| <<<Back 1 day (to 2015/05/25) | 20150526 |
HD | Hi all good morning... | 09:46.55 |
| @paulgardiner: Any update for me? | 09:47.16 |
Robin_Watts | HD: You're a free, unsupported user, right? | 09:47.54 |
| I mean, you've not got a support contract or anything, right? | 09:48.07 |
HD | YEs | 09:58.19 |
Robin_Watts | HD: Right. So paulgardiner has other work at the moment. Work for paying customers that takes priority. | 10:00.31 |
| You've opened a bug - just watch that for updates. | 10:00.56 |
paulgardiner | Unless it is important enough to you that you might wish to fund the work. | 10:02.58 |
Robin_Watts | Indeed. If you were a commercial (or supported) customer, then the priorities change. | 10:03.42 |
kens | chrisl I suspect that problem from Kanazawa to be font related, the font is not embedded int he PDF file and the area where the error occurs (printed on the form) uses text. Regular GS lods msmincho.ttc as a replacement for MS-Mincho | 10:30.02 |
chrisl | kens: okay, thanks. I'll double check it - when I feel so inclined | 11:00.34 |
kens | Probably sooner than I feel like looking at Alex's report..... | 11:00.54 |
chrisl | You'd think he'd know enough to submit a viable bug report! | 11:01.42 |
kens | Its moe the fact that the example file is a 165Mb EPS file | 11:02.10 |
chrisl | Well, I'd consider that part of being a "viable" report.... | 11:03.38 |
kens | Lets say I'm not impressed | 11:03.53 |
sebras | tor8: are my gif patches up to speed now? | 11:45.45 |
tor8 | sebras: yes, LGTM but maybe we want Robin_Watts to take a look as well? | 11:48.19 |
| sebras: hm, you seem to mix rtol and ltor in the prototype and implementation. | 11:49.35 |
| sebras: maybe rename it reversebits or something similar? | 11:49.54 |
Robin_Watts | let me look now. | 11:53.06 |
| LZW_CLEAR used to be 256 | 11:55.49 |
| oh, I see. | 11:55.59 |
| yeah, tor is right. You should be consistent about rtol/ltor. | 11:57.58 |
| other than that one tiny ltor->rtol change, lgtm. | 12:00.37 |
sebras | tor8: Robin_Watts: ah, yes, reverse_bits is probably better. | 12:02.18 |
| or maybe we are not underscope people in this project? | 12:02.33 |
tor8 | sebras: mistell? | 12:02.59 |
sebras | tor8: underscore. | 12:04.16 |
tor8 | we are underscore people these days; forget about our barbarian runtogetherwords past ;) | 12:04.48 |
sebras | tor8: ah, ok. reverse_bits it is. | 12:05.13 |
bigsky | hi all | 13:32.58 |
tor8 | hi | 13:42.27 |
ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 13:42.27 |
henrys | meeting in 10 minutes | 14:19.46 |
marcosw | morning all | 14:28.42 |
kens | Morning Marcosw | 14:28.49 |
henrys | hi marcosw | 14:29.07 |
| not much to say this morning for me, well a couple of things: | 14:29.40 |
| we'll skip next Tuesday's meetings | 14:29.51 |
Robin_Watts | This is Tor8's last meeting for a while. | 14:31.01 |
rayjj | we'll miss you tor8 | 14:31.15 |
henrys | mvrhel, chrisl: should we try and collect web stats for gsview for a bit and see how it is doing? | 14:31.37 |
| tor8: you will be missed there is still time to make a reservation if you change your mind ;-) | 14:32.02 |
fredross-perry | morning | 14:32.10 |
henrys | hi fredross-perry | 14:32.22 |
chrisl | henrys: we'll probably have to involve marcosw if we want to do that - I wouldn't know where to start...... | 14:32.28 |
henrys | fredross-perry: we traditionally skip the Tuesday meeting before the in person meeting so no meeting next week. | 14:32.45 |
Robin_Watts | chrisl: Where are the gsview binaries stored? | 14:32.46 |
chrisl | Robin_Watts: on casper at the moment | 14:32.58 |
fredross-perry | all righty thanks | 14:33.06 |
Robin_Watts | chrisl: eek. Amazon is not the cheapest place for download hosting. | 14:33.26 |
