| | <<<Back 1 day (to 2015/04/30) | 20150501 |
| mvrhel_laptop | rayjj: for the logs. | 04:50.56 |
| | after our conversation, I dug a bit more with Acrobat DC | 04:51.10 |
| | they have a setting which lets me set the transparency blend space. | 04:51.38 |
| | I need to investigate further, but it appears that this should allow me to replicate what gs is giving I hope. that is by specifying the use of an rgb blend space for when we are going out to an rgb device and a cmyk blend space for when we are going out to a cmyk device. this would only happen when the blend space is not explicitly set in the source document | 04:53.45 |
| | damn it does not seem to work with this file. | 04:57.28 |
| ManDay | If anyone has time, could they look at that picture http://postimg.org/image/5be73zn5n - I use translate to move the content to the right and increase the MediaBox so as to create a left margin, but the content appears trimmed on the right. Why is that so? Are there more things that need to be adjusted besides the Media- and CropBox (maybe some kind of page size)? | 06:33.14 |
| | Hi kens chrisl, since you just showed up allow me to repeat my question in case you did not see it | 06:54.05 |
| | If anyone has time, could they look at that picture http://postimg.org/image/5be73zn5n - I use translate to move the content to the right and increase the MediaBox so as to create a left margin, but the content appears trimmed on the right. Why is that so? Are there more things that need to be adjusted besides the Media- and CropBox (maybe some kind of page size)? | 06:54.08 |
| | (the generating script is http://dpaste.com/32THDRQ as before) | 06:54.47 |
| chrisl | ManDay: what are you using to view the results? | 06:59.08 |
| ManDay | zathura chrisl | 06:59.23 |
| chrisl | What does Ghostscript do with it? | 06:59.50 |
| ManDay | chrisl: what do you mean? | 07:00.42 |
| chrisl | Oh, hang on, are you narrowing the page size there? | 07:01.04 |
| ManDay | I don't understand the question. Does "page size" refer to something specific? | 07:01.31 |
| chrisl | Erm, the size of the page..... | 07:01.48 |
| ManDay | You can see the PS at the end of the script | 07:01.49 |
| | If "page size" refers to a specific variable in PS, then no. As you can see in the code I'm only modifying the MediaBox and the CropBox and translate in /BeginPage | 07:02.29 |
| chrisl | You're making the MediaBox narrower | 07:02.50 |
| | PDF is a static page layout format - it does not "reflow" to suit different media sizes. | 07:03.45 |
| ManDay | chrisl: You really don't have to explain to me that PDF doesn't reflow :-p | 07:17.02 |
| | I'm not making the MediaBox narrower, what makes you think that? | 07:17.16 |
| | The Media-Box is bigger than the crop box | 07:17.38 |
| | in this case, they are the same | 07:17.43 |
| chrisl | Which is the original? | 07:17.58 |
| ManDay | hold on, i'm collecting the data | 07:18.09 |
| | Original: MediaBox: [0 0 536 697] CropBox: [12.96 20.16 517.28 697] - After modification: MediaBox: [0 0 636 697] CropBox: [0 0 636 697] | 07:22.12 |
| | it should work on any pdf. you can just try it with ./crop.sh -t pdf -100 -100 600 700 or something like that | 07:28.48 |
| chrisl | ManDay: It doesn't seem to have any effect at all on the PDFs I'm trying | 07:40.03 |
| | Oh, it does change the page size, but doesn't seem to affect the viewing area | 07:41.00 |
| ManDay | chrisl: could you be more specific? | 07:42.13 |
| | I don't know what "page size" and "viewing area" are | 07:43.00 |
| chrisl | Hang on, I've got a file that I do see a difference with..... | 07:44.04 |
| ManDay | Here is a document which we can both try: https://docs.sony.com/release/DPTS1_basicoperation.pdf - I call ./crop.sh -t DPTS1_basicoperation.pdf -200 0 708 907 | 07:47.28 |
| | (optionally with -v mupdf) | 07:47.44 |
| chrisl | Looks to me like the contents aren't being translated correctly | 07:53.57 |
| | ManDay: So, you're translating the page contents by 200,0 - that's moving the page contents 200 units to the right.... is that really what you want? | 07:57.09 |
| ManDay | chrisl: yes, of course. | 07:57.25 |
| | i want to create a 200 margin on the left | 07:57.32 |
| | thats how kens explained it to me | 07:57.43 |
| chrisl | Okay, just checking..... | 07:58.22 |
| | Hmm, sorry, I confess, I don't know what's going here.... my only thought was that it's running into the Ghostscript default page size, but that doesn't seem to have an effect :-( | 08:04.21 |
| ManDay | thanks, I'm glad it's not a beginners mistake :) | 08:08.25 |
| | kens: To my rescue? | 08:53.37 |
| kens | ManDay I'm busy, if you want my help you're going to have to do some work to simplify the problem for me to look at. Firstly I'm not going to run a bash script, I'm not a Linux user on the whole, you need to give me a sample file and a Ghostscript command line. Secondly you need to simplify your test file so you can see what's going on. A 7 page PDF file with a mixture of content is too complex to usefully look at. Make a PDF fi | 08:57.40 |
