| | <<<Back 1 day (to 2014/09/08) | 2014/09/09 |
| kens | LOL Microsoft wanting to know if we support Ghostscript 8.5 (!!) on various Windows platforms :-) | 09:21.25 |
| tor8 | kens: well, it *is* a fair bit newer than Windows XP, which they just recently stopped supporting :) | 09:23.38 |
| kens | Well, I just told them they didn't have support for 8.50 ever, because they're a free user.... | 09:24.20 |
| chrisl | I like the "If you are interested in obtaining a licence for Ghostscript........" | 09:25.23 |
| kens | I elt I shoudl mention it :-) | 09:25.39 |
| chrisl | It leaves me a little conflicted.... I dislike M$, but they could pay us a *lot* of money ;-) | 09:26.22 |
| kens | I very much doubt they will pay us anything :-( | 09:26.41 |
| chrisl | No, this feel more like an opening gambit for "can we interest you in M$ certification for your application.....?" | 09:27.29 |
| kens | Well, it could be, but I checked out the email and he does work for MS, in India, and its conceivable they use Ghostscript for something. Possibly for printing to non-PostScript printers on Windows systems. | 09:28.34 |
| | Hmm, someone in Russia trying to hack my Microsoft Live account | 13:07.44 |
| | chrisl I'm thinking its time to stop shipping the .inf file with the Windows version of Ghostscript.... | 13:20.18 |
| davesleep | Error: /rangecheck in --run-- there's the error? | 13:30.04 |
| | is there anything else i need? | 13:30.07 |
| kens | for what ? | 13:30.16 |
| davesleep | to debug the error | 13:30.23 |
| | im trying to figure out what im supposed to be googling for in this error | 13:30.33 |
| kens | Nothing I wouldn't think. Its telling you your PostScript program has thrown a rangecheck error.Some parameter passed to an operator was outside the valid range | 13:31.12 |
| | A rangecheck in 'run' probably menas that some external file was executed, and that was the file which threw the error. | 13:32.03 |
| davesleep | i got more error but it looks like gibberish to me | 13:32.59 |
| kens | THat wil be the operand stack dump | 13:33.16 |
| | Ah, so to add some context back into the problem, you were using ImageMagick and getting an error ? | 13:37.22 |
| | The only way we can likely help you here is if you point us to a copy of the file you are using as input, tell us the command line you are using and which version of Ghostscript (and which OS/word size) | 13:38.18 |
| | Looks like (from your conversaation on Firday) you are using a PDF file as input, so pretty much all bets are off regarding errors, because the PDF interpreter tries ot ignore errors (ie broken PDF files) and carry on. Sometimes it can't but the error it throws may not be related to the real error, whch occured some time previously and was ignored. So the only way we can offer you any assistance is to see the original input file. | 13:41.37 |
| kens | wonders if he's talking to himself..... | 13:45.22 |
| bashseb | good day, everyone. I'm currently trying to debug an unhandled exception in `pdf_open_document_no_run`. Can I ask some questions? | 14:07.25 |
| kens | THIs is MuPDF ? | 14:07.37 |
| tor8 | kens: looks like it, from the function name | 14:07.46 |
| Robin_Watts | bashseb: Sure. First off, are you using the latest version from git ? | 14:07.56 |
| bashseb | yes, I'm accessing it through zathura though. | 14:08.01 |
| | yes | 14:08.10 |
| kens | bashseb : feel free to ask, but be aware we have a meeting here in 20 minutes, and it might get confusing | 14:08.18 |
| bashseb | ok, thanks | 14:08.28 |
| Robin_Watts | bashseb: OK, so none of us know anything about zathura. | 14:08.29 |
| tor8 | mupdf uses its own exception mechanism based on setjmp | 14:08.37 |
| | we have a set of macros that you need to use to wrap function calls to mupdf at the top level | 14:09.03 |
| | if you forget that, then the code doesn't know where to return error messages, and you see 'unhandled exception' instead | 14:09.27 |
| | I haven't looked at the code, but I suspect that either you or zathura have forgotten to wrap a call in fz_try/fz_catch | 14:10.04 |
| bashseb | I can step into mupdf and once the excpetion in `source/pdf/pdf-xref.c` is raised, It is not caught anymore | 14:10.19 |
| Robin_Watts | bashseb: See mudraw.c line 1038 | 14:10.38 |
| | The call to fz_open_document there is surrounded by an fz_try(ctx) { ... } | 14:10.54 |
| tor8 | bashseb: yeah, at least one of the functions in the stack trace from that needs to have used fz_try. can you paste a stack trace? | 14:10.59 |
| Robin_Watts | and followed by an fz_catch(ctx) { ... } | 14:11.07 |
| | When an exception is raised within fz_open_document control instantly jumps back to the fz_catch() for then next highest enclosing fz_try(). | 14:11.58 |
| | If there isn't one, you get an 'unhandled exception' error. | 14:12.09 |
| bashseb | @Robin_Watts: line 1038 in current git? | 14:12.47 |
| Robin_Watts | I suspect that zathura doesn't have an fz_try(ctx) { } around it's call to fz_open_document_no_run | 14:12.49 |
| | in my current source, which is pretty much current git. | 14:13.01 |
| | Look for fz_open_document | 14:13.15 |
| bashseb | zathura doesn't do it. correct. | 14:13.15 |
| Robin_Watts | So that's your problem. | 14:13.21 |
| | Add an fz_try(ctx) { ... } fz_catch(ctx) {} and you'll be sorted. | 14:13.35 |
| bashseb | understood. I'll give it a shot. | 14:13.44 |
| | reporting back in a couple of mins | 14:13.49 |
| Robin_Watts | rayjj: MuPDF does not honour TR at all. In SMask or otherwise. | 14:14.37 |
| rayjj | Robin_Watts: I'm not surprised. Since I confirmed that it is a problem from a real (gs) customer file, I went ahead and opened a bug. | 14:15.25 |
| | Robin_Watts: I assume there is an FTS where mupdf gets it wrong if it doesn't support TR at all | 14:15.51 |