chrisl | Robin_Watts: I wasn't expecting high volume for the early "releases" | 14:34.07 |
Robin_Watts | I could enable webaliser or AWstats for downloads.ghostscript.com | 14:34.23 |
| and we could make a downloads.ghostscript.com/gsview directory ? | 14:34.36 |
mvrhel | henrys: yes I was curious about that | 14:34.42 |
henrys | since this is the last meeting before the in person meeting please review the agenda items with your names next to them since you will be asked about those items at the meeting. | 14:35.19 |
chrisl | I think we have stats for the godaddy site, but I have no idea whether it's just the whole site, or something more detailed | 14:35.21 |
mvrhel | The write-up in the news letter that Scott sent out was pretty brief about gsview. I don't think it even said what it was | 14:35.45 |
henrys | Robin_Watts, fredross-perry what is the status of the certificates? | 14:35.48 |
| jogux said there was an issue with apple and qt? | 14:36.01 |
chrisl | Robin_Watts: TBH, I'd think mupdf would be a considerably higher bandwidth load than gsview (for now) | 14:36.24 |
Robin_Watts | henrys: Non-apple ones all working. They are in my casper home dir (Certificates) | 14:36.24 |
fredross-perry | I got Apple certs installed and rebuilt gsview, seems to work. | 14:36.54 |
mvrhel | the installer for gsview is using the certificate | 14:37.01 |
rayjj | mvrhel: the news letter is mostly for customers and potential customers on his contact list, and there may not be much interest in gsview among customers | 14:37.10 |
marcosw | speaking of the in person meeting, it's been unseasonably cool and cloudy in the Bay Area for the last few weeks, so pack a sweater (having said that it will probably end up being warm and sunny next week). | 14:37.23 |
mvrhel | true. | 14:37.31 |
rayjj | If we want wider notification, that's probably something our VP of Marketing should tackle | 14:37.55 |
| has Ted started yet ? | 14:38.14 |
henrys | mvrhel: I'll put that on the agenda if you want | 14:38.23 |
| rayjj: yes | 14:38.28 |
Robin_Watts | mvrhel: When you say the installer for gsview is using the cert, you don't mean that you've built the signing into the build process, do you? | 14:38.48 |
mvrhel | no | 14:38.55 |
jogux | henrys: I tried Fred's latest gsview and certificate is working | 14:38.58 |
mvrhel | I mean you did it | 14:38.59 |
| Robin_Watts: | 14:39.02 |
Robin_Watts | right. I manually signed it. | 14:39.12 |
jogux | fredross-perry: btw, I added a problem I did see with gsview on OS X to bugzilla, hopefully it emailed you or something? | 14:39.18 |
henrys | rayjj: he's going to be working SO for the most part. | 14:39.30 |
fredross-perry | yes, I saw it thanks. | 14:39.35 |
henrys | rayjj: at least for now | 14:39.40 |
Robin_Watts | the download stats for godaddy aren't broken down by area currently. | 14:40.10 |
| just "50Gig a day" style graphs. | 14:40.24 |
henrys | about the meeting we should freshen the high priority projects for miles' benefit I don't use the list much myself but he likes to know what everyones primary focus is, so I'll put something there for all of you sometime today and if you want it to say something different let me know ;-) | 14:41.18 |
kens | I'm (hopefully) nearing completion of the FirstPage and LastPage work. The code currently works, and to test the method I have two types of subclass, a First/Last page and a object filtering (text, image, linework) class | 14:41.46 |
| To test the interaction I've been loading both and using the cluster as well as local testing. TO date this all works well, but I have no doubt there will be a significant bug tail. | 14:41.46 |
| Additionally the PCL interpreter now uses this code for the 'monochrome palette' instead of replacing the colour mapping procs in the device. | 14:41.46 |
