| <<<Back 1 day (to 2014/01/15) | 2014/01/16 |
tor8 | Robin_Watts: argh. I've forgotten what csi->obj is for :( | 00:14.39 |
Robin_Watts | When we parse an object, it goes into csi->obj | 00:15.04 |
| i.e. it's "the current object" AIUI. | 00:15.14 |
tor8 | henrys: perl has a mix of dynamic and static scoping, the "local" and "my" keywords | 00:16.46 |
Robin_Watts | Well, if perl has it, it must be a good idea. | 00:18.15 |
tor8 | Robin_Watts: right. so that should be for the marked content like stuff and inline image dictionaries? | 00:19.09 |
| Robin_Watts: indeed! just ask marcos for some more perl evangelism ;) | 00:19.33 |
Robin_Watts | In this case it was the array before a TJ | 00:19.33 |
tor8 | Robin_Watts: those shouldn't be... the csi->in_text special cases the arrays before a TJ | 00:19.55 |
| it might have been a naked TJ outside BT/ET | 00:20.12 |
Robin_Watts | tor8: no, this was inside a BT/ET. | 00:20.27 |
tor8 | Robin_Watts: that shouldn't happen... | 00:20.35 |
henrys | tor8:good point about perl I had forgotten about that. | 00:21.04 |
tor8 | pdf_run_stream, if it sees a '[' inside a BT/ET it sets "in_array" and directly calls pdf_show_string/space without creating an array object | 00:21.20 |
| Robin_Watts: sorry about the delay. I was coaching sebras in how to write a compiler/interpreter. | 00:22.01 |
| our javascript interpreter can now run code, with proper static scoping and closures and all. | 00:22.20 |
Robin_Watts | pdf_show_text is called with pdf_obj *text, and it runs through calling pdf_array_get then pdf_show_string on each char in turn. | 00:22.24 |
| pdf_show_text is always passed csi->obj as text. | 00:22.43 |
| hence if pdf_show_string destroys csi->obj, we have a problem. | 00:23.00 |
| tor8: you and sebras are writing a javascript interpreter? | 00:23.42 |
marcosw1 | ray_laptop: for the logs. I just replied to cust 351/850 re: bug 694750, you hadn't closed the bug so it didn't show up in my daily "customers bugs that have been closed but not verified and therefore need me to do something bug search. | 00:23.55 |
tor8 | Robin_Watts: yup. he didn't like quad wheel and wanted to know how to write an interpreter so I agreed to help him out. | 00:24.15 |
Robin_Watts | cool. | 00:24.30 |
tor8 | I figured, if the original JS interpreter was written in a week, how hard can it be? | 00:24.31 |
Robin_Watts | I think the language has moved on a bit since then :( | 00:24.51 |
tor8 | anyway, it parses the full syntax, compiles most of it to a bytecode (just a few features missing) and now runs the bytecode with recursion, variables, scoping and objects | 00:25.02 |
Robin_Watts | Can it cope with eval ? | 00:25.16 |
tor8 | Robin_Watts: a fair bit, but not a whole lot. it's still a big warty simplified mess. | 00:25.19 |
Robin_Watts | I guess if it's an interpreter, then yes. | 00:25.35 |
tor8 | Robin_Watts: it's designed to, hence the stupid variable scoping handling. | 00:25.43 |
| every single variable access hits a dictionary :( | 00:25.55 |
| but that's what you have to do to comply with the spec and handle eval and "with", without writing a lot of complicated code | 00:26.14 |
| still, on fibonacci it's only 5 times slower than spidermonkey :) | 00:26.30 |
Robin_Watts | tor8: yeah. I wrote an actionscript interpreter (so javascript with the parsing side of the problem removed) | 00:26.36 |
tor8 | https://github.com/ccxvii/libjs/ | 00:27.01 |
| I expect it'll still be a while until we can use it in mupdf, but it might get there eventually | 00:28.46 |
Robin_Watts | Well, I have the FBS updated so that it should be able to make us an android exe. | 00:30.29 |
tor8 | Robin_Watts: I'm still trying to figure out when we end up calling TJ with an array | 00:30.38 |
Robin_Watts | It now gets all the way to the final link and then fallsover. | 00:30.41 |
tor8 | Robin_Watts: might it be a (xxx) TJ ? | 00:31.17 |
| string then TJ, not array then TJ | 00:31.25 |
Robin_Watts | hold on, let me check. | 00:31.46 |
| Nope. | 00:32.14 |
| Let me upload you my minal test file. | 00:32.22 |