| henrys | wow we'll be coming to the UK in the midst of a civil war... will it be safe? | 14:15.56 |
| kens | What civil war ? | 14:16.07 |
| Robin_Watts | There are various files. | 14:16.11 |
| rayjj | henrys: well, we'll be far south of the "border" | 14:16.21 |
| kens | Can't be any more dangerous than visiting Texas anyway...... | 14:16.32 |
| Robin_Watts | we'll be shipping kens and chrisl back north in 2 weeks time :) | 14:16.47 |
| kens | still expects a no vote.... | 14:17.01 |
| henrys | kens: can't imagine that it would. | 14:17.02 |
| pedro_mac | Robin: the borders are closed ;) | 14:17.37 |
| kens | I'd be happy to head back North, but Steall wouldn't agree. | 14:17.54 |
| | STella* | 14:17.57 |
| Robin_Watts | pedro_mac: Well, you'll be shipping Matt and Joseph south again, right? :) | 14:18.06 |
| henrys | are the scots going to use the euro? | 14:18.13 |
| rayjj | any fighting would be with clubs and bows and arrows (like the first time) since guns are (relatively) hard to come by | 14:18.20 |
| Robin_Watts | The Euros won't have 'em. | 14:18.25 |
| kens | henyrs, who knows ? :-) | 14:18.25 |
| henrys | if the vote goes through | 14:18.29 |
| kens | Robin_Watts : doens't stop them using it. | 14:18.34 |
| pedro_mac | I reckon I couldnât deal with doing customs on the train, so Iâm out ;) | 14:19.01 |
| kens | The Irish used the pound for ages before they established the punt (or however they spelled it) | 14:19.09 |
| rayjj | they want the stirling is what I heard (who would want to be Euro based?) | 14:19.18 |
| kens | Stirling is a place, perhaps you mean Sterling ? | 14:19.34 |
| rayjj | kens: oh, yeah, that sterling | 14:19.48 |
| kens | oR MAYBE IT WAS A PUN :-) | 14:19.49 |
| | oops caps | 14:19.53 |
| Robin_Watts | kens: The Scots can pick any currency they want and use it informally, but such informal currency unions (aka "sterlingisation") are a dangerous game. | 14:19.56 |
| kens | Robin_Watts : yep, but it doesn't stop them using it.... | 14:20.12 |
| Robin_Watts | kens: It doesn't stop them doing it, no. The runaway cost of borrowing that would ensue might however. | 14:20.47 |
| kens | shrugs | 14:21.00 |
| | not my problem | 14:21.04 |
| henrys | I've read some about but it isn't clear what the scots want? Political decision making? | 14:21.21 |
| rayjj | kens: not until you have to get a new passport ;-) | 14:21.22 |
| Robin_Watts | rayjj: I'm pushing for any new scottish currency to be called the groat. | 14:21.23 |
| | henrys: They want the moon on stick. | 14:21.40 |
| kens | rayjj : I'm a UK citizen currently, that won't change. I could probably apply for dual citizenship though | 14:21.47 |
| | henrys, severl things | 14:22.02 |
| tor8 | henrys: freedom from people like boris and farage, I'd guess :) | 14:22.09 |
| kens | Mostly not being told what to do by the English | 14:22.15 |
| Robin_Watts | quietly explodes. | 14:22.36 |
| | We are part of a union. This means that random parts of the country are always going to get a government they don't want. | 14:23.30 |
| henrys | kens: that sounds like everything I've read and motivated my original question about it not being clear. | 14:23.35 |
| Robin_Watts | The scots don't vote tory, hence whinge constantly when they have a tory government. | 14:23.57 |
| sebras | Robin_Watts: how is this different from you not wanting to be part of EU? | 14:24.13 |
| kens | No, mostly the Scots dislike being ignored, or treated as useful experimental animals | 14:24.19 |
| sebras | walks away.... | 14:24.20 |
| henrys | anyway 5 minutes to the meeting. need some coffee | 14:24.26 |
| sebras | ;) | 14:24.27 |
| Robin_Watts | On the other hand, the english countryside doesn't vote labour, so we endure governments we don't want. | 14:24.38 |
| | It's all part of being a union. | 14:24.44 |
| kens | Yes, and if you don't like it, you can leave :-) | 14:25.00 |
| Robin_Watts | sebras: I would consider it a shame if the union broke down, but if the scots want out, then so be it. | 14:25.13 |
| tor8 | kens: isn't that what started this whole discussion? :) | 14:25.18 |
| kens | :-) | 14:25.23 |
| rayjj | I suspect they want to gain control over the oil fields. Sort of why No CA wants to split from the south, since they have more of the water | 14:25.35 |
| Robin_Watts | What I would consider an absolute travesty is Devo Max. | 14:25.36 |
| pedro_mac | actually likes Boris - heâs great value for money | 14:25.42 |
| kens | Boris is good entertainment value | 14:25.54 |
| Robin_Watts | Either we're all part of a union or we're not. It's wrong that the scots get double representation. | 14:25.57 |
| kens | Pictures of him dangling on a wire amused me for ages | 14:26.07 |
| Robin_Watts | It's wrong that they get 22% per head more spent on them (by the barnett formula). | 14:26.17 |
| pedro_mac | Robin: absolutely - I donât think any of these arguments are specific to Scotland | 14:26.32 |
| Robin_Watts | The barnett formula is pretty bloody specific to scotland. | 14:26.47 |
| pedro_mac | if we want devolved power within the UK then eevry region should be entitled to it | 14:26.52 |
| Robin_Watts | right. | 14:27.01 |
| | and I would disapprove mightily of that. | 14:27.08 |
| pedro_mac | nods | 14:27.14 |
| bashseb | RobinWatts: et al. Thanks Robin and others, I could solve my issue. Now I need to solve the resulting double free. Have a nice meeting. | 14:28.04 |
| Robin_Watts | Adding extra layers of government has never produced better more efficient governance. It's just produced more bloody civil servants, more politicians, more waste,... | 14:28.06 |
| | bashseb: Fab. | 14:28.20 |
| rayjj | Robin_Watts: sort of what we have here with state governments and federal | 14:28.34 |