| I'm using the same code for the StudioRip work, but this turns out to be insufficient (The interpreter does 'stuff' that we simply cannot intercept). So I've had to make additional changes in the interpreters for ths, sadly this is less neat than I had originally hoped. | 14:41.46 |
| I plan to document what I've done (and why) and the StudioRip work, and then send an emaikl to tech, it would be great if everyone could take a look at this before the staff meeting, its a lot easier to deal with face-to-face and I do expect there will be some comments :-) | 14:41.46 |
marcosw | wow, kens has learned to type quickly. | 14:42.11 |
kens | Miranda doesn't send anything till you press 'return' :-) | 14:42.33 |
mvrhel | is it OK to have the customer name up there? | 14:42.39 |
rayjj | probably so. | 14:43.16 |
henrys | mvrhel: let's not worry about that customer | 14:43.20 |
mvrhel | :) | 14:43.24 |
rayjj | we used to mention them all the time when they were using ghostscript before deciding to become a customer | 14:43.45 |
kens | I probably should have just said Zoltan :-( | 14:45.11 |
rayjj | chatzilla has a mode that lets you prepare several lines in advance and then send them in a flood as well | 14:45.17 |
| kens: don't sweat it | 14:45.27 |
henrys | I think IRC should be opened up we don't have these secrets anymore i.e. http://www.artifex.com/page/printer-products.html | 14:45.34 |
| anyway | 14:45.49 |
| how do the certificates work into the regular release of gsview. Is there something special that has to be done each time, is it in the wiki so someone else can do it? | 14:46.29 |
rayjj | henrys: there may be some customers that haven't given Miles permission, and it's hard to keep track of which they are, so it's easier to just avoid customer names unless we are sure | 14:46.40 |
fredross-perry | regarding gsview/mac, Iâve got a âhow to build itâ doc whic Iâll update with the cert instructions. | 14:47.42 |
Robin_Watts | henrys: When we have an exe, we run a command on it. | 14:47.48 |
rayjj | but I did mention the customers on the printer-products web site on IRC last week (intentionally) | 14:47.50 |
henrys | rayjj: I'm good with that too, if that's everyone's preference, ain't mine though ;-) | 14:47.50 |
Robin_Watts | I've got a README file in ~robin/Certificates. | 14:48.05 |
mvrhel | Robin_Watts: so should we be signing the exe and the dlls? | 14:48.38 |
| as well | 14:48.46 |
| as the installer? | 14:48.50 |
Robin_Watts | mvrhel: Arguably. | 14:48.52 |
fredross-perry | I would think so | 14:48.58 |
mvrhel | ok | 14:49.01 |
henrys | mvrhel, Robin_Watts well a gsview release procedure for the twiki seems like the right thing to do, yes? | 14:49.28 |
Robin_Watts | henrys: Sure, but there is information we can't put on the twiki (like passwords etc) | 14:49.47 |
| but yes, a release procedure there would make sense, with a pointer to where you can get the certs and passwords. | 14:50.11 |
henrys | Robin_Watts: sure | 14:50.21 |
mvrhel | sounds like a good idea | 14:50.22 |
Robin_Watts | Should I make a GSView web on the twiki? | 14:50.30 |
| Or do we roll it into either gs or mupdf ? | 14:50.53 |
kens | Can one of you point me at the certificate please ? I need to have a go at signing the .inf file. | 14:51.00 |
Robin_Watts | kens: ~robin/Certificates | 14:51.09 |
kens | Thanks, I'llgo fetch it now | 14:51.19 |
henrys | mvrhel, Robin_Watts I'd prefer something separate in the twiki? | 14:51.20 |
Robin_Watts | Sure. Will add that now. | 14:51.28 |
mvrhel | separate yes | 14:51.48 |
henrys | fredross-perry: can you do a release procedure for apple and email it to robin for inclusion in the wiki? | 14:52.07 |
fredross-perry | surel | 14:52.16 |