tor8 | Robin_Watts: please do. I agree with the patch regardless, but I want to understand it completely. | 00:32.38 |
Robin_Watts | b8e_2.pdf in my home dir. | 00:32.40 |
| Bedtime for me. Talk to you tomorrow... | 00:32.58 |
tor8 | Robin_Watts: that file is missing the BT | 00:33.48 |
| there is an ET but no BT | 00:33.52 |
Robin_Watts | Mail from miles - crap! | 00:33.59 |
tor8 | Robin_Watts: oh bollocks! :( | 00:34.31 |
Robin_Watts | indeed. | 00:34.43 |
tor8 | Robin_Watts: does this mean you're going to abandon the JNI project? | 00:35.30 |
| or go ahead full steam and show them hahaa, should've chosen differently!? | 00:35.48 |
| Robin_Watts: still no word on the travel arrangements? | 00:36.13 |
henrys | tor8, Robin_Watts: might have been for the best. | 00:36.18 |
tor8 | Robin_Watts: patch for 694894 LGTM | 00:37.00 |
Robin_Watts | tor8: Thanks. | 00:37.27 |
| tor8: No, no word on travel arrangements. | 00:37.36 |
| henrys: if you speak to Miles can you ask him about our proposed flight back? And whether I should go ahead and sort a car rental ? | 00:37.58 |
| henrys: Well, it means we won't have to worry about BSDing it. | 00:38.29 |
henrys | Robin_Watts: yes I'll talk to Miles but I think he left for the day | 00:38.54 |
Robin_Watts | tor8: Dunno about JNI. I guess we can take a few days to figure out whether to keep pushing on that. | 00:39.06 |
tor8 | Robin_Watts: I think it should still be worthwhile, to get rid of all the people misusing the MuPDFCore stuff | 00:39.27 |
| should hopefully shut up the incessant whin(g)ing on stackoverflow | 00:39.48 |
Robin_Watts | I suspect it's probably still worthwhile, as it forces us to tweak the core into a more nicely bindable format. | 00:40.03 |
| And the JNI bindings should be useful for desktop java too. | 00:40.35 |
| ho hum. really is bedtime. | 00:41.05 |
tor8 | Robin_Watts: good night! | 00:41.17 |
| I should probably start thinking about bed too... | 00:41.27 |
sebras | tor8: Robin_Watts: me too... | 00:56.07 |
kens | Does anyone know where to find the source code for gsprint ? | 09:06.49 |
| I can't find it on the net anywhere..... | 09:07.01 |
chrisl | I thought it was included with the gsview sources? | 09:09.09 |
kens | Ah, maybe that's where I'm going wrong, I was looking for it separately | 09:09.29 |
| OK I found the source for GSView 5 at least, lets see.... | 09:10.05 |
| well its got a gsprint.htm file at least :-) | 09:10.27 |
chrisl | Yes, IIRC, the sources aren't well differentiated. I wonder if I still have it there..... | 09:11.00 |
kens | that's an understatment | 09:11.08 |
| THi looks kind of horrible | 09:11.18 |
| It also looks like it uses some ancient version of MS C++ | 09:11.37 |
chrisl | Well, I was able it build it (quite some time ago) so it must be possible to build with at least VS2005 | 09:12.25 |
kens | Looks like it uses makefiles not projects | 09:12.44 |
chrisl | I'd guess you want to start looking in srcwin/win.mak | 09:14.14 |
kens | There are instructions in readme.htm | 09:14.29 |
| Have to copy win.mak and edit it | 09:14.37 |
chrisl | Well, that makes things a little easier..... | 09:14.54 |
kens | No way to figure out which files go with gsprint as far as I can see | 09:14.57 |
chrisl | Just need to work through the targets | 09:15.27 |
kens | :-( | 09:15.32 |
| If I have to work on this its getting a solution and the Unix/OS2 targets die | 09:15.57 |
| Of course it would help pif Russell said what DEVBASE was for | 09:17.07 |
| OK looks like its the location of VS | 09:17.35 |
chrisl | I doubt there is a gsprint for Unix, and gsview for Unix hasn't actually even built for years | 09:17.38 |
kens | And OS/2 is dead, so ..... :-) | 09:17.51 |
| Well there is no gsprint target in the makefile (no reference to it at all actually) | 09:18.39 |
| Oh goody gsprint is a .cpp file | 09:19.52 |
| and it looks like gvc.mak builds gsprint | 09:20.06 |
| Hmm, maybe not, it just copies the files | 09:20.23 |
chrisl | wincom.mak | 09:20.34 |
kens | ah yes, I see it now | 09:20.49 |