| | le with a simple rectanlge in a single colour. THen you can look at the output and see what's happened to your input. | 08:57.40 |
| | I 'suspect' the problem is that you haven't altered the PageSize though. | 08:58.08 |
| ManDay | kens: Well, that would be it | 08:58.18 |
| | I haven't | 08:58.20 |
| kens | Well then,there's your problem. | 08:58.28 |
| ManDay | I suspected that I need to do that but couldnt find any reference to it | 08:58.31 |
| | Could you tell me where (BeinPage Endpage) and how (which operator) perhaps? | 08:58.48 |
| kens | <</PageSize [width height]>> setpagedevice | 08:59.10 |
| ManDay | oh | 08:59.26 |
| | uhm | 09:00.13 |
| kens | THere are also seeral ways to modify the media size from teh Ghostscrip command line, see the documentation for details, and you will almost certainly want ot use those and the FIXEDMEDIA switch | 09:01.30 |
| ManDay | media size= Page size? | 09:01.59 |
| kens | Yes | 09:02.12 |
| ManDay | ok, so i dont need to write this into the PS code, you mean. i can use the commandline | 09:02.29 |
| kens | Yes | 09:02.36 |
| | I suspect you will *have* to in order to use FIXEDMEDIA | 09:02.59 |
| | Though it should be possible from BeginPage | 09:03.16 |
| ManDay | so i understand correctly that the whole document has one fixed page size. it's not set on a per-page basis (like the Media- and Cropbox)? | 09:04.29 |
| kens | No. | 09:04.37 |
| | You can emit a setpagdevice during the course of a PostScript program | 09:04.52 |
| ManDay | ah, so i could emit a setpagedevice from within an /EndPage to have the next page get another size | 09:05.19 |
| | i guess /beginpage would be too late, right? | 09:05.29 |
| kens | Not in an EndPage, that would be a bad idea. The reason is left as a research topic for the student. | 09:05.41 |
| ManDay | ok | 09:05.49 |
| | I'm calling gs -o "outfile" -sDEVICE=pdfwrite -dDEVICEWIDTHPOINTS=808 -dDEVICEHEIGHTPOINTS=907 -dFIXEDMEDIA -c ... "infile" where "..." sets the media and cropbox to 808x907pts. and translates the content by 100 to the right, but the content is still clipped 100 on the right. | 09:23.35 |
| | Is it at least correct in theory? | 09:23.58 |
| kens | Possibly, it depends on what's in infile. | 09:24.30 |
| ManDay | * that is -f "infile.pdf" | 09:24.36 |
| kens | Why would you use -f ? | 09:24.48 |
| | Not thsat it matters since it has no effect | 09:25.02 |
| ManDay | Because it comes after the -c ? | 09:25.05 |
| kens | What -c ? | 09:25.18 |
| | See, you haven't given a full command line | 09:25.37 |
| | Now I don't know what you're talking about | 09:25.44 |
| ManDay | http://dpaste.com/1FVVHQN this is the full cli | 09:26.32 |
| kens | You seriously need to lose whte space | 09:26.48 |
| ManDay | I've got tabs displayed as two spaces and I'm not familar with how to indent PS or reverse polish in general | 09:27.23 |
| | now I was indenting it by how much the stack grows | 09:27.34 |
| | makes it easier to count? | 09:27.39 |
| kens | Well looking at that page there is monstrous indentation | 09:27.41 |
| ManDay | the stack grows a lot | 09:27.49 |
| kens | ] doesn't increase the stack size, it decreases it | 09:28.25 |
| ManDay | if I push something on the stack i call that increasing the stack | 09:28.57 |
| kens | ] doesn't push anything on the stack, its an operator | 09:29.15 |
| ManDay | that's why it decreases the stack | 09:29.40 |
| | and the indention is removed | 09:29.44 |
| kens | It removes objects from the stack back to the previous mark, creates an array of sufficient size to hold those objects, puts the objetcs in the array and then pushes the array onto the stack | 09:29.54 |
| ManDay | yes | 09:30.03 |
| kens | Well you can indent it any way you like, but its awfully hard to read | 09:30.30 |
| | And given the stack depth of most PostScript programs you're going to be doing a load of scrolling | 09:30.51 |
| ManDay | yes, ok | 09:30.59 |
| | it was just easier to count for me | 09:31.14 |
| kens | What effect do you expect form '0 add' ? | 09:31.47 |
| ManDay | anyway, using pdfinfo.ps the mediabox and cropbox are okay. and my commandline is okay for setting the page size, no? | 09:31.50 |
| kens | Looks OK, can't say more than that without trying it | 09:32.11 |
| ManDay | the code is generated. as useless as 1 mod. but in different cases there is not 0 and not 1 | 09:32.14 |
| | kens: do you have a pdf that is simple for trying it that you could give me? | 09:35.50 |
| kens | Nope, but you cna make one | 09:36.04 |
| ManDay | Ugh. Can I also take an easy one like that http://www.education.gov.yk.ca/pdf/pdf-test.pdf | 09:36.16 |
| kens | That's not simple, I see text, images and annotations i that file, and that's just by visual inspection. | 09:36.47 |
| ManDay | http://www.krf.no/globalassets/testfiler/test.pdf :-D | 09:37.52 |
| kens | I would make a coloured rectangle, the PDF operators are much simpler | 09:38.10 |
| | Text has its own matrix in PDF which will complicate matters. | 09:38.46 |
| ManDay | oh man | 09:38.54 |
| | can we please try with the text? | 09:39.05 |