| y | 14:52.19 |
Robin_Watts | or, if you give me a few minutes you can edit the twiki yourself. | 14:53.20 |
henrys | I'll put some of this stuff on the agenda just so we don't let anything slip. Hopefully we'll just check it as done at the meeting. | 14:53.40 |
| tor8: before you go what do you think about bumping up the priority of XFA - we have a couple customer requests and it seems to be supported by other pdf products? | 14:54.52 |
jogux | robin: it might be worth standardising on using git-private (or some other similar shared location) for certificates etc? | 14:55.07 |
Robin_Watts | jogux: yeah. | 14:55.28 |
henrys | I didn't have anything else this time. | 14:56.20 |
mvrhel | ok. I will be back in a bit | 14:56.35 |
henrys | off to skype ... | 14:56.58 |
Robin_Watts | mvrhel, fredross-perry: OK, if you go here: http://twiki.ghostscript.com/do/view/GSView/WebHome | 15:00.11 |
| You'll see (on the left hand side) 2 links, one for each of you. Click on it, and edit to your hearts content. | 15:00.35 |
mvrhel | ok. thanks Robin_Watts | 15:01.21 |
fredross-perry | please add a third, for OSX. Thanks. | 15:01.38 |
rayjj | and does kens need a place to document the steps to sign the .inf ? | 15:02.18 |
kens | Not until I find out how to do it..... | 15:02.32 |
chrisl | That probably goes in with Ghostscript | 15:02.43 |
kens | I suspect ts somethign we'll only do once anyway | 15:02.50 |
| And yeah, it goes with Ghostscript only | 15:02.58 |
chrisl | Well, maybe more than once, but not with every release | 15:03.09 |
kens | I'd guess the next time will be 3 years when the certificate expires | 15:03.32 |
| Its not like the .inf file changes | 15:03.41 |
chrisl | I changed the comments not that long ago - and we'll have to change them again once we can sign it | 15:04.09 |
kens | Really ? Why ? | 15:04.24 |
chrisl | At the moment it mentions all the hacky stuff to install on Win8 | 15:04.56 |
kens | The .inf file mentions that ? | 15:05.22 |
chrisl | In the comments | 15:05.30 |
kens | Oh, never noticed | 15:05.37 |
rayjj | chrisl: we'll have to change the comments, then sign the changed one, right ? | 15:07.24 |
chrisl | rayjj: yes | 15:07.34 |
tor8 | henrys: I don't know ... our forms support is still a bit incomplete as it stands | 15:09.50 |
| but I don't think supporting XFA should be all that onerous, we already have an XML parser and mapping between that and the PDF objects shouldn't be too much work | 15:10.51 |
| but I fear that a lot of our current forms work is all very much pdf-object specific | 15:11.06 |
Robin_Watts | fredross-perry: Done. | 15:11.43 |
| tor8: I know almost nothing about xfa. Is XFA of use purely for printing ? | 15:12.17 |
| i.e. is there a usecase for it in just being able to display xfa files without the editing? | 15:12.51 |
fredross-perry | thanks | 15:14.16 |
tor8 | Robin_Watts: I have forgotten or repressed my memory of most of the details of XFA | 15:15.31 |
| that said, I think it's a way to store forms data outside of the PDF file in an XML format | 15:15.43 |
kens | chrisl Hin-Tak sticking his nose in again | 15:16.13 |
chrisl | With a comment as useful as usual :-( | 15:17.23 |
kens | Yes, see my reply :-) | 15:17.41 |
tor8 | henrys: actually, I take that back. I just downloaded the XFA spec ... 1600 pages | 15:17.52 |
kens | XFA is a PDF file where some or all of teh content is actually XML | 15:18.19 |
| Increasinlgly we are seeing these, and the content is almost totally XML, the only part which is PDF is the bit that says 'if you can see ths your viewer does not support...' | 15:18.51 |
tor8 | with scripting... | 15:18.55 |
| this does not look like something in the scope of mupdf | 15:19.15 |