| I'm beginning to think this customer might actually be using redmon to print from a windows client to a Windows non-PostScript printer. There must be some method to this madness, but right now I can't see what it is..... | 09:33.42 |
| chrisl : Robin_Watts I've just sent a proposed mail to both of you, its for the customer with the gsprint problem. | 09:42.05 |
| Would you mind having a quick read please, particularly to make sure I@mbeing polite, but also for an inaccuuracies. | 09:42.31 |
| Good grief my typing is bad today.... | 09:42.46 |
| back in a moment | 09:43.03 |
chrisl | kens: "but its not at all clear to me how you are using" -> "but its not at all clear to me how you are using it" | 09:43.59 |
kens | chrisl, oops I thought I had caught all those, thanks | 09:59.24 |
chrisl | kens: easy to miss. The only thing I would say is that I'd almost suggest putting more emphasis on getting a clear description of their workflow and their problem. BUT maybe that's not such a good idea given the delay in responding. | 10:00.53 |
kens | And given that I've asked before. Last time when I asked they simply disappeared..... | 10:01.16 |
chrisl | gsprint doesn't use mswinprt2? Really? | 10:01.26 |
kens | No it really doesn't | 10:01.33 |
| It has GS write a BMP file then opens the pritner itself | 10:01.44 |
chrisl | Crikey that seems.... crazy...... | 10:01.52 |
kens | and draws teh BMP file on the printer DC | 10:01.55 |
| Russell claims this is faster, I cannot for the life of me see why it would be | 10:02.14 |
chrisl | Maybe on Windows 3.11? | 10:02.33 |
kens | ROLF | 10:02.39 |
| The only way to tell would be to try it I guess | 10:02.48 |
| I'm not sure I care that much for a legacy product we plan to dtich and replace | 10:03.05 |
chrisl | No, I'd agree. I am surprised there's no way to pass printer settings via the API, though, that seems short sited even for M$ | 10:03.43 |
kens | I'm not sure that is the case actually, but I didn't want to say anything to teh customer on that score | 10:04.06 |
| Also it may very well be dependent on the vcersion of Windows, I have no idea what versio nRussell used to develop this stuff | 10:04.44 |
| Just as well Norbvert doesn't try building GS on warning level 4 :-) | 10:05.03 |
chrisl | Yeh, good point! | 10:05.38 |
| Mind you, how the hell that font library builds with that is beyond me! | 10:06.44 |
kens | Beats me too. | 10:06.57 |
| BTW I'm assuming you'll reply to Sudheesh ? | 10:06.58 |
| I admit I have no idea what he is babbling about | 10:07.09 |
| Its amazing how often people think a Unicode encoding has something to do with fonts | 10:07.30 |
chrisl | No, I'm going to leave that to henrys, I can explain how to include extra fonts, but I have *no* idea whether PCL actually supports any of that | 10:07.48 |
kens | It seems unlikely to me that you can put Unicode in a PCL file. I would assume you would have to encode the font in some manner, which probably means downloading a custom subset, thereby negating the question about making the font available to GhostPCL | 10:08.51 |
chrisl | The problem is, there may be PCL extensions out there (like the PCL 5J ones) for this stuff. Henry is more likely to be aware of that, and answer appropriately | 10:09.47 |
kens | Certainly the case yes | 10:10.09 |
chrisl | Crap like that is one of the big reasons I dislike PCL, and why flogging to big, important customers always worries me..... what exactly do *they* mean by PCL? | 10:11.12 |
kens | Yes, which particular flavour form which division of HP or which clone ....... | 10:11.39 |
| OK I'm going to send that mail now. | 10:13.07 |
| I'd really like to get a picture from them of what they are doing, and at least as importantly, why...... | 10:13.47 |
chrisl | I think I'm going to have to rewrite how U*ST integrates with our file i/o - since they changed to unbuffered i/o it's becoming a pain :-( | 10:14.20 |
kens | Oh great :_( | 10:14.34 |
chrisl | They now (obviously) pass files around as integer file descriptors, where we need pointers..... | 10:15.17 |
kens | sigh..... | 10:15.31 |
| It never improves much as a library does it ? | 10:15.54 |