| | it's supposed to work on text in the end | 09:39.11 |
| | would take me hours to learn how to make a pdf myself :-/ | 09:39.20 |
| kens | But you are complicatng the issue, it will be harder to see the problem if there are more operations involved. Simplicity is the key to debugging problems. | 09:39.44 |
| | You can make a PDF file with Ghostscript....... | 09:39.53 |
| | 3 operators and you cna make a simple file with a coloured rectangle | 09:40.14 |
| | setrgbcolor, rectfill, showpage | 09:40.45 |
| ManDay | heh that was easy | 09:44.45 |
| kens | fetches more coffee | 09:45.42 |
| ManDay | kens: it doesn't happen with the rectangle | 09:48.15 |
| chrisl | ManDay: so try some very simple text..... | 09:49.34 |
| | To your existing Postscript, (before the showpage) add something like: /Helvetica 16 select font 20 400 moveto (a long string of text) show | 09:51.16 |
| | oops, should be: /Helvetica 16 selectfont 20 400 moveto (a long string of text) show | 09:51.36 |
| | And pad out the "a long string of text" so it gets far enough to show the problem | 09:52.18 |
| ManDay | that text doesn't wrap chrisl | 09:54.52 |
| chrisl | Why would you expect it to.....? | 09:55.35 |
| ManDay | It still doesn't happen, even with the text | 09:56.00 |
| | But it does happen with http://www.krf.no/globalassets/testfiler/test.pdf | 09:56.47 |
| kens | Now try adding a CropBox to the PDF when you create it. | 09:56.49 |
| | Using the pdfmark you already know how to use | 09:57.01 |
| | I don't mean using EndPage or your script or anything, just add it to the PostScript to create the file, before the showpage | 09:57.46 |
| ManDay | i know | 09:57.54 |
| | i got it :) | 09:57.58 |
| kens | I htought I'd better be clear | 09:58.05 |
| ManDay | indeed. it now happens | 09:58.12 |
| kens | Yeah, thought it might | 09:58.21 |
| ManDay | with both the rectangle and the text | 09:58.26 |
| kens | The PDF interpreter is applying the CropBox before you translate the content | 09:58.39 |
| | Or so it appears anyway | 09:59.05 |
| ManDay | So I should place a cropbox into /BeginPage ? | 09:59.27 |
| kens | You can probably evade that by changing the PageSize in the BeginPage | 09:59.30 |
| | |Not the CropBox. | 09:59.39 |
| ManDay | "probably"? | 09:59.57 |
| kens | Ths is *PostScript* you;re dealing with, it has no concept of a CropBox | 10:00.00 |
| | Yes, probably. | 10:00.06 |
| chrisl | Erm, *can* you change the PageSize in BeginPage?? | 10:00.08 |
| kens | chrisl I don't see why not | 10:00.17 |
| | BeginPage can call anything, so I can't see why it can't call setpagedevice | 10:00.31 |
| | But that is why I said 'probably' :-) | 10:00.41 |
| chrisl | Because, IIRC, setpagedevice will cause BeginPage to be called again..... | 10:01.00 |
| ManDay | but if the interpreter is applying cropbox before I transate, can I not fix this by fixing the cropbox in /BeginPage BEFORE translate? | 10:01.14 |
| kens | Indeed, which is why you need to check the existing PageSize to see if its what you want before you call setpagedevice | 10:01.28 |
| chrisl | Ah, right :-) | 10:01.44 |
| kens | ManDay, the PDF CropBox is converted into a PostScript clip by the PDF interpreter. Using pdfmark to set the CropBox for the output isn't going to help | 10:01.58 |
| | Although you could do a 2 step process. I wouldn't reccomend it | 10:02.33 |
| ManDay | can I not either prevent the PDF interpreter from creating the clip or already removing it again in /BeginPage? | 10:02.58 |
| | How would chaging the PageSize in /BeginPage help? | 10:03.11 |
| kens | No, by that time its already applied I thnk. | 10:03.16 |
| ManDay | I think a two step process seems more reasonable for my purpose | 10:03.38 |
| kens | You run the risk of multiple cionversions knackering your PDF file | 10:03.52 |
| | THe more conversion you run the higher the likelihood that the output wil be of poor quality (or wrong) | 10:04.16 |
| ManDay | well, 1 conversion or 2 conversion, what's the differentce | 10:04.18 |
| kens | ^^ | 10:04.23 |
| ManDay | first pass I'll fix the *Boxes, 2nd pass I translate, shouldn't be too bad | 10:04.39 |
| | it's not like this is google translate :-P | 10:05.02 |
| kens | Each time the PDF input file is interpreterd and converted to marking operations, each tmie you do that, ther is potential for loss of informatoin | 10:05.17 |
| | And yeah, it is kind of like Google translate | 10:05.27 |
| ManDay | sure, but i hope all loosable information has already been lost with the first pass. otherwise there would be previously not-loosable information be stored in loosable information during the conversion | 10:06.11 |
| | which i hope doesn't happen | 10:06.17 |
| kens | No, doesn't work that way | 10:06.21 |
| ManDay | have you tried? passing a pdf through GS n times? | 10:06.47 |
| kens | Especially if, for example you have JPEG compressed images, each pass will decompress the images and potentially reapply JPEG quantisation | 10:06.50 |
| | You've seen what happens if you JPEG a JPEG image ? | 10:07.26 |
| ManDay | yes, but that's an exception | 10:07.36 |