kens | There's a lot to it, one of the reasons I keep saying we don't want to go near it with GS | 15:19.22 |
| If you avoid the form filling and just display the form that might be enough for some people, it seems to be what they expect from GS | 15:19.52 |
| Robin_Watts : where did you get hold of the Windows Driver kit ? | 15:21.52 |
Robin_Watts | kens: the what now ? | 15:23.07 |
kens | Hmm, maybe you don;t need it for signing executables | 15:23.26 |
Robin_Watts | signtool was on my machine. It's probably in one of the VS versions. | 15:23.43 |
kens | I need inf2cat as well | 15:23.53 |
rayjj | my impression of XFA is that it looks pretty horrendous | 15:23.53 |
kens | But I don't seem to have signtool either | 15:24.02 |
| Actually I do have signtool, seems to be part of SV 2012 | 15:24.44 |
Robin_Watts | signcode, sorry. | 15:26.26 |
kens | Well, I have signtool :-) | 15:26.38 |
| but not signcode | 15:26.54 |
Robin_Watts | yeah, I have signtool too in VS2005 (and probably others) | 15:27.21 |
kens | I have it al the way back to v 6.0a | 15:27.39 |
rayjj | kens: it says: Download Windows Driver Kit (WDK) 8.1 Update to get the tools to build, test, debug, and deploy drivers for Windows Windows 8.1 Update, Windows 8.1, Windows 8, and Windows 7. on this site: https://msdn.microsoft.com/en-us/windows/hardware/gg454513#drivers | 15:27.41 |
kens | rayjj I was hoping Robin_Watts already had it :-) | 15:27.57 |
rayjj | kens: and the download link is: http://go.microsoft.com/fwlink/p/?linkid=393659 | 15:28.08 |
kens | rayjj it says I need to install VS 2013 before installing WDK 8.1, I htnk that lets me out, I don't have VS 2013# | 15:28.58 |
rayjj | kens: I think all you need is VS 2013 Express. | 15:30.13 |
kens | It stoo late in the day to download both, I'll look again tomorrow | 15:30.34 |
rayjj | mvrhel: has the full VS 2013, iirc. | 15:31.10 |
kens | Yeh, I was thinking it might be easier for him to do this, as he has the full release | 15:31.27 |
mvrhel | what exactly do you want me to do? | 15:31.43 |
kens | Sign a .inf file so that we don't get the UAC stuff | 15:31.58 |
rayjj | mvrhel: figure out how to sign the inf | 15:32.02 |
kens | https://technet.microsoft.com/en-us/library/dd919238%28v=ws.10%29.aspx#bkmk_sub4b | 15:32.17 |
mvrhel | so there is something more to it then what Robin_Watts did with the gsview installer? | 15:32.26 |
Robin_Watts | mvrhel: Signing code (or dll) is trivial. | 15:32.40 |
kens | Its different | 15:32.40 |
Robin_Watts | I haven't looked at signing a .inf though. | 15:32.50 |
kens | You need to convert the .inf to a .cat using inf2cat | 15:32.52 |
| Then use signtool to sign the .cat | 15:33.12 |
Robin_Watts | I will try to download that stuff later and obtain the relavent bits for kens to play with :) | 15:33.14 |
kens | I'd do it myself but I'm off out for dinner shortly | 15:33.32 |
| I won;t have time to download and install all the bits until tomorrow anyway | 15:33.47 |
mvrhel | I am running on fumes today. We drove in from Bend Oregon at 1:30am last night | 15:33.57 |
| So I am not sure I can figure this one out | 15:34.05 |
| right now | 15:34.09 |
kens | Well, me neither, but it *looks* simple | 15:34.18 |
| If you have inf2cat and signtool | 15:34.26 |
| Of which I have one.... | 15:34.32 |
| but not the other.... | 15:34.38 |
mvrhel | I see | 15:34.42 |
kens | So I'd have to download the WDK at least, anad maybe VS 2013 as well, just to get inf2cat :-(( | 15:35.11 |
rayjj | can't mvrhel just send you the inf2cat.exe ? | 15:49.28 |
kens | Its not in VS 2013, its in the WDK | 15:49.45 |
| But the WDK says that VS 2013 is a pre-requisite for isntalling the WDK | 15:49.59 |
| SINCE i HAVE NETIEHR, i'D NEED TO DOWNLOAD AND INSTALL BOTH | 15:50.18 |
| Oops caps sorry | 15:50.24 |