| Robin_Watts | rayjj: Yes. Though actually I think the US is a model of how to do it more or less right. | 14:29.01 |
| | Where Europe is a classic example of how to do it wrongly. | 14:29.22 |
| rayjj | Robin_Watts: I hardly think so | 14:29.25 |
| | that the US has it right -- "state rights" create a morass of conflicting laws | 14:29.55 |
| Robin_Watts | rayjj: Yes, I see that. | 14:30.11 |
| rayjj | and improbable things like legal marijuana in some states, but seriously illegal according to the feds | 14:30.38 |
| henrys | rayjj: a necessary evil most would agree. How is one govt going to govern alabama and california. | 14:31.05 |
| rayjj | and myriad taxing authorities | 14:31.06 |
| Robin_Watts | rayjj: Yes, it's imperfect, but on the whole it mostly works. | 14:31.06 |
| kens | Meeting time ? | 14:31.18 |
| henrys | kens: yes I was going to give michael and marcos a few more minutes but let's being without them | 14:31.47 |
| kens | chrisl is also not here | 14:31.56 |
| | No problem waiting a bit more | 14:32.05 |
| henrys | Robin_Watts: nothing from our ebub customer? | 14:32.12 |
| Robin_Watts | henrys: No, nothing at all. | 14:32.23 |
| henrys | or epub but I kind of like ebub... | 14:32.31 |
| Robin_Watts | ebub is the hillbilly version. | 14:32.41 |
| kens | Back to Texas again..... | 14:32.44 |
| henrys | ebubba | 14:32.48 |
| rayjj | epub makes me think of ordering beer over the net :-) | 14:32.49 |
| henrys | hi marcosw | 14:33.02 |
| marcosw | morning henrys | 14:33.08 |
| henrys | releases! woo hoo! | 14:33.14 |
| marcosw | yes, testing in progress, should be done on Friday but I don't expect any issues. | 14:33.35 |
| henrys | chris is out right? | 14:33.39 |
| kens | Seems so | 14:33.45 |
| | But the RC is available | 14:33.51 |
| marcosw | there hasn't been a last minute rush to get stuff in, so 9.15rc1 is basically the same code we've been testing for weeks. | 14:34.13 |
| kens | Well.... I did stick a few things in hurrieldy, but they are low risk | 14:34.38 |
| henrys | marcosw: well that's good kens had a few big recent changes though | 14:34.49 |
| rayjj | I notice that the only "critical" bugs are P4 :-) | 14:34.59 |
| mvrhel_laptop | morning. Sorry to be late | 14:35.37 |
| kens | morning mvrhel_laptop | 14:35.45 |
| rayjj | we should explain on bugzilla that "critical" means "trivial" to us | 14:35.54 |
| henrys | rayjj: how goes the boards, miles always bugs me about that, if we aren't making progress we should come up with a new strategy like getting them to help us. | 14:35.59 |
| | rayjj, marcosw : can we fix the critical bug ? | 14:36.37 |
| | I mean make it not critical or P1 | 14:36.51 |
| rayjj | henrys: getting them to help would be nice, but I can't even get them to give me passwords to some of the third party docs (the network stack). | 14:36.53 |
| mvrhel_laptop | sounds like perhaps miles needs to make a call then | 14:37.39 |
| henrys | rayjj: that was the problem last week right? Did you call them? mvrhel_laptop has a connection right? | 14:38.00 |
| mvrhel_laptop | not there | 14:38.07 |
| | at company M | 14:38.12 |
| | everyone I knew at C moved to M | 14:38.18 |
| rayjj | henrys: and the latest is that the free eval license for the toolchain (from ARM) is elapsed, so I have to try to switch stuff over to a different toolchain. I've requested an extension and/or licensing quote from ARM with no response | 14:38.23 |
| henrys | mvrhel_laptop: oh he did work for C though right? | 14:38.37 |
| mvrhel_laptop | yes | 14:38.41 |
| rayjj | Company M is A--OK. I'm on their extranet | 14:38.43 |
| | company C is the problem | 14:38.53 |
| Robin_Watts | rayjj: ARMs licensing is extortionate. | 14:38.59 |
| henrys | rayjj: so maybe just focus on the M board and that's what Miles gets. | 14:39.04 |
| | tor8: have you started epub? | 14:39.28 |
| rayjj | Robin_Watts: yeah, but I can't even get a quote. Searching the web says that a node locked license is $7,500 | 14:39.40 |
| Robin_Watts | rayjj: Yeah. | 14:39.51 |
| rayjj | Robin_Watts: but they won't even respond :-( | 14:40.08 |
| Robin_Watts | In the past I've always worked with the codesourcery stuff. | 14:40.08 |
| marcosw | done. no more 'critical' bugs (as noted none of them were actually critical). | 14:40.16 |
| Robin_Watts | rayjj: This all sounds very familiar from when I was working on their boards. | 14:40.24 |
| henrys | moving on to the mupdf release - miles wants a paid android app sadly. Can that be concurrent with the regular mupdf release? Robin_Watts, tor8 | 14:40.37 |
| | marcosw: thanks | 14:40.58 |
| rayjj | Robin_Watts: I thought you were working on their linux build system | 14:41.02 |
| Robin_Watts | Making the standard app paid for is trivial I think, we just use the same apk, but tweak the android developer console settings,I think. | 14:41.26 |
| henrys | marcosw: did you get my email about PAM? I didn't google around for it. Do you want me to? | 14:41.29 |
| Robin_Watts | rayjj: I was dealing with linux, but my point was that they don't respond to emails in a useful way. | 14:41.48 |
| rayjj | marcosw: thanks. I realized that they weren't really critical | 14:41.50 |
| tor8 | henrys: yes, I have a started a branch for doing epub | 14:41.56 |
| | and integrated the existing parsers into the mupdf framework | 14:42.09 |
| rayjj | Robin_Watts: right. I only have one manager contact at company C -- it would be MUCH nicer to have an engineer contact, but I suspect they don't have anybody left (probably just a team outsourced across the Pacific from me) | 14:43.25 |
| Robin_Watts | tor8: New version the abort commit on robin/master | 14:43.44 |
| tor8 | henrys: if we make the mupdf app paid, won't just one of the clones come eat our lunch and mupdf completely disappear into oblivion? | 14:43.46 |