chrisl | No, but this seems like a highly retrograde step. It's not like they added a bunch of custom buffering in to compensate, so..... | 10:16.44 |
kens | Interesting article on El Reg: | 10:17.44 |
| http://www.theregister.co.uk/2014/01/16/microsoft_buries_the_sinofsky_era_and_jumps_on_the_coffin_lid/ | 10:17.44 |
| "Even today the RT application store is a sad and desolate place if you blind-tested it on a user alongside BlackBerrys store for its defunct PlayBook, you wouldnt be able to tell much difference." Ooh cutting.... | 10:17.59 |
chrisl | "'Windows 9' will wipe the slate clean" - what, again? I thought that's what Win8 was all about! | 10:18.34 |
kens | Seems like maybe they really mean it this time, there havebeen some rumours for a few months that 8.1 wasn't enough (which it isn't) and MS might actuallu U-turn again | 10:19.06 |
chrisl | I can't believe they didn't do enough pre-release "user experience" testing on 8 to get the idea of what the general reaction would be | 10:20.21 |
kens | I'm sure they did, and then decided to carry on anyway in the best traditions of Margaret Thatcher | 10:20.43 |
Robin_Watts | Morning kens | 10:24.17 |
kens | Hi Robin_Watts | 10:24.21 |
Robin_Watts | Just looking at that email now. | 10:24.30 |
kens | Too late, I sent it ;-) But thanks | 10:24.42 |
Robin_Watts | From reading the logs so far, I reckon you *should* be asking for a clear statement... oh. | 10:24.52 |
kens | Robin_Watts : I am trying to get a clear picture of their workflow and usage but as I said to chris above, I tried that 2 years ago and they went silent | 10:25.26 |
Robin_Watts | kens: I understand. Nevertheless, I would consider it to be the primary thing to pursue. | 10:26.00 |
kens | All I feel I cna do is gently prod them and now I'm involved again I'll keep poking them | 10:26.02 |
| Agreed, until I understand what they are doing, (and what they expect) I'm kind of fumbling in the dark | 10:26.33 |
chrisl | I'd also be concerned about being too forthright, given the circumstances - we don't want to give the impression we're blaming them for this delay | 10:26.37 |
Robin_Watts | And it's a stick to beat them with. If they come back in 2 years time and say "we still haven't had a solution", we can justifiably say "look, we've asked to to give us a clear description, and you've never done that. How can we be expected to solve for a use case you won't describe?" | 10:26.59 |
kens | Yeah, which was why I didn't mention that the reason they didn't get a reply for 2 weeks was becuase they stopped CC'ing support and mailed Marcos direct.... | 10:27.06 |
| Robin_Watts : I did allude to my previous document, which contains a series of questions they never answered.... | 10:27.33 |
Robin_Watts | People never read back. | 10:28.04 |
kens | Trying to walk the line of saying 'you need to tell us more' wihtout saying 'you're a bunch of idiots'..... | 10:28.13 |
chrisl | "Bzzzed - I'm sorry, you have failed the exam to continue as a Ghostscript customer, please exit quietly....." | 10:28.26 |
kens | I wonder where these people are based | 10:28.48 |
| Hmm Galveston texas | 10:29.13 |
chrisl | Get Scott to "buzz" them in his plane...... | 10:29.40 |
kens | If this drags on I could offer to drop in at the next meeting and actually *see* what thye do | 10:29.53 |
| back in a moment, electrician here | 10:30.07 |
| Actually it seems like its not texas, its minnesota | 10:38.40 |
| Maybe, time for google maps | 10:38.53 |
| Aha its Georgia | 10:39.40 |
| So not really amenable to a visit | 10:40.33 |
| Its kind of ner Atlanta | 10:40.48 |
Robin_Watts | Morning paulgardiner | 11:05.03 |
paulgardiner | Hi | 11:05.52 |
Robin_Watts | I have an exe! | 11:23.43 |
paulgardiner | Oh amazing | 11:23.52 |
Robin_Watts | I had to run the final link by hand (remove -lpthread from the arguments). | 11:23.54 |
| It's 37Meg. | 11:23.58 |
| but that's probably unstripped. | 11:24.03 |
kens | O.O | 11:24.05 |
| Even so.... | 11:24.10 |
Robin_Watts | and I have no idea if it actually works :) | 11:24.20 |