| kens | No, its an example | 10:07.43 |
| ManDay | lossy compressd data is an exception | 10:07.49 |
| | that's not the general data model here | 10:07.54 |
| kens | You are certain you won;t get JPEG compressed images ? | 10:08.10 |
| ManDay | hardly any. and on those I'm willing to take the risk | 10:08.36 |
| | rather, i think this thing with the cropbox being applied before the transform is a bug | 10:08.56 |
| kens | In any event there are also problems (albeit harder to explain) surrounding the embedding and subsetting of fonts, potential issues with colour spaces and transparency | 10:09.05 |
| | ManDay : That's the way our PDF interpreter works. | 10:09.17 |
| | Its not 'applied before the transform', its applied at a zsensible time in the processing of the PDF file. | 10:09.48 |
| | And I have no idea what you mean by 'transform' here. | 10:10.14 |
| ManDay | but if it's mangled into a "clip" at that time already, how come /EndPage can still modify the cropbox (and thus make previously clipped content visible again)? | 10:10.28 |
| | s/transform/translate | 10:10.34 |
| kens | Its not 'mangled' what do you thnk a CroipBox is, except a clip ? | 10:10.43 |
| | You are confusing teh output (modified by pdfamark) with teh input. | 10:11.05 |
| | The CropBox of the input file is applied before we interpret the page content, and produce marking operations. | 10:11.30 |
| | The pdfamrk alters 'metadata' that is in the final PDF file, it has no effect on the graphcis state | 10:11.59 |
| | pdfmark never affects the graphics state of the interpreter. | 10:12.24 |
| ManDay | why does that cropbox of the input file have to be applied before? | 10:12.33 |
| kens | (Actually that's a simplification, but never mind...) | 10:12.42 |
| | ManDay : When would you like us to apply the CropBox ? | 10:12.53 |
| | It has to be applied before we render graphcs | 10:13.08 |
| | Ohterwise how do we know where to stop ? | 10:13.17 |
| | You are too focused on PDF otuput, you need to remember that Ghostscript also renders to bitmaps | 10:13.45 |
| ManDay | First point being there is obviously something flawed in the logic. It should be straightforward to re-draw the pdf anywhere on any sized-page (basically what we want to do here), how? (In the given framework) Here is how I think it should work and I don'T understand fully what you're saying but I think we're only a short way from this "correct" manner: | 10:17.09 |
| kens | ManDay : If ths were a specific PostScript (or PDF) er purely designed to output to PDF you might have a point. But tha't snot what you're dealing with. | 10:17.57 |
| ManDay | I do realize that. | 10:18.07 |
| | Let me talk pls. | 10:18.10 |
| | It's fully general and a fully general problem | 10:18.24 |
| kens | I disagree, however | 10:18.34 |
| ManDay | All drawing (wether onto a PDF surface or onto any other surface, like a PNG) should be offsetted by the "translate" amount. This includes the creation of a clip (as it is done by the PDF interpreter). Then, the only thing that remains for us to be done is call "translate" AFTER we know a new page should start but BEFORE the PDF interpreter goes about its business. | 10:21.17 |
| | So I think the problem is that /BeginPage in this case isn't really at the beginning of it all, but rather after the PDF interpreter has already started messing with the page | 10:21.55 |
| | aight? | 10:21.59 |
| kens | And indeed it is. If you set the CTM before you draw the page, that's what will happen. | 10:22.08 |
| ManDay | CTM is the clip? | 10:22.26 |
| kens | No CTM is the Current Transformation Matrix | 10:22.39 |
| ManDay | But the problem was not that the CTM was applied before you draw the page. The problem is that the original CropBox was not translated because it was created before we translated! | 10:23.29 |
| | So we need a way to inject a translate before the PDF interpreter creates a clip from the cropbox | 10:23.54 |
| kens | Well, the code is all available, off you go. | 10:24.05 |
| | Its all in PostScript. | 10:24.15 |
| ManDay | so it's not possible at the moment? | 10:24.22 |
| kens | The PDF interpreter works the way it works, if you want to change it, then by all means go ahead, but I won't answer any questions about it. Tht code is not something regard as 'user servicable'. My solutoin is to put the PageSize in the BeginPage. I cannot say for sure that thi will work wihtout testing it, but I believe it is likely. I have explained why a 2 pass approach is less than ideal, but if you think it isn't a proble | 10:27.28 |
| | m, that's up to you. Your alternative is to spend some time (actually make that a *LOT* of time) understaning why the clip occurs where it does. | 10:27.28 |
| ManDay | kens: As for the 2 pass problem: Is there a device which can hold the intermediate data so that tehre is no conversion problem any more? | 10:27.28 |
| | basically an intermediate file format which holds all information so that GS can read it in without loss again | 10:28.31 |