| | but we've been over this a dozen times before... *shrug* | 14:44.20 |
| Robin_Watts | AIUI, the amount we've made from making the ios app chargeable is stupidly small. | 14:44.24 |
| | so we have that as evidence that making the apps chargable doesn't work as a revenue generator. | 14:45.07 |
| tor8 | Robin_Watts: indeed. why would you pay for mupdf when you can use the built-in viewer for free? | 14:45.13 |
| henrys | android isn't as damaging we can have a free apk on mupdf.com and folks will use it. | 14:45.25 |
| tor8 | and for android, there are dozens of mupdf-based open source viewers on the app store | 14:45.30 |
| | henrys: yeah, but that will be invisible from the android marketplace where 99% of people look for apps | 14:45.56 |
| | henrys: but it is true for those of us who know about side-loading apps | 14:46.13 |
| rayjj | are any of the other free ones actually AGPL compliant ? (i.e., provide source) ? | 14:46.18 |
| tor8 | rayjj: I believe sebras has occasionally gathered up a list of the non-compliant ones and sent to miles | 14:46.50 |
| Robin_Watts | It means we can still point potential customers at a free version, but let's face it, if the potential customer contact is a manager of similar technical competency to miles, do you think they could figure out how to install an apk from our website? | 14:47.09 |
| henrys | my current plan is to do this for a bit show him a few months of numbers and hope to talk him out of it. I don't see what else I can do. | 14:47.35 |
| Robin_Watts | henrys: That seems reasonable. | 14:47.43 |
| | Can we all wear "Free MuPDF" T-shirts in december? :) | 14:48.10 |
| henrys | ;-) | 14:48.28 |
| | tor8: do we have a mupdf release ready to go? | 14:48.45 |
| Robin_Watts | The color vs greyscale detector for MuPDF is now pretty much as good as it's going to get. | 14:49.05 |
| | When the last commit goes in, we should meet the customers performance requirements. | 14:49.17 |
| rayjj | maybe the ploy is to establish a 'value' in order to go after the violators, since we can there is established 'damages' per unit they ship | 14:49.23 |
| henrys | shall we give paulgardiner a break from sot to do an android release ? | 14:49.45 |
| tor8 | henrys: no, but the code is in a pretty stable state so we should be able to just slap a number on it and release as is | 14:50.19 |
| henrys | rayjj: I believe if we can show it doesn't make money we can go back. | 14:50.37 |
| tor8 | Robin_Watts: I see nothing new when I pull | 14:50.44 |
| rayjj | rayjj: that, too | 14:50.49 |
| Robin_Watts | tor8: Try now, sorry. | 14:51.16 |
| kens | rayjj talking to yourself now ? :-) | 14:51.17 |
| henrys | anything else for this meeting? I'm done. | 14:51.31 |
| | bbiam | 14:52.52 |
| mvrhel_laptop | nothing from me. other than let me get the gsview merged in today before the mupdf release occurs | 14:53.05 |
| tor8 | Robin_Watts: looks almost acceptable -- two questions though: FZ_ABORT enum in device.h isn't used anywhere, and I don't see where the abort field in pdf_csi_s in pdf-interpret-imp.h is used | 14:53.09 |
| Robin_Watts | The abort field in the cookie? It's not used, and it shouldn't be. | 14:53.28 |
| sebras | reads logs. | 14:53.32 |
| tor8 | Robin_Watts: not cookie, pdf_csi struct | 14:53.42 |
| Robin_Watts | oh, was that a hangover from my first try I wonder? | 14:53.59 |
| marcosw | henrys: I did see your pam stuff, but wasn't sure what it meant. Google says that the file /etc/pam.d/sshd contains permissions for ssh logins, but I don't know why that would have changed. I compared my pam.d/sshd files on my Mac OS X 10.8.5 and 10.9.3 systems and they are identical. | 14:54.07 |
| tor8 | Robin_Watts: that would make the most sense :) | 14:54.16 |
| Robin_Watts | You're right about FZ_ABORT | 14:54.26 |
| mvrhel_laptop | brb | 14:54.34 |
| Robin_Watts | and about abort. I will fix those. Thanks. | 14:54.44 |
| tor8 | Robin_Watts: there are two commits on tor/master as well | 14:55.16 |
| chrisl | henrys: sorry, missed the meeting - an extra 30 miles on UK roads is a time killer.... :-( | 14:57.32 |
| Robin_Watts | tor8: New commit on robin/master then. Looking at yours now. | 14:57.33 |
| kens | chrisl, I assumed it was deliberate :-) | 14:57.56 |
| tor8 | Robin_Watts: LGTM | 14:58.18 |
| Robin_Watts | Ta. | 14:58.24 |
| henrys | marcosw: i'll look some more but that dns stuff does not work or the cron was wrong it was spawning many instances of itself and bad stuff started happening. | 14:58.24 |
| chrisl | kens: well, my only contribution was going to be that the (first) RC is done......# | 14:58.44 |
| kens | Ah well, I mentioned that for you | 14:58.57 |
| chrisl | Ta! | 14:59.04 |
| sebras | tor8: rayjj: I did send a list of apps that didn't appear to deliver source to miles, yes. | 14:59.14 |
| henrys | chrisl: I figured you were off conspiring with the scots | 14:59.17 |
| aksr | hi guys, is it possible to rescale an a4 pdf to a6 pdf or something like that? | 14:59.32 |
| kens | Yes | 14:59.41 |
| chrisl | henrys: Ha! No, there was a power outage at the squash club, so we went to another club, but I hadn't taken into account the extra travelling time | 14:59.53 |
| aksr | kens: hi, how good is it? | 14:59.54 |
| kens | Its as good as any PDF->PDF processing with Ghostscript and pdfwrite | 15:00.23 |
| marcosw | henrys: I understand but that code hasn't changed in years and I run the same stuff on most of my machines (including my macbookpro) and have never seen the behavior you describe. Presumably a recent change on your macpro affected the ddclient software, which I'd be happy to diagnose if I could log in :-) | 15:00.35 |
| aksr | kens: i'm trying to read something on my kindle | 15:00.56 |