kens | chrisl I've added a new 'PDFUseCIEColor' control that determines whether the CIE transformation takes place or not. This is partly because of the detection in pdfwrite, and partly because some of the rendering looks wrong with GS when thed CIE remapping is done. | 11:29.45 |
| And, of course, files which contain a Default* colour space are comparatively rare. | 11:30.09 |
| Hmm, actually I think I should also set the control to true if UseCIEColor is true in the device dictionary | 11:31.12 |
| Ah, no, if its true then the transform takes place, so no worries there. | 11:31.42 |
| I wonder if I need this control atg all | 11:31.51 |
chrisl | Er, so why do we need this at all?? | 11:34.40 |
kens | I'm wondering exactly that :-) | 11:34.53 |
| Actually I htink I do | 11:35.20 |
| Its because if UseCIEColor is true then *all* colours go through the transform, even if the page has no Default* colour space | 11:35.43 |
| If UseCIEColor is false, then only pages that have the DefautlTransform (and an appropriate one for the given colour space) do it | 11:36.15 |
chrisl | I'd have assumed we'd always want the latter | 11:37.26 |
kens | chrisl, as it stands, we can't control the trasnfrom | 11:38.01 |
| *If* the page has a Default* colour space then we *always* perform the transformation. There is no other control | 11:38.20 |
chrisl | Oh, so you want a control for *just* that? | 11:38.48 |
kens | well given that at least one PDF file renders incorrectly at the moment, yes maybe. | 11:39.06 |
chrisl | Doesn't UseCIEColor potentially break all overprinting? | 11:39.46 |
kens | yes and no.... | 11:40.03 |
| well mostly yes | 11:40.28 |
| I think I need to consider this some more | 11:41.36 |
chrisl | Well, in theory, we've performed a color transform, so shouldn't apply overprinting. In practice, we've made often quite considerable changes to the color values, so overprinting is unlikely to have the desired results - I'd have thought. | 11:41.47 |
kens | There is a test file which has overprinting and a Default* space, acrobat shows it without overprints, even when 'simulate overprint' is true | 11:42.39 |
chrisl | For once, I would say that's actually in-line with the spec! | 11:43.08 |
kens | Essentially this is just ot avoid the problem of pdfwrite emitting a warning whenever the input is a PDF file | 11:43.22 |
| Yeah, it is correct according to the spec | 11:43.33 |
| The problem is that the PDF interpreter unconditionally sets UseCIEColor to true | 11:44.12 |
chrisl | Yes, which was something of a shock to me | 11:44.35 |
kens | It then sets up a default DefaultGray, DefaultRGB and DefaultCMYK which are *actually* the Device* spaces | 11:44.50 |
| when we see a colour space we substitute the Devioce* space with the relevant Default* space (if USECIEColor is true) | 11:45.35 |
| Now there's an optimisation in setcolorspace which jumps out if the new space is the same as the current space. | 11:45.56 |
| So in the default condition, we don't actually do a transformation to CIE even though UseCIEColor is true (can you say ugly hack ?) | 11:46.24 |
chrisl | So, we set UseCIEColor true, then add a hack so it doesn't actually apply to device colors? | 11:46.38 |
kens | Yes, exactly. | 11:46.49 |
chrisl | That's, erm, "sub-good"...... | 11:47.15 |
kens | Then if the page has a Default* space we use that instead. Because its not the device space, teh action of settgin the colour space actually performs the transformation. | 11:47.27 |
| So the problem is that I've added code to pdfwrite to detect UseCIEColor and say 'don't do that, m'k ?' | 11:48.04 |
| But it can't tell that we've hacked the colour transform machinery so that it isn't actually doing anything..... | 11:48.27 |
| I'm also concerned that (when using pdfwrite) people may actually want to preserve the original colour spce *even if* the PDF file has a Default* space. | 11:49.14 |
chrisl | Yeh, I see. That makes sense | 11:50.04 |
kens | So I'm trying to think of a way to do that, and the only solution I can come up with is a separate control for PDF files that turns on the CIE transform | 11:50.50 |
