | 08:58.51 |
| kens | And the default settigns are 'don't change anything unless you have to' | 08:59.03 |
| Asuran | pdf is just ps but with addiotional stuff? | 08:59.05 |
| kens | Whereas the PDFSETTINGS are 'change things like ths' | 08:59.13 |
| | No | 08:59.17 |
| Asuran | k | 08:59.21 |
| kens | PDF is a page description language, PostScript is a programming language | 08:59.31 |
| | You cannot write programs in PDF, you can in PostScript | 08:59.46 |
| Asuran | ok thanks kens | 09:00.45 |
| | for all | 09:00.47 |
| | ^^ | 09:00.48 |
| kens | Right coffee time | 09:00.53 |
| Asuran | not for me hehe, its now idk time... | 09:01.11 |
| | shall i dislike Adobe? | 09:01.15 |
| | i ask this to myself maybe | 09:01.23 |
| | i tend to dislike them | 09:02.04 |
| FjordPrefect | Hey guys. I recently discovered mupdf and have fallen in love with it. Kudos for the good work! The only thing I'm missing is the indexing of internal links in a pdf. Will it be a feature at any point in time? | 12:47.25 |
| kens | I'm not a MuPDF developer but what do you mean by 'indexing' internal links in a PDF ? | 12:49.06 |
| | Are you referring to Outlines ? Link annotatoins ? Something else ? | 12:49.19 |
| | What do you mean by 'indexing' them ? | 12:49.32 |
| FjordPrefect | Like in the adobe reader where the TOC pops up on the side. | 12:50.02 |
| kens | Thats the Outlines tree, also referred to as bookmarks | 12:50.22 |
| FjordPrefect | Ah, okay | 12:50.31 |
| | Is it possible to replicate that in mupdf? | 12:50.49 |
| kens | MuPDF already reads it I believe, but the current example implementation viewers don't display it. Its possible the latest GSView does. | 12:50.55 |
| | Robin_Watts : or tor8 may know better than me | 12:51.13 |
| Robin_Watts | FjordPrefect: It's already there. | 12:51.43 |
| tor8 | FjordPrefect: the mupdf-gl viewer has an outline view. open it with 'o' for outlines. | 12:51.52 |
| Robin_Watts | On what platform are you using MuPDF? | 12:51.58 |
| kens | is out of date as usual | 12:52.01 |
| FjordPrefect | I'm on Ubuntu 14.04 | 12:52.13 |
| kens | lunch time...... | 12:52.50 |
| FjordPrefect | tor8: as in mupdf -o filename ? | 12:53.59 |
| Robin_Watts | FjordPrefect: No. mupdf-gl filename then hit the o key. | 12:55.24 |
| | FjordPrefect: What version of mupdf are you using ? | 12:55.34 |
| FjordPrefect | Robin_Watts: 1.3-2 | 12:57.06 |
| Robin_Watts | FjordPrefect: We're at 1.8 | 12:57.15 |
| | almost 1.9. | 12:57.19 |
| | So you'll need to update to get that feature. | 12:57.26 |
| FjordPrefect | Yikes. The package manager installs 1.3 by default. I'll switch to the latest one. | 12:58.19 |
| | Robin_Watts: Got 1.8 and all is well. Thanks for the help! | 13:14.16 |
| Robin_Watts | no worries. | 13:14.25 |
| FjordPrefect | Robin_Watts: One more thing. Did my specifying mupdf and not mupdf-gl lead you to ask me what version i was using? | 13:18.51 |
| Robin_Watts | No. | 13:19.01 |
| | Well, not by itself. | 13:19.13 |
| FjordPrefect | Haha, okay. Normal debugging process then. | 13:20.50 |
| kens | Grrr | 14:12.56 |
| Asuran | kens: it has huge impact if color profile is embedded or not, the text is sharper with color profile, without blurry | 14:20.40 |
| | it seems really theres a problem with youre PDFA_def.ps | 14:20.58 |
| kens | No | 14:21.02 |
| Asuran | i can proof it to you, i made twice the pdf, one with and without | 14:21.23 |
| kens | PDFA_def.ps does not control the conversion into ICCbased colour for the content of the file. | 14:21.28 |
| | Asuran : I am too busy at the moment | 14:21.34 |
| Asuran | im talking about embedding the profile hehe | 14:22.05 |
| kens | If PDFA_def.ps does not produce a valid PDF/A-ab file then you can open a bug report | 14:22.11 |
| | The profile *is* embedded. | 14:22.23 |
| Asuran | yes maybe, but its not used since the PDFA_def.ps is maybe outdated or buggy at all, idk whats the problem, but it looks really its in there | 14:23.13 |
| | it does report a valid one with the validator i used | 14:23.34 |
| kens | I seriously do not have the time to discuss ths right now. If the PDF/A is valid then its valid. | 14:23.51 |
| Asuran | but you clearly see text blurryness in the pdf created with ghostscript | 14:23.53 |
| | kens, no, that means the parser says it is | 14:24.10 |
| kens | 'blurriness' is more likely due to the colour conversion. | 14:24.31 |
| | It is nothign to do with the PDFA_def.ps file | 14:24.42 |