| henrys | tor8: so can the mupdf'ers figure out who is doing the android release? | 15:01.17 |
| marcosw | Are we having a sot meeting now? If so I need to change computers. | 15:01.21 |
| henrys | marcosw: that was the plan. I'll try to be quick | 15:01.35 |
| Robin_Watts | henrys: We don't always release the android ones at the same time. | 15:01.46 |
| aksr | kens: will you give me one working example (a4->a6)? | 15:01.50 |
| chrisl | henrys: has the macpro ended up with a circular symlink in the file system - I've seen that cause infinite grep problems...... | 15:01.51 |
| Robin_Watts | oh, but miles wants it chargable. | 15:01.55 |
| | I can do new builds if you want. | 15:02.03 |
| tor8 | henrys: paulgardiner or Robin_Watts usually do the android release, but sometimes a week or so after the main release. | 15:02.16 |
| paulgardiner | I've released the iOS version before, but not the Android one I think | 15:03.01 |
| henrys | do we know how to make it a paid app? | 15:03.25 |
| kens | aksr, no I can't give you a worked example. What I will tell you is read the documentation on setting the mdeia size (there are several ways to achieve this) and -dFIXEDMEDIA as well as -dPDFFitPage, you will need all to set the target media size, make it fixed, and have the interpreter scale the PDF to the target size. | 15:03.25 |
| aksr | kens: tried both things, doesn't work so far | 15:03.57 |
| tor8 | aksr: which kindle? | 15:04.06 |
| aksr | k3 | 15:04.09 |
| | keyboard | 15:04.12 |
| kens | You need all 3, not 'both' | 15:04.14 |
| tor8 | aksr: ah, the crappy one that can't 'fit page', IIRC | 15:04.33 |
| aksr | don't be foolish, i googled i tried a dozen examples so far... | 15:04.41 |
| henrys | marcosw_: and what happened to using our old artifex DNS address? Why do we need more DNS? | 15:04.47 |
| aksr | tor8: :( | 15:04.48 |
| Robin_Watts | henrys: I think it's just a setting in the developer console. I will look into it. | 15:04.58 |
| kens | foolish eh ? *plonk* | 15:05.06 |
| aksr | ;) | 15:05.12 |
| | be helpful | 15:05.16 |
| | :D | 15:05.19 |
| tor8 | aksr: you could try converting the document to a CBZ of the appropriate resolution (600x800) and then convert that to an ebook | 15:05.31 |
| marcosw_ | henrys: I'm not sure what you are asking. what other dns? | 15:05.48 |
| rayjj | henrys: what's the DNS issue ? | 15:05.50 |
| tor8 | you won't be able to zoom or select text, but if you just want to read that should work | 15:05.56 |
| aksr | tor8: i tried k2pdfopt | 15:06.16 |
| | i'm not content with it | 15:06.25 |
| | ..at all... | 15:06.29 |
| Robin_Watts | tor8: Your commits lgtm. | 15:06.36 |
| aksr | it breaks down equations etc. | 15:06.40 |
| tor8 | Robin_Watts: Ta. | 15:06.41 |
| Robin_Watts | in so far as I understand them. | 15:06.42 |
| henrys | marcosw_: you set up this henry-artifex.homeip.net, why do I need another name? | 15:07.03 |
| tor8 | aksr: render to a bitmap using gs or mupdf, say 'mudraw -w 600 -h 800 -o page%03d.png input.pdf' | 15:07.25 |
| | aksr: zip up the pages into a CBZ: 'zip out.cbz page*.png' | 15:07.48 |
| Robin_Watts | henrys: Is the ddclient thing not the thing that keeps henry-artifex.homeip.net up to date? | 15:07.56 |
| tor8 | aksr: then let calibre convert that into a MOBI | 15:07.58 |
| aksr | tor8: i'm trying to avoid calibre | 15:08.17 |
| tor8 | aksr: ebook-convert from the command line, and you won't have to touch its atrocious UI | 15:08.40 |
| henrys | Robin_Watts: I don't know marcos said he's now using henrys.ghostscript.com | 15:08.45 |
| aksr | tor8: yes, i converted a dozen so far | 15:08.52 |
| henrys | which I assume is associate with ddclient | 15:09.02 |
| aksr | if book contains equations, it's more than bad | 15:09.12 |
| | but not bad for novels :) | 15:09.23 |
| | tor8: yes, that gui is horrible | 15:09.37 |
| Robin_Watts | henrys: something needs to run on your machine (or on your network at least) to update henry-artifex.homeip.net whenever your IP address changes. Presumably the same thing can also update henrys.ghostscript.com | 15:09.37 |
| tor8 | yeah, the calibre internal representation of text is, ahem, not perfect | 15:09.46 |
| aksr | kens: i tried this so far: gs -sDEVICE=pdfwrite -sPAPERSIZE=a6 -dFIXEDMEDIA -dPDFFitPage -dCompatibilityLevel=1.4 -o output.pdf input.pdf | 15:10.06 |
| tor8 | but if you do what I said, you'll use mupdf to pre-render at the kindle resolution then create an ebook of images | 15:10.07 |
| | which should render all fancy pdf content properly | 15:10.31 |
| aksr | hm | 15:10.37 |
| kens | aksr, can't comment without seeing your PDF file also | 15:10.41 |
| tor8 | and PDF is a much more general format than HTML which is what mobi contains | 15:10.59 |
| | equations drawn in pdf don't necessarily have a useful translation to HTML | 15:11.22 |
| henrys | Robin_Watts: I thought there was something else for the artifex address. Is there marcosw_ ? I don't want 2 notifications running. | 15:11.24 |
| aksr | tor8: i know, that's why | 15:11.31 |
| | let me try it your way | 15:11.43 |
| tor8 | aksr: you might want to use a smaller size, depends on how much space the margins and page indicator take up | 15:12.30 |
| | 600x800 is the physical screen size of the old kindles | 15:12.53 |
| marcosw_ | henrys: the *-artifex.homeip.net names are used for dynamic dns, via dyn.com. when you ip address changes the ddns gets updated (if you look in /etc/bind/master/ghostscript.com on casper you'll see there is a cname pointing to henry-artifex.homeip.net). there entrys for various *-artifex.homeip.net systems in that file (note that miles.ghostscript.com is no longer one of these, since that's now an actual static ip). it's been s | 15:13.24 |