Robin_Watts | kens: It's also in the WindowsSDKDir\bin\ | 15:50.30 |
kens | Is that under one of the VS installs ? | 15:50.49 |
rayjj | Robin_Watts: where's that | 15:50.53 |
henrys | tor8, kens : so you see no hope of just implementing the parts that customers would likely encounter. Jeong claims to have done that in his product and we can have that code. | 15:50.57 |
Robin_Watts | possibly. | 15:51.06 |
kens | henrys, it would require adding an XML parser to GS, seems like a load of effort for something which *stil* wouldn't allow you to actually fill the form in | 15:51.36 |
rayjj | henrys: that's code to do XFA ? | 15:51.37 |
henrys | rayjj: yes | 15:51.43 |
Robin_Watts | henrys: I think the answer (if that's about xfa) is that we'd need to look into it. I don't think any of us know enough to comment at the moment. | 15:51.51 |
| kens: We're talking mupdf, not gs. | 15:52.09 |
kens | I thought henrys included me for a GS opinion. No point talking to me about MuPDF -) | 15:52.36 |
| Robin_Watts : where did you find WindowsSDK ? I don't seem to have that anywhere | 15:53.00 |
chrisl | Thing is, if we're only talking about printing, then we might as well include GS...... | 15:53.02 |
henrys | kens: yes it would be in mupdf but that doesn'l leave you out. You have useful things to say about XFA itself. | 15:53.13 |
kens | henrys, I don't really speak XML..... | 15:53.27 |
Robin_Watts | kens: I believe "Windows Platform SDK" is a separate download that can be had for earlier versions. | 15:53.33 |
kens | Robin_Watts : Ah, well I can try downloading that tomorrow instead | 15:53.46 |
chrisl | IIRC, we're had more queries about XFA in GS than mupdf so far | 15:53.49 |
Robin_Watts | but leave it with me, and I'll try and get inf2cat from somewhere. | 15:53.52 |
henrys | we do have an XML parser for GS XPS, can't think that's an issue with this problem. | 15:53.59 |
kens | My worry would be that the older verison wouldn't sign Windows 8.1 drivers | 15:54.09 |
| henrys, we don't include XPS parsing in Ghostscript | 15:54.33 |
| Only in GhostPDL | 15:54.40 |
| Since our PDF is in Ghostscript we'd need to move the XML parser into Ghostscript itself | 15:54.56 |
| I'm not saying that can't be done mind you | 15:55.05 |
rayjj | kens: right, but at least we have the makefile to build expat :-) | 15:55.08 |
chrisl | Adding expat to GS would be trivial...... | 15:55.08 |
henrys | kens: it will be avaliable in the language switch build ;-) the point is the parser can be added that's not the big problem. | 15:55.20 |
kens | is less than sure about that :-) | 15:55.35 |
| Finger on nose | 15:55.42 |
rayjj | chrisl just volunteered, didn't he ? | 15:55.57 |
rayjj | finger on nose, too | 15:56.10 |
kens | THe worst case would be a hybrid XFA file where portions of the content are in PDF and portions in XML | 15:56.10 |
chrisl | I can add it to the build in about ten minutes - hooking it into the PDF interpreter is *way* more work | 15:56.29 |
henrys | actually paulgardiner is currently tasked with XFA for MuPDF and I am waiting to hear back from Jeong. We'll talk more about it after we have some code to look at. | 15:56.43 |
rayjj | I didn't think it hooked into the PDF interpreter, I thought it just draws stuff on top of the PDF content | 15:57.31 |
kens | On top of/as well as/ instead of..... | 15:58.01 |
rayjj | but then I haven't dug into it. The spec was rather daunting for casual reading | 15:58.04 |
chrisl | rayjj: we'd need some kind of hook into the PDF world so the interpreter could kick it off | 15:58.14 |
kens | https://technet.microsoft.com/en-us/library/dd919238%28v=ws.10%29.aspx#bkmk_sub4bThe PDF itnerpreter would just hand the XML to ExPat I thnk | 15:58.40 |
| Grr stupid paste.... | 15:58.48 |