| Asuran | https://www.docdroid.net/file/download/GPQERNF/readme.pdf and https://www.docdroid.net/file/download/D13WhtV/tfz.pdf see the difference when you got time | 14:25.02 |
| kens | It will not be today | 14:25.10 |
| Asuran | well just open them and look when you got time and you see the one where adobe reader dc says something about colorprofile looks sharper then the one i made with ghostscript | 14:27.21 |
| tor8 | Robin_Watts: agh. upgraded some packages and, oops, it upgraded to java 8 and now the android dex tool barfs on the class files >.< | 15:20.39 |
| Robin_Watts | tor8: Yeah, doan do tha. | 15:21.53 |
| | I *thought* that Android pretty much worked with jdk 8 now though. | 15:22.23 |
| | What version of the sdk are you using? | 15:22.34 |
| | I have both 1.7 jdk and 1.8 jrt on this machine | 15:23.18 |
| tor8 | Robin_Watts: the latest... | 15:32.11 |
| | might depend on what target is set for android | 15:32.23 |
| | I just downgraded back to java 7 | 15:32.26 |
| | but that collides with the UNUSED macro you added a while back | 15:33.38 |
| | I've gone and purged that macro, apparently added for clang warnings | 15:33.49 |
| | I've since gone and added -Wno-unused-parameter so that warning macro isn't needed anymore | 15:34.09 |
| | and fixed a bug where an explicit type cast was converting one enum to another... apparently to fix a warning but that warning was a bug... | 15:34.43 |
| | commits on tor/master for those fixes | 15:34.56 |
| Robin_Watts | Does that stop marcosw's warning checker whinging about those warnings ? | 15:35.23 |
| tor8 | well, if those warnings reappear we have bigger problems. there are a TON of unused parameters everywhere (all the fz_context ones that may or may not be used in any given function) | 15:36.06 |
| | so I hope marcos is also squashing those warnings | 15:36.23 |
| Robin_Watts | All lgtm. | 15:36.47 |
| | tor8: sebras has an updated commit (or several maybe) to be looked at. | 15:37.08 |
| tor8 | Robin_Watts: ta. I've rewritten the platform/java makefile to use GNUMake syntax and do proper dependency tracking. | 15:37.11 |
| Robin_Watts | tor8: aargh. | 15:37.27 |
| tor8 | Robin_Watts: yeah, I should look at those. I was hoping to do that face to face tomorrow before the meeting. | 15:37.27 |
| Robin_Watts | So you've just stopped it being buildable on windows. | 15:37.43 |
| kens | THat sounds bad :( | 15:38.01 |
| tor8 | Robin_Watts: it wouldn't be buildable on windows anyway... the -I/usr/include/java-7-sdk/include stuff can't possibly have worked | 15:38.13 |
| Robin_Watts | My previous changes meant I could build it using nmake, I think. | 15:38.22 |
| tor8 | or do you mean just the 'javac' and make a jar file commands? | 15:38.25 |
| Robin_Watts | that, yes. | 15:38.34 |
| tor8 | if the latter, then just add a makejar.bat file? | 15:38.43 |
| Robin_Watts | tor8: can we have a makefile.jar and then have makefile include makefile.jar? | 15:39.14 |
| | then I can nmake -f makefile.jar | 15:39.25 |
| tor8 | do you really need to use nmake or would a batch file suffice? | 15:39.52 |
| Robin_Watts | I would prefer make. | 15:40.03 |
| tor8 | java plays badly with make style dependency tracking | 15:40.05 |
| Robin_Watts | A .bat file would do I guess, but it's a pain that it will rebuild everytime. | 15:40.43 |
| | cos I want to add it to the platform/android stuff. | 15:41.00 |
| | actually, no a .bat won't build on linux :) | 15:41.19 |
| | Ideally I want to do the normal platform/android build and have that do the mupdf.jar build as a subtask of that. | 15:41.46 |
| tor8 | a .bat will run as a .sh on linux ;) | 15:42.37 |
| | I took the liberty of ditching the src/ and build/ subdirectories; the java toolchain is really crappy so this simplifies quite a bit | 15:43.19 |
| | libmupdf.jar: | 15:43.29 |
| | javac com/artifex/mupdf/fitz/*.java | 15:43.29 |
| | jar cvf com/artifex/mupdf/fitz/*.class | 15:43.29 |
| | that will not get dependencies remotely correct though | 15:43.58 |
| | for that we'll need $(wildcard) or explicit lists like you made | 15:44.09 |
| Robin_Watts | yeah, explicit lists was a small price to pay, for correctness I thought. | 15:44.40 |
| tor8 | Robin_Watts: okay. I'll see if I can make something that works with nmake. | 15:50.25 |
| | I think the only real problem is the use of the 'subst' function to change the path from com/artifex/foo to com.artifex.foo for javah | 15:51.35 |