| | this way for 2 or 3 years. | 15:13.25 |
| tor8 | aksr: http://www.amazon.com/gp/feature.html?ie=UTF8&docId=1001103761 might be useful, but caveat emptor since I haven't tried it | 15:13.56 |
| marcosw_ | henrys: I don't have access to dns for artifex.com, so there won't be an entry (unless someone else set it up). | 15:14.21 |
| | Robin_Watts: I don't think we want to run our own ddns service, I think using dyn.com and a cname in master/ghostscript.com is the better solution. | 15:15.02 |
| henrys | marcosw_: and ddclient does this updating on my side? | 15:15.05 |
| kens | For me gs -sDEVICEWIDTHPOINTS=306 -sDEVICEHEIGHTPOINTS=396 -dFIXEDMEDIA -dPDFFitPage -sDEVICE=pdfwrite -sOutputFIle=out.pdf annots.pdf produces a PDF file where hte media is scaled to 4.25x5.50 inches the original being A4 | 15:15.35 |
| marcosw_ | henrys: yes. And again, I run them on about 7 or 8 computers and have never seen the behaviour you describe. It just checks your external ip address every 15 minutes and sends an update to dyn.com if it's changed (or once a month, if it hasn't changed). | 15:16.13 |
| henrys | marcosw_: okay I'll try and fix the PAM thing | 15:16.57 |
| Robin_Watts | marcosw_: I was not suggesting we should run our own ddns service. | 15:17.26 |
| aksr | kens: i used -sPAPERSIZE | 15:17.51 |
| Robin_Watts | I was just saying that it was entirely possible that the dnsclient stuff on henrys machine *was* the stuff that we were relying on. | 15:17.54 |
| aksr | will try that too | 15:18.08 |
| kens | aksr, I didn't, mine worked :-P | 15:18.15 |
| | Maybe you should try it my way | 15:18.15 |
| marcosw_ | Robin_Watts: sorry, misread your suggestion that we should update ghostscript.com dynamically as that. | 15:18.20 |
| aksr | kens: ;) | 15:18.44 |
| marcosw_ | I don't know enough about dns to be certain that using a cname is the correct thing to do in this case, but it seems to work (famous last words). | 15:18.46 |
| henrys | marcosw_: you probably don't need to set up your other computer - I was just going to ask about the testing stuff? Is that on hold? | 15:23.14 |
| marcosw_ | you mean ats? It's all setup in my garage an afaik running correctly. I updated the twiki with instructions on how to access it (the same as the current setup with a different ssh tunnel). | 15:24.09 |
| aksr | kens: i get this: Unrecoverable error: typecheck in .putdeviceprops | 15:24.53 |
| kens | Possibly you used -d instead of -s for DEVICEHEIGHTPOINTS or DEVCIEWIDTHPOINTS ? | 15:25.31 |
| | Err no actually thats atypo on mjy part, they should both be -d | 15:25.59 |
| aksr | i copy-pasted your example | 15:26.01 |
| kens | They are -d because they are numeric only | 15:26.19 |
| | I wouldn't copy/paste command lines from me off of IRC, I do typos a lot here | 15:27.29 |
| aksr | kens: gs -dDEVICEWIDTHPOINTS=306 -dDEVICEHEIGHTPOINTS=396? | 15:27.33 |
| kens | Yes. | 15:27.41 |
| | And all the rest of the command line obviously | 15:27.49 |
| aksr | kens: of course, now, this: **** Unable to open the initial device, quitting. | 15:28.03 |
| kens | I don't think that's A6 though, maybe close to A5 | 15:28.05 |
| aksr | doesn't matter, just wanted to try it | 15:28.20 |
| kens | I just wanted a test which was different to thte source file | 15:28.23 |
| aksr | me too | 15:28.33 |
| | kens: >>> **** Unable to open the initial device, quitting. | 15:29.08 |
| | what's now the problem? | 15:29.21 |
| kens | SOunds like the filename is wrong, possible capitalisation in -sOutputFile ? | 15:29.26 |
| | Yes my command line has a capital I in File | 15:29.43 |
| rayjj | sounds like aksr didn't include the rest of the command line | 15:29.51 |
| kens | Like I said, don;t copy/paste my command lines | 15:29.54 |
| aksr | kens: you're right, wrong capitalisation | 15:30.51 |
| | rayjj: maybe it sounds like that, but, it ain't true :D | 15:31.08 |
| kens | Its not hard to guess when you mistype it as often as I do | 15:31.17 |
| aksr | i'm pretty much tired | 15:32.10 |
| | i can't think straight | 15:32.19 |
| | tor8: cbz to mobi with ebook-convert, right? | 15:34.31 |
| | kens: rescaling should make it more readable, right? | 15:37.54 |
| kens | Eh ? No why would it have that effect ? All it does is scale it to the size you asked for | 15:38.25 |
| | The data is (usually) unchanged | 15:38.43 |
| aksr | i know, i mean on kindle | 15:39.12 |
| kens | Thoough the usual caveats about complete interpretation to graphics primitives and reassembly form graphics primitives to PDF apply here | 15:39.20 |
| | aksr I can't really comment on what happens on the kindle. | 15:39.32 |
| aksr | i know, just asked for your opinion | 15:40.10 |
| kens | But since you are scaling down, the resulting text will be smaller (on the same media at the same resolution) so it seems unlikely wit would be more readable | 15:40.18 |
| | Bigger text is usually easier to read I find, as long as you don;t go to silly sizes | 15:40.46 |
| | On your Kindle each glyph will clearly cover fewer pixels, so the text will be 'coarser' which will be less readable, IMO | 15:41.38 |
| | Of course, anti-aliasing or other effects can cause text to be more or less readable | 15:42.09 |
| | Possibly I'm missing your point though | 15:42.40 |
| aksr | i have a few books in LaTeX, formated for kindle. they are more than good. | 15:43.29 |
| | pdf, formated for kindle is the real deal. | 15:43.39 |
| chrisl | PDF should be device independent..... | 15:45.35 |
| mvrhel_laptop | off to dentist. hopefully back shortly | 15:45.43 |
| aksr | chrisl: it is, i meant papersize | 15:46.21 |
| | chrisl: think about it: a4 pdf readable on 6" kindle | 15:46.52 |
| | letters are too small as is everything | 15:47.05 |
| | too much strain on the eyes | 15:47.17 |
| kens | Well yes. | 15:47.19 |