| The PDF interpreter would have to dig out the XNL from the file, and pass it to ExPat for processing | 15:59.07 |
| I don't recall how the mixed mode works, when we have PDF content as well as XML content | 15:59.26 |
rayjj | right, just a 'runXML' operator that gets passed a stream | 15:59.32 |
chrisl | As long as that is all the interaction there is, then yes | 15:59.40 |
kens | rayjj yes for simple cases (and may files are *really* simple that way) | 15:59.48 |
chrisl | And we'd also have to work out how pdfwrite would deal with this stuff | 16:00.11 |
kens | The majority of XFA files we see have nothign useful inthe PDF, just a boilerplate 'your viewer isn't Adobe' message. All the content is in the XML. But a mixed mode is possible | 16:00.30 |
rayjj | yuck! | 16:00.30 |
kens | Well, pdfwrite already handles XPS.... | 16:00.50 |
chrisl | Initially, at least, I guess pdfwrite would just "flatten" it to PDF | 16:00.50 |
kens | chrisl I would never plan to do anything else | 16:01.04 |
chrisl | I was thinking whether we'd preserve.... | 16:01.09 |
rayjj | but as long as the XFA paints stuff onto the page using the graphics library, pdfwrite would just get the "flattened" content | 16:01.12 |
kens | No way would I tr5y to turn it into a PDF Form | 16:01.14 |
rayjj | agrees wholeheartedly to that | 16:01.32 |
| I don't think Adobe even does tht | 16:01.41 |
chrisl | I wondered if we could just do an XML passthrough | 16:01.45 |
kens | Sure, define it as a spec op | 16:02.11 |
| Turning an XFA file into an XFA file doesn't seem useful to me, but what do I know ? :-) | 16:02.58 |
chrisl | Whereas turning a PDF into a PDF..... | 16:03.14 |
kens | Yeah, I keep telling people not to do that, but they don't listen :-( | 16:03.30 |
rayjj | does PDF/A allow XFA content ? | 16:03.33 |
kens | No | 16:03.38 |
rayjj | OK, 'cause that's often a reason people do pdfwrite of a PDF (PDF->PDF/A) | 16:04.13 |
| so pdfwrite going to PDF/A would flatten the XFA, but other PDF flavors would do the passthrough ? | 16:05.21 |
kens | I'd still prefer to just turn it into a straight PDF | 16:05.45 |
| Anything else would be an enhancement, so I could ignore it :-) | 16:06.03 |
chrisl | Personally, I'd be tempted to just always flatten it - I just think it's a decision we would want to make and document clearly..... | 16:06.33 |
rayjj | kens: the whole project is an enhancement :-) | 16:09.46 |
kens | True. | 16:10.03 |
| I have to go get ready to go out. Goodnight all | 16:10.40 |
mvrhel | This stuff that Junghen is sending me is wacky | 16:45.05 |
| he hacks the mupdfnet code making it impossible for the display list to be freed and wants to know why there is a leak | 16:48.38 |
| By the time I down load and unzip the 500Meg project he hands me and figure out what is wrong my morning is 1/2 gone | 16:49.47 |
| on to something usefull | 16:50.05 |
Robin_Watts | mvrhel, ray, kens: Windows users may want to look at ConsoleZ. | 18:57.46 |
| It's an updated version of Console2 that I've raved about before. | 18:57.55 |
| https://github.com/cbucher/console/wiki/Downloads | 18:58.26 |
mvrhel | thanks | 18:58.50 |
Robin_Watts | In particular it gets the "high dpi" stuff right for windows 8.1 which stops text appearing blurry. | 18:59.15 |
| (if you're not using console 2 already, then you're missing out on having lots of command prompts/git bash/cygwins etc running in multiple tabs within a single window, and sane copy/paste etc) | 19:00.04 |
henrys | rayjj: I thought you may know the history to this one: there seems to be a read from the tty in gs when it is run in the background. So if you run with -dNOPROMPT using unix job control "&" the job will stay in the background until you press a key which "stops" the job. Is there some postscript reason for this read of the tty at the end of a program? | 20:05.28 |