| kens | No, because the problem isn't 'holding the intermediaste data' its the fact that you *DON'T KNOW* what the input is, all you get is what the PDF interpreter passes after its interpreted it. | 10:28.33 |
| ManDay | I think you misunderstood me. What I mean is PDF ---change_boxes---> PS PS ---translate--->PDF | 10:29.41 |
| kens | That's not what happens | 10:30.13 |
| ManDay | I mean during any pass the PDF is completely mangled into PS anyway, right? There is no loss if I store it as a PS and then read the PS in the second pass | 10:30.28 |
| kens | No, the PDF is not 'mangled' at all | 10:30.39 |
| ManDay | In total, it will still only be one conversion forth and one back | 10:30.42 |
| kens | You can't store a PDF as PostScript, for starters the graphcs models are no longer the same | 10:31.06 |
| | At the moment I still don't see whnby you need to translate the output anyway, I thought I gave a solution whch added margins around the PDF simply by altering the MediaBox with negative values. | 10:31.52 |
| ManDay | kens: You said it required bleeding edge GS which it appears it does, 14th of April or something. Also, I thought this would basically amount to the same thing I'm doing right now | 10:32.41 |
| | but well, i guess given the problem right now you're right and I'll wait for that to hit stable | 10:33.06 |
| kens | Yes, it requires current code, no its not at all the same thing you are doing now. The solution I gave does not incvolve translating the content | 10:33.20 |
| | Of course, you will probably have to mess with the CropBox values with that solution, if you want to set both. | 10:34.02 |
| ManDay | well, effectively it does :-P | 10:34.08 |
| | kens: yes | 10:34.24 |
| kens | Again, I disagree, it does not involve translating the content. | 10:34.32 |
| ManDay | when I call pdfinfo.ps on such a thing it will still show the mediabox having 0 0 lower left corner, despite my specifying it as negative during the parse, right? | 10:35.43 |
| kens | No | 10:35.54 |
| ManDay | oh, well that's suggests it would also require a compatible viewer to view this correctly | 10:36.22 |
| kens | Yes. | 10:36.29 |
| ManDay | is this negative mediabox business in the spec? | 10:36.37 |
| chrisl | A *compliant* viewer..... | 10:36.43 |
| ManDay | all right | 10:36.48 |
| kens | ManDay : You cna read the spec | 10:36.53 |
| ManDay | you're being a bit grumpy | 10:37.14 |
| kens | I've tried to explain, and I *do* have work to do | 10:37.34 |
| ManDay | you did explain and I don't want to keep you from work | 10:37.50 |
| | but you must admit the thing with the cropbox being applied before /BeginPage is a subtle issue | 10:38.09 |
| kens | Not especially, if you want to meddle succesfully with the way a PDF interpreter works, you mjust first understand how the interpreter works. | 10:38.55 |
| | WHch is why I wanted you to use a nce, simple PDF file to investigatge the problem. | 10:39.18 |
| ManDay | frankly speaking, this shouldn't depend on implementation details of the PDF interpreter given the architecture as far as I understand it. The PDF interpreter interprets PDFs into PS. GS creates a drawing area for the PS onto which the interpreted commands and my own commands are applied. If I issue a translate, everything (including the cropbox) should be translated. Any deviation from that is a | 10:41.02 |
| | semantic error in my opinion | 10:41.04 |
| kens | It works perfectly to implement teh specification. Its only a problem for you because you want to subvert the operation. | 10:41.38 |
| ManDay | I suspect very much that the problem you're referring to with the cropbox needed to be applied before comes from an flawed structure of how GS figures out the needed page dimensions | 10:41.43 |
| kens | The PDF interpreter does not interpret the PDF into PostScript | 10:41.52 |
| ManDay | kens: What I described can HARDLy be called subverting the operation | 10:42.02 |
| kens | Certainly it is, you want to alter teh clip of the PDF, as defined by the PDF. | 10:42.24 |
| chrisl | ManDay: Ghostscript was not designed as a tool for manipulating PDF files | 10:42.40 |
| kens | That's not what you're supposed to do with the CropBox | 10:42.44 |
| ManDay | kens: CropBox aside. Do you think that Ghostscript should be able to do the following semantic operation: You have a PDF (say A4) and you want to draw it onto a larger, empty A3 canvas, centered in the middle? | 10:43.33 |
| | I think, although GS is not a tool for manipulating PDFs, it wants to be able to do that | 10:43.52 |
| kens | It can, and does. But that is not what the PDF asked for, so once again you are subverting the PDF interpreter | 10:44.05 |
| ManDay | Because basically, all this amounts to is translating the whole PDF anywhere onto the canvas | 10:44.08 |
| | kens: yes | 10:44.35 |
| | so you're suggesting the only operation that does not subvert the pdf interpreter is reproducing the exact same output (possibly in a different file format) | 10:46.48 |
| | fine, I was trying to subvert the pdf interpreter then, I agree | 10:47.08 |
| kens | Yes, that's wjat the word 'Portable' in Portable Document Format meams | 10:47.15 |