| | but if you don't run javah, that might be okay | 15:51.45 |
| | Robin_Watts: does the makefile on tor/jni work for you on windows? | 15:52.54 |
| | if not I'll see about moving the $(subst) call into the recipe | 15:53.07 |
| Robin_Watts | I would like to be able to run javah. | 15:53.08 |
| tor8 | Robin_Watts: okay. | 15:53.16 |
| Robin_Watts | Let me try. | 15:53.24 |
| | makefile(3) : fatal error U1033: syntax error : '=' unexpected | 15:54.28 |
| tor8 | omg nmake is *that* incompatible? | 15:55.29 |
| | okay, try changing all the := to = | 15:55.53 |
| | but let me guess, $@, $^ and $< and $? don't work in nmake either... | 15:56.26 |
| malc_ | tor8: missing jam yet? | 15:57.04 |
| tor8 | malc_: how about even bloody bsd make?! | 15:57.24 |
| Robin_Watts | doesn't like the $^ or the subst | 15:57.44 |
| malc_ | tor8: hah, real men use shake anyways | 15:57.52 |
| Robin_Watts | NMAKE : fatal error U1071: cycle in dependency tree for target 'com/artifex/mupdf/fitz/Annotation.java' | 15:58.48 |
| | tor8: This is why the gs makefiles are the way they are. We have 'top level' makefiles for msvc or unix that set stuff up, then they both include the same platform independent ones that list all the targets etc. | 15:59.47 |
| tor8 | Robin_Watts: I guess it can't do $(FOO:%.c=%.o) style substitutions either? | 15:59.55 |
| Robin_Watts | All the clever stuff is in the top level ones. | 15:59.55 |
| | I'm guessing not. | 16:00.03 |
| tor8 | then you'll have to use batch files or install a real make | 16:00.20 |
| | because I'm *not* breaking dependency tracking in a real makefile just to deal with nmake's brokenness | 16:00.38 |
| Robin_Watts | Or we could just list the files explicitly. | 16:00.46 |
| | which is what I had before and worked on both nmake and make with correct dependencies. | 16:01.14 |
| tor8 | Robin_Watts: hmm. | 16:01.47 |
| | so, does $@ and $^ work or do I need to make those explicit in every rule too? | 16:02.00 |
| | it did *not* have correct dependencies | 16:02.09 |
| Robin_Watts | $^ does not. | 16:02.12 |
| tor8 | it kept rebuilding and getting stale results on gnumake | 16:02.18 |
| Robin_Watts | $@ might be OK. | 16:02.20 |
| tor8 | and rm -f ? | 16:03.37 |
| | or does that need to use 'del' with some other funky flag | 16:03.55 |
| | or is $(RM) set properly? | 16:03.58 |
| Robin_Watts | tor8: I dunno. | 16:04.05 |
| tor8 | javah doesn't update the .h file if the result is the same, so we end up always rebuilding it if make thinks it should be update | 16:04.27 |
| | what about touch, does that exist on windows? | 16:04.37 |
| Robin_Watts | $RM is not set. | 16:04.43 |
| | no, touch does not exist. | 16:04.52 |
| | OK, let's just use a .bat file :) | 16:05.05 |
| | and we'll figure out how to integrate that in the platform/android stuff later. | 16:05.33 |
| tor8 | make jar can invoke the bat file | 16:05.45 |
| | or add a Makejar makefile that you can call with nmake -f that just does the batch file invocation in a recipe | 16:06.18 |
| | that should work for both windows nmake and linux | 16:06.25 |
| kens | Asuran for the logs, your files look absolutely identical to me when opened with Acrobat Professional X on WIndows. Not only that, but the only significant difference between the 2 files is that they have different DestOutputProfiles | 16:56.14 |
| | One is 'spacRGB' and one is 'mntrRGB' teh 'spacRGB one will not be valid. THat's the readme file. | 16:57.03 |
| | However the actual content of the PDF is the same. | 16:57.19 |
| | So if its being view3ed differently then its down to the viewer, and Acrobat does not show a difference that *I* can see. | 16:57.41 |
| | I also fail to see how the Acrobat DC (which I'm not using, I'm using Professional X) message being 'sharper' can be anything to do with the PDF file, since that's a part of the screen furniture drawn by Acrobat. | 16:59.04 |
| | Oh and both files were made with PDF24, whcih uses Ghostscript 9.14 apparently. Neither appears to have been made by you | 16:59.35 |
| Robin_Watts | Unless he's involved in the development of PDF24 in which case, "get a license". | 17:02.08 |
| kens | Yeah I htnk thye may be techncially in breach | 17:02.29 |
| | They don't poit to our installer, they have built one of their own | 17:02.43 |
| | It doesn't include the source, and their '9.18.1' version is not the same as our 9.18. THey also don't include a link to the source | 17:03.11 |
| | Forward 1 day (to 2016/02/16)>>> | |
IRC Logs