| chrisl | But scaling it to a6 isn't going to help that..... | 15:47.56 |
| kens | If you want the text reflowed, you are out of luck. That's not something you can reasonably do starting with a PDF file. | 15:48.00 |
| aksr | i had to try | 15:48.14 |
| chrisl | If the device doesn't let you change the scaling for the best viewing, I would change the device! | 15:49.03 |
| kens | Some PDF files can be converted to another format (eg HTML) and reflowed there. Results are highly variable and never guaranteed to be correct. | 15:49.22 |
| | But this is not what I understand 'rescaling' to mean | 15:49.39 |
| aksr | i guess i missed the correct term | 15:49.55 |
| kens | So the correct answer to your original question is 'no, you can't reflow a PDF file onto smaller media' | 15:50.41 |
| aksr | a full sentence?! - well, thank you! :P | 15:51.08 |
| tor8 | aksr: actually, you probably don't want calibre's ebook-convert to convert CBZ to MOBI -- it trashes the image quality thoroughly :( | 15:51.14 |
| kens | That#'s quite contrary to the orignal purpose of PDF, which was intended to be a frozen layout that looks the same on all viewers | 15:51.22 |
| Robin_Watts | "no, you cannot reflow a PDF file onto smaller media without doing some really hairy heuristics that will break down lots of the time and leave you with a crap looking document" | 15:51.28 |
| aksr | Robin_Watts: no doubt about that | 15:51.48 |
| kens | Well I did say it would work sometimes | 15:51.56 |
| Robin_Watts | aksr: MuPDF will extract a PDF document to HTML for you. | 15:52.10 |
| aksr | how good is it? | 15:52.21 |
| Robin_Watts | That's an example of hairy processing that will break down lots of the time. | 15:52.32 |
| aksr | tor8: anyhow, i'm trying it atm | 15:52.46 |
| Robin_Watts | aksr: Depends on the input file. | 15:52.51 |
| aksr | i doesn't matter, it's of no use; i can't proofread files every time | 15:53.28 |
| | s/i/it | 15:53.32 |
| kens | Its in the nature of a PDF file not allow reflow, its supposed to look the same every time, so the layout is frozen as part of the document | 15:54.25 |
| henrys | marcosw_: hey great news on the testing stuff. | 15:54.47 |
| kens | So you will never be able to reliably reflow a PDF file, no matter what you do with it. | 15:54.49 |
| aksr | kens: i understand completely. | 15:54.55 |
| | i just can't understand why ebooks went with mobi/epub etc. | 15:55.19 |
| kens | Presumably because those formats *can* be reflowed ? | 15:55.34 |
| aksr | i know it's *easier* with those formats... | 15:55.43 |
| kens | So for text only documents they work better on variable size devices | 15:55.53 |
| aksr | kens: not enough, they also ruin good typography practise | 15:56.21 |
| Robin_Watts | aksr: Eh? Those documents *must* be flowed in order to be displayed. It's fundamental to them. | 15:56.27 |
| kens | O.O Where do we use WRT_wfloat in Ghostscript ? | 15:56.31 |
| Robin_Watts | Or at least that's true of epub 2. | 15:56.39 |
| | epub 3 is moving towards unflowed, for some unknown insane reason. | 15:56.53 |
| aksr | Robin_Watts: anyhow, some better format should have been invented for this... | 15:57.08 |
| | html is handicapped in general | 15:57.32 |
| | especially, the technical books... | 15:57.47 |
| kens | Ah, Tim meant WRF_wfloat, not WRT_wfloat..... | 15:58.32 |
| aksr | ..also readers of some optimal size should be enough for most books | 15:58.53 |
| henrys | marcosw: on your macs do you need to be a user in Sharing->Remote Login. I don't have you in there, maybe this was bug before and they fixed it. | 15:58.54 |
| aksr | but 6" is just too small | 15:58.59 |
| chrisl | aksr: I think it's pretty much an insoluble problem: such technical books rely on a close relationship between figures, tables and the main text body - but reflowing body text, by implication, loosens that relationship..... | 15:59.58 |
| aksr | loosens, but how much? and what direction | 16:00.39 |
| | *and in what | 16:00.45 |
| | it all could be _somehow_ solved | 16:00.57 |
| Robin_Watts | aksr: Go for it. | 16:01.12 |
| aksr | in short, it could be better than now | 16:01.13 |
| | Robin_Watts: i'll think about it | 16:01.32 |
| kens | Heading off out, goodnight all | 16:01.32 |
| aksr | kens: good night and thank you! | 16:01.45 |
| Robin_Watts | night kens | 16:01.46 |
| chrisl | I'm sure it could be *better* but I think there will always be corner cases that defeat any solution | 16:01.50 |
| aksr | chrisl: it doesn't matter; it's the same thing with pdfs | 16:02.12 |
| chrisl | No, not the same thing with PDF - PDF is specifically designed to retain the structure defined by the original author | 16:02.49 |
| | Retro-fitting text reflow to PDF is a *really* bad solution to the problem | 16:03.32 |
| aksr | retain the structure and retain the format | 16:03.33 |
| chrisl | PDF is supposed to look exactly the same when viewed on *any* (capable) device | 16:04.38 |
| aksr | are there a better solutions than pdf? | 16:05.44 |
| Robin_Watts | PDF is a "good" (not perfect, but good) solution to the problem of making files that will render as they are authored on any device. | 16:06.21 |
| | PDF is not a good solution to the problem of making files that will render readably on any device. | 16:06.55 |
| aksr | sure, but do you know or is there any better solution proposed or otherwise? | 16:07.03 |
| chrisl | Define "better"? | 16:07.34 |
| Robin_Watts | For which problem? | 16:07.38 |
| aksr | in general | 16:08.16 |
| Robin_Watts | gives up. | 16:08.34 |
| aksr | it doesn't matter. | 16:08.51 |
| chrisl | As we've already said, for accurate reproduction on a range of devices, PDF works pretty darned well | 16:08.52 |