| | You shoudl get the same appearance no matter where you interpret the PDF | 10:47.31 |
| | Anythign else is contrary to the intention of the original specification., | 10:47.54 |
| | I'm aware that some people have taken 'Portable' to mean that its some ind of way of carrying data around that they can endlessly manipulate, but that was never the intention | 10:48.34 |
| ManDay | fwiw, I don't see any point in the existance of PDF at all. | 10:49.50 |
| kens | Yuo seem to be in a minority | 10:50.15 |
| ManDay | if you want portable documents (in that sense of looking the same everywhere), use a image file or, if you need resolution independence as well, SVG | 10:50.29 |
| kens | A lot of people seem to like the idea that what you see in one place is what someone else sees in another | 10:50.30 |
| ManDay | kens: sure i am | 10:50.32 |
| kens | image formats are large, and non-scalable, SVG simply sucks for text | 10:50.54 |
| ManDay | kens: yes, and as i said SVG and other file formats serve that purpose well | 10:50.55 |
| | it's a trivial exstension to SVG to make it not suck for text | 10:51.18 |
| kens | ROFL | 10:51.31 |
| ManDay | kens: or what do you mean by suck? | 10:51.53 |
| kens | I mean its awful, you can't have proper fonts for starters | 10:52.12 |
| | And you cna't have anythign except RGB so no spot colours | 10:52.47 |
| ManDay | Ok, the color issue may be more serious, but it's also easily rectifiable | 10:53.18 |
| kens | But it hasn't been, nor has text. | 10:53.31 |
| chrisl | These issues have already been addressed.... in PDF | 10:53.44 |
| kens | Anyway, I'm going to go and do proper work now. | 10:53.51 |
| ManDay | anyway i'm not saying SVG is a dropin replacement for PDF out of the box, but the only real difference between PDF and a picture is that PDF creates the illusion so that people think they still have semantically structured content while it's technically only a picture | 10:54.04 |
| | kens: ok | 10:54.18 |
| kens | THat is complete rubbish | 10:54.19 |
| | And the point of PDF was never to include semantic structure, the appearance has always been all that it is intended for. It was never intended as an editable format. | 10:55.10 |
| ManDay | a fancy picture. but a picture. a picture looks the same everywhere. and with sufficient aux data you can embed color information and everything you want | 10:55.13 |
| kens | A picture is worth 1000 words. | 10:55.31 |
| chrisl | Does svg still not allow font embedding? | 10:56.03 |
| kens | I believe not, though it 'may' have changed since I last looked | 10:56.17 |
| ManDay | chrisl: i don't know. but extending an XML language is as easy as pie. | 10:56.19 |
| chrisl | ManDay: then you're no longer producing svg..... | 10:56.39 |
| pedro_mac | stops watching daytime TV and tunes into this chat | 10:56.45 |
| ManDay | pedro_mac: you're too late, we're just finishing | 10:56.57 |
| chrisl | pedro_mac: I wouldn't bother..... | 10:56.57 |
| pedro_mac | :) | 10:57.05 |
| ManDay | chrisl: extended svg, also xsvg :-P | 10:57.09 |
| chrisl | ManDay: then you need to convince other people to implement it, get it ratified etc - good luck with that | 10:57.45 |
| pedro_mac | I think its fair to say SVG is a handy format for scalable web content, but was never suited to proper typography or print processing | 10:57.57 |
| kens | As an example, see Microsoft's XPS, now an ISO standard and XML based. Loads of penetration there | 10:58.56 |
| | Hmm, why does Miranda only let me ignore 'events' I wonder | 10:59.37 |
| | And where has my list of users gone :-( | 10:59.53 |
| chrisl | I'm trying out xchat for IRC - not really convinced, so far..... | 11:00.51 |
| kens | I must have hit a shortcut key with my usual wild typing :-( Wading through the options tryign to find out how to turn it back on again | 11:01.27 |
| ManDay | you would want weechat if you're on console | 11:01.48 |
| | and linux *cough* | 11:01.52 |
| kens | Hmm, no that check box is still checked | 11:01.54 |
| ManDay | anyway, i guess i'm a bit biased towards PDF. my aversion against it is probably less with the format (it's only a spec, what's the matter) but that every one seems to use it these days where they better should not | 11:02.30 |
| kens | Hmm, back in a moment | 11:03.03 |
| ManDay | every one publishes their stuff in PDF although they really don't give a damn about how it looks | 11:03.08 |
| chrisl | People use it because it works, it's reliable and widely supported | 11:03.17 |
| ManDay | they just think PDF sells well | 11:03.22 |
| | HTML works good, too. it may not always look as nice but at least I can still reflow it | 11:03.40 |
| | e-reader guy speaking, i guess | 11:03.49 |
| kens | Ah, that's better, now my user list is back | 11:04.01 |
| chrisl | But most people author documents to look a particular way, and they want it to look that way consistently wherever it is viewed | 11:04.38 |
| ManDay | chrisl: Maybe it's just me but I don't think the obsession about every pixel fitting right is worth the loss of information when you loose semantics. for books maybe, but not for everyday documents | 11:05.43 |