rayjj | henrys: it's probably reading for the >>showpage, press <return> to continue<< (which isn't printed because of the -dNOPROMPT). To not have it pause after each page, you also need -dNOPAUSE | 20:07.51 |
henrys | rayjj: right but then it wouldn't stay in the background ... think of a display device you want to leave on, I guess it isn't a big deal | 20:09.04 |
rayjj | henrys: So if you want to leave the display "on", it has to read in order to know when to go to the next page | 20:10.12 |
rayjj | doesn't quite understand what the question is, I guess | 20:10.45 |
| gs doesn't know or care if it is running in the backgrgound | 20:11.37 |
| but when it's running in the background, it still gets stdin from somewhere | 20:12.36 |
sebras | tor8: Robin_Watts: so hopefully the patch series at sebras/master is ok now..? | 20:12.38 |
henrys | rayjj: I am wondering if there should be -dNOREAD option. The particular application is to run Xvfb then gs in the background, then run another program after that grabs the framebuffer while gs is running in the background in the same script. Haven't thought much of multiple pages yet. | 20:16.57 |
| not a big deal ... just thinking about how xvfb can be incorporated into the cluster testing | 20:17.59 |
rayjj | henrys: you need to synchronize the capture of the page to when the 'showpage' is finished (>>showpage, press <return> to continue<< prompt). Otherwise how will you know when the page is completely rendered ? | 20:18.50 |
| then you will need to send the '\n' to gs to let it go on to the next page. | 20:19.25 |
henrys | rayjj: I was thinking of some sort of scheme with FirstPage and LastPage, and not worry about postscript | 20:20.17 |
rayjj | the 'app' that is invoking gs must read the stdout and look for the >>showpage, press <return> to continue<< string, then capture the frame buffer (or analyze it or whatever), _then_ send the '\n' to go on to the next page | 20:21.05 |
| if you want the app to be a shell (or perl) script, then using named pipes for stidn and stdout is possible | 20:22.15 |
| but I don't see any other way to know when to capture the frame buffer | 20:22.56 |
| otherwise you will get a partially rendered (or empty, if transparency is used) frame buffer | 20:25.00 |
henrys | rayjj: yeah I thought a "sleep" but that is kinda bogus. | 20:26.09 |
rayjj | henrys: yeah -- open loop. Sometimes it'll sleep too long (for simple pages) and sometimes not long enough | 20:27.28 |
mvrhel | henrys: are you there? | 20:50.44 |
henrys | mvrhel: back now yes | 21:03.28 |
mvrhel | ignore the email I just sent you | 21:03.39 |
| I was mistaken on the issue that I thought was going on | 21:03.50 |
| Digging a bit further | 21:04.02 |
| bbiab | 21:07.04 |
| henrys: are you there still? | 23:26.09 |
henrys | I am | 23:26.51 |
mvrhel | now I managed to burn up the afternoon. turns out this guy had made another change in the mupdfnet code where he did a modification which was dropping a displaylist | 23:26.53 |
| the ref counts end up all screwed up in this case | 23:27.10 |
| I am not sure what to say to him really | 23:27.23 |
| I don't quite understand what he is trying to accomplish | 23:27.49 |
henrys | tell him you can't debug his code and only send you problems he can reproduce with our code in a pristine state. | 23:29.31 |
mvrhel | ok that is a good idea. doing a one line change in our code and asking me why it does not work is not a good use of time | 23:30.04 |
| especially when I don't know where the change was | 23:30.17 |
| thanks henrys | 23:30.31 |
henrys | mvrhel: sure | 23:30.38 |
| Forward 1 day (to 2015/05/27)>>> | |