| aksr | i understood that, and i'm aware of it. just asked are there any other solutions. that's it. | 16:09.18 |
| Robin_Watts | For fixed format reproduction across multiple devices? No, nothing comes close to PDF, IMHO. | 16:09.52 |
| pedro_mac | aksr: if you want a format which guarantees that you can view the doc laid out the way it was intended by the author then PDF is great. If you want to adjust content layout to fit various device sizes/orientations, HTML is a good option. Most other formats fit somewhere in between and their suitability depends on your use case and the type of content you want to view | 16:09.55 |
| aksr | pedro_mac: sure | 16:10.32 |
| | ..just, i don't think html is a good option.... | 16:10.55 |
| | ..even acceptable... | 16:11.19 |
| chrisl | Seems to work quite well for thehttp://www..... | 16:11.41 |
| Robin_Watts | html is a horrible bloated mess with far too much historical baggage. | 16:12.14 |
| | but used properly, with css, it is at last almost possible to get the same rendering on different engines for the same size output screens. | 16:12.51 |
| pedro_mac | it is, but its also specifically designed for reflowability. Depends what you want to present and how much effort you âre prepared to put in to make it work on different sized output devices | 16:13.17 |
| Robin_Watts | The problem is when you bring in the fact that most html pages don't obey the rules, and the hairy cruft that is javascript, it's amazing that it does as well as it does. | 16:13.45 |
| | pedro_mac: Absolutely. | 16:13.54 |
| pedro_mac | and thats the rub - getting a very good output on a wide variety of devices depends on good authoring | 16:14.24 |
| | and we have no guarantee of that | 16:14.38 |
| chrisl | pedro_mac: but that's true even with PDF, to some extent | 16:14.47 |
| pedro_mac | chrisl - absolutely | 16:16.42 |
| bashseb | Robin_Watts, tor8: changes were merged into zathura. Thanks again. http://bugs.pwmt.org/issue435 | 16:42.09 |
| Robin_Watts | bashseb: No problems. | 16:42.20 |
| | Are you a zathura dev then ? | 16:42.29 |
| bashseb | no, just fixed this thing | 16:42.38 |
| Robin_Watts | ok, cool. | 16:44.09 |
| aksr | Robin_Watts: agreed, html is horrible | 16:53.41 |
| rayjj | mvrhel_laptop: well, the change for SMask None is working, sort of. Now the clist writer is confused about the num_components (it's writing 1 component when it should have written 3) :-( | 18:06.36 |
| mvrhel_laptop | rayjj: oh there may have been some color space installed when the push occurred | 18:07.06 |
| rayjj | of course, this only happens on the complete, complex file -- not my simplified case | 18:07.20 |
| mvrhel_laptop | ugh | 18:07.25 |
| rayjj | I _should_ be able to catch it when writing the fill_trapezoid and num_components == 1 | 18:08.30 |
| | just venting (again) | 18:08.42 |
| | first, time for a lunch break | 18:09.00 |
| davesleep | using ghostscript do i need to create the file before i output to it? | 19:08.49 |
| | so the only way i can figur eout the error is by linking the pdf file | 19:56.37 |
| | one moment | 19:56.51 |
| | is there a thing i can read to see what the errors actually meaN/ | 19:57.38 |
| | wow.. my ghostscript is old.. i take it 8.70 is very old? mayb eupdating it could help? | 20:00.09 |
| Robin_Watts | yes. update to 9.14 or (better) the 9.15 release candidate and try that. | 20:00.45 |
| davesleep | im kind worried i may mess up everything else | 20:05.50 |
| | yup, it's definitely the fonts | 20:08.35 |
| | it seems | 20:08.38 |
| | even though it claims to replace them in the message.. it then throws an error afterwards.. however when i use a PDF that has normal fonts it's completely fine | 20:08.55 |
| | im using centOS 6.5 and im not sure how safe it'd be to install ghostscript 9.x | 20:36.35 |
| | considering i have imagick and read i'd need to reinstall imagick aswell | 20:36.46 |
| henrys | marcosw_: I added you to the remote login list under Sharing maybe this is required now. Reading the docs it should be required. | 22:34.10 |
| marcosw_ | henrys: yes, that worked. thx. | 22:36.50 |
| mvrhel_laptop | blah. finally found this bug I had in the text selection of gsview | 23:18.50 |
| | now I will get that committed the folder rename and the branch merged | 23:19.10 |
| | henrys: then we need to throw the beta of gsview on gsview.com | 23:19.37 |
| | and then its back to SOT for me | 23:19.43 |
| henrys | mvrhel_laptop: sounds good, I meant to ask about fred this morning at the meeting and forgot, have you heard from him? | 23:20.46 |
| mvrhel_laptop | henrys: yes we talked for 90 minutes last week. | 23:20.59 |
| | I was going to give him a call tomorrow to see how it is going | 23:21.07 |
| henrys | mvrhel_laptop: sorry I was hoping we could keep his work on the side. | 23:21.31 |
| mvrhel_laptop | I also need to add some page / print selection options on gsview. | 23:21.34 |
| | henrys: it was OK. I think he is up and running now | 23:21.50 |
| | it was needed | 23:21.56 |
| | there was actually more code in the gsview project than I realized | 23:22.17 |
| | I was productive last spring | 23:22.22 |
| | makes me feel bad about how little I have gotten done on SOT..... | 23:22.40 |
| | just a different type of coding | 23:23.07 |
| | ground up dev. vs. working with an existing beast | 23:23.19 |
| | I have been using gsview heavily as my default pdf viewer. the only thing I feel is inadequate is the printing | 23:23.59 |
| | again due to page selection primarily | 23:24.09 |
| | that is when I switch to AR | 23:24.16 |
| | also form filling would be good to add | 23:24.51 |
| | since all the logic for text selection is there, I dont think it would be too hard to do | 23:25.31 |
| | with mujs | 23:25.39 |
| | Forward 1 day (to 2014/09/10)>>> | |
IRC Logs