| | most prominent example would be specs | 11:06.18 |
| chrisl | it's just you..... | 11:06.30 |
| ManDay | so you have the postscript spec as pdf. looks so brilliant i want to hang it to my wall. but I can't properly search it, I can't reflow it, I can't mark it up and add links etc... | 11:07.24 |
| | what's the gain really? | 11:07.27 |
| kens | Several | 11:07.44 |
| chrisl | I can search mine.... and with the right tools, I can annotate it. | 11:08.20 |
| kens | For example i can refer people to a page number, I *can* seaqrch it, even in different languages. It looks the same as the author intended, etc | 11:08.21 |
| pedro_mac | guaranteed portabililty and fidelity. Thatâs what its all about | 11:08.25 |
| ManDay | kens: it's a digital document. you should forget about page numbers - they belong in printed press. | 11:09.28 |
| | if you want to link into a digital document, use a link, refer to a section if you must | 11:09.43 |
| kens | Its much easier to point to a page number | 11:09.59 |
| | And you are assuming that your 'digital document' has soem way of jumping staright to a point in the middle of the doucment, and that point is the same no matter what device I use | 11:10.38 |
| | ManDay : You are probably arguing in the wrong channel, we don't agree with you, we aren't likely to be swayed by any arguments you can present. To my mind ths comes perlously close to trollign and I would ask you to stop please. | 11:11.48 |
| ManDay | as for your question: yes, i'm assuming that. | 11:13.32 |
| rayjj | ManDay (for the logs): I saw some of the logs on using BeginPage, etc. using 'setpagedevice' in there is NOT a good idea (as was mentioned). For changing the CTM, using Install is better since it affects the initial CTM (so 'initgraphics' or grestoreall will use the modified CTM). | 13:48.42 |
| kens | rayjj if you want to chaneg the initial clip, you have little choice | 13:51.40 |
| | But I haven't investigated to see why the CropBox is being set like that, I suspect it something to do with setting the PageSize though | 13:52.54 |
| rayjj | what the heck. my gs-devel list delivery was disabled due to excessive bounces, but the email address is correct :-( | 13:58.49 |
| kens | We all had that last week | 13:58.59 |
| | THere are comments form me chris and Robin in the IRC log | 13:59.12 |
| | We believe that GMail (silently) bounced a mail with an attachment and we all got kicked off gs-devel | 13:59.45 |
| rayjj | and the 're-enable' reply gave me Bad confirmation string | 13:59.46 |
| | mine just happened today | 14:00.09 |
| | but, it does mention a bounce from Apr 24 | 14:01.49 |
| chrisl | I think that was the day it happened | 14:02.26 |
| rayjj | but the email telling me about it didn't happen until today, and so the confirmation string to re-enable is bad (probably because it's stale, but it doesn't say) | 14:08.02 |
| chrisl | Check in the headers? Maybe it got delayed somewhere? | 14:09.01 |
| rayjj | chrisl: I did -- all the dates in the headers are Fri May 1 | 14:10.04 |
| chrisl | Strange... anyway, it seems (based on my brief look) to be gmail's fault. | 14:11.19 |
| rayjj | chrisl: I re-entered my address, but don't know if I'm still disabled or not | 14:11.56 |
| | Robin_Watts: so what laptop did you finally decide on ? | 14:14.35 |
| Robin_Watts | Lenovo T450s. | 14:15.04 |
| | I went for the top processor option (all 3 processors have essentially the same power usage, so no meaningful battery life impact) | 14:15.44 |
| | It's got a 500Gig HD with a 16Gig SSD cache, plus I have another 256Gig SSD to put into it as a boot drive here. | 14:16.15 |
| rayjj | Robin_Watts: good luck with it. Do you actually have it ? | 14:16.16 |
| Robin_Watts | rayjj: no, it'll be another 8 working days or so. | 14:16.32 |
| rayjj | Robin_Watts: it has room for more than one HD ? | 14:16.41 |
| Robin_Watts | rayjj: No, just 1 HD. | 14:16.55 |
| | Google for M2 2242 SSD | 14:17.10 |
| | They are SSDs that are 22mm by 42mm. | 14:17.24 |
| | I was unaware of the M2 form factor until I started looking at this laptop. | 14:18.36 |
| | The SSDs are the attractive part that sets the laptop apart from the one chrisl got (which seems otherwise very similar) | 14:20.01 |
| | The twin batteries (one of which is hot swappable) was what attracted me to the model in the first place. | 14:20.21 |
| chrisl | I was planning to swap in an existing SSD anyway..... | 14:20.42 |
| Robin_Watts | chrisl: Yeah, but I want both speed and space. | 14:20.57 |
| chrisl | So how much faster are those SSDs then? | 14:21.22 |
| Robin_Watts | chrisl: Than a hard drive? lots. | 14:21.43 |
| | They are not faster than normal SATA based SSDs. | 14:22.00 |
| | (in fact, notionally a fraction slower, I think) | 14:22.09 |
| rayjj | Robin_Watts: so you are going to remove the 500Gb drive? | 14:22.20 |
| Robin_Watts | rayjj: For now, I'll live with a 256Gig SSD for the OS/tools etc on, and I'll leave the 500GB drive in there for data. | 14:22.50 |
| chrisl | I figured 500Gb was enough for the laptop.... | 14:23.00 |
| Robin_Watts | I may possibly up the 500GB drive at some point in the future. | 14:23.01 |
| rayjj | Robin_Watts: so it has room for the M2 SSD as well as the HD ? | 14:23.43 |
| Robin_Watts | rayjj: Yes. | 14:23.51 |
| | The M2 SSDs are tiny. | 14:24.03 |
| | less than 2 square inches, and as thin as a RAM module. | 14:24.24 |
| chrisl | rayjj: for gs-devel you are listed as being still a member, but disabled, let me check if i can reenable.... | 14:25.19 |
| rayjj | chrisl: thanks | 14:25.41 |
| | Robin_Watts: I looked at the "tech specs" on the 450s and it doesn't mention (that I could see) having an M2 capability | 14:28.54 |
| Robin_Watts | rayjj: Go to configure, and you should see the NGFF SSD option. | 14:29.41 |
| | http://shop.lenovo.com/us/en/laptops/thinkpad/t-series/t450s/#tab-customize | 14:30.51 |
| rayjj | ha ha. They offer to downgrade to to Win 7 64 Pro for an additional $50 | 14:31.11 |
| Robin_Watts | "Micro Hard Drive: M.2 Solid State Drive Single" | 14:31.22 |
| | and "System Expansion Slots: 3rd M.2 SSD Slot" | 14:31.38 |
| | technically, it's an upgrade to Win 8.1 Pro (and then a downgrade using the downgrade rights you get with that) | 14:31.56 |
| | rayjj: Annoyingly, in the UK, the prices are the same dollar to pounds :( | 14:32.13 |
| chrisl | rayjj: see private chat - if that doesn't work, might try the web interface: http://www.ghostscript.com/mailman/options/gs-devel and if that doesn't work, you might need marcosw to do it via the list admin web page | 14:57.46 |
| | rayjj: you're still listed as having delivery disabled :-( I don't know how often the database gets updated, though | 15:09.56 |
| ManDay | Given our previous discussion, I take it it's not possible to use GS to extract annotations from a PDF and copy them into another? | 15:27.32 |
| Robin_Watts | mvrhel_laptop: I've just bought a laptop with windows 8.1. | 15:36.12 |
| mvrhel_laptop | oh cool | 15:36.21 |
| Robin_Watts | As such, I could probably run MSVC2013 now. | 15:36.32 |
| mvrhel_laptop | yes | 15:36.39 |
| | kens are you still here? | 16:22.34 |
| kens | more or less... | 16:22.48 |
| mvrhel_laptop | I have a text/transparency question for you | 16:23.08 |
| kens | Not sure I will know the answer, but go ahead | 16:23.25 |
| mvrhel_laptop | Trying to understand section 5.2.7 text knockout | 16:23.29 |
| kens | 1 moment | 16:23.35 |
| | OK I hav it open | 16:24.01 |
| mvrhel_laptop | right now,have a single glyph that is getting filled with a pattern that has a shading | 16:24.03 |
| | it is getting drawn in a transparency group | 16:24.11 |
| kens | OK | 16:24.12 |
| mvrhel_laptop | and it seems to not do any blending with the underlying group. | 16:24.39 |
| | in acrobat | 16:24.53 |
| | but does in gs | 16:24.57 |
| kens | I don't htnk text knockout is anythjing to do with that | 16:24.59 |
| | GS could be wrong because of how we handle TK of course | 16:25.14 |
| mvrhel_laptop | reading the last section though in 5.2.7 | 16:25.16 |
| kens | With a single glyph TK has no effect | 16:25.22 |
| mvrhel_laptop | After the implicit transparency group for the text object has been completely evaluated, the group results are composited with the backdrop, using the Normal blend mode and alpha and soft mask values of 1.0. | 16:25.29 |
| kens | What;s the value of TK ? | 16:26.13 |
| mvrhel_laptop | well my setting it has no effect. | 16:26.24 |
| kens | I believe it should not, no | 16:26.31 |
| mvrhel_laptop | ok | 16:26.34 |
| kens | It only affects glyphs which overlap | 16:26.41 |
| | So for one glyph there can be no overlap | 16:26.50 |
| mvrhel_laptop | ok | 16:26.58 |
| | I am trying to understand why this glyph never blends with anything in Acrobat | 16:27.11 |
| | I have torn this file apart. | 16:27.19 |
| kens | TBH that sound swrong, but it depends on what's in the group | 16:27.37 |
| mvrhel_laptop | I did a rect fill in the same group and that is blending | 16:27.47 |
| | but the glyph just blows knocks out everything | 16:28.05 |
| kens | The alpha applied to the text group is 1.0, which is pure opaque isn';t it ? | 16:28.07 |
| mvrhel_laptop | yes | 16:28.20 |
| | that is what the last part of section 5.2.7 says | 16:28.38 |
| | but I may be applying this wrongly | 16:28.48 |
| kens | OK so we should draw the text group and blend that with the 'backdrop' as opaque | 16:28.53 |
| mvrhel_laptop | yes | 16:28.59 |
| kens | But, if that 'implicit' group is inside another group, then the result of that needs to be blended with any underlying eisting content, according to the enslocing group | 16:29.33 |
| | But I freely admit I don't understand transparency any more | 16:29.51 |
| mvrhel_laptop | right. oh let me draw a rect in the next level up and see if that is blending properly | 16:30.08 |
| | I have to go to daughters school now | 16:30.21 |
| | thanks for the chat kens! | 16:30.25 |
| | this stuff is really starting to piss me off | 16:30.38 |
| kens | NP | 16:30.45 |
| | Forward 1 day (to 2015/05/02)>>> | |
IRC Logs