| Of course that will trigger a warning from pdfwrite, but presumably that wqill be OK since the user took action to turn it on, we can assume they know what they are doing. | 11:51.31 |
chrisl | I can't help but feel that the pdf interpreter should handle the presence Default*** spaces in the input better than this :-( | 11:52.30 |
kens | I'm inclined to agree but I'm not sure how that should be (could be) done. | 11:52.50 |
| Actually I wonder if redefining cs and CS might be the way to go | 11:53.31 |
| Oh and 'k' and 'g' and whatever rgb is | 11:53.53 |
chrisl | It would probably mean quite a lot of work, *and* the house of cards that is the pdf interpreter..... | 11:54.13 |
kens | I'll look at that this afternoon. Getting away from UseCIEColor has to be better than this | 11:54.13 |
| chrisl yes that's what worries me | 11:54.26 |
Robin_Watts | OK, I now have an exe without needing to do the final link myself, and without having (knowingly) broken any of the more standard builds. | 14:12.23 |
| And when I strip it it's 2.3Meg. That seems more plausible. | 14:14.40 |
kens | better anyway | 14:14.47 |
| must be lots of symbols there | 14:14.56 |
| Hi norbertj | 14:18.02 |
norbertj | kens: Hi , I just want to mention that we too have switched to warning level :( | 14:18.36 |
| kens; warning level 4 , VS2012 | 14:18.48 |
kens | we couldn't build GS with that | 14:19.25 |
norbertj | to circumvent some problems: I added some -wdxxxx flags (e.g. -wd4047) | 14:19.48 |
kens | I htink we would need lots of those | 14:20.10 |
| enough tot make it pointless I suspect | 14:20.19 |
norbertj | I can give you my list? | 14:20.21 |
Robin_Watts | So you've upped the warning level in general, but kept specific ones down at the level they were before. | 14:20.42 |
| norbertj: Sure. Might be worth putting them on an enhacement bug ? | 14:21.03 |
| That way we can track the progress and they won't get lost. | 14:21.13 |
kens | chrisl is the build maintainer | 14:21.17 |
chrisl | I'm not getting saddled with fixing warnings..... | 14:21.35 |
Robin_Watts | I'll have a look at some of them in my idle moments. | 14:21.53 |
kens | I found that fixing warnings on VC tended to introduce them in gcc anbd vice versa | 14:22.38 |
norbertj | on pcl6_msvc.mak UFST_CFLAGS=-DGCCx86 -wd4047 -wd4022 (to disarm a warning about: SL32 <-> void * | 14:25.25 |
| in cgif.c | 14:25.36 |
chrisl | norbertj: I'm just looking at the euro vs currency glyph thing - it looks to me like the difference is because fonts genuinely have different glyphs at that index. | 14:28.29 |
norbertj | chrisl: you mean the UFST FCO has a different glyph then the urwfonts? | 14:30.54 |
chrisl | norbertj: yes, exactly. I get exactly the same output from the old PCL/UFST interface code, with UFST 5.3 | 14:31.44 |
norbertj | chrisl: so probably I/we should contact Monotype for this? | 14:32.12 |
Robin_Watts | sorry, the warnings thing - is that for UFST or for gs ? | 14:32.59 |
norbertj | Robin_Watts: b.t.w. I mailed the ccf32.tr (with a list of -wd####) to support. | 14:33.19 |
Robin_Watts | I have just seen it. | 14:33.26 |
chrisl | norbertj: Well, the MT fonts do have a Euro glyph. I'm cutting down the test now to a more usable size | 14:33.43 |
| norbertj: when I have a smaller test job, maybe the next thing would be to get henrys to send the PCL to his HP printer, and see what it produces? | 14:35.04 |
norbertj | chrisl: perhaps the codepage for sweden is not correct? However, the pcl compatibility guide shows that Sweden has a currency symbol (at ascii 36), Which matches also the output of my LJ4345 | 14:36.41 |
| chrisl: I ment the ISO Substitution Table on B-34 Internal Symbol Set Charts. | 14:37.25 |
henrys | chrisl: let me low if you want me to look at it. | 14:37.34 |
| hi norbertj | 14:37.39 |
norbertj | hi henrys | 14:38.04 |
chrisl | norbertj: sorry, do you mean that the UFST output matches your LJ, or the URW output matches it? | 14:39.01 |
norbertj | chrisl: I double checked; UFST matches LJ4345 (with currency symbol), URW has a euro-symbol, so would be not according to PCL-compatibility guide. | 14:40.22 |
chrisl | norbertj: so the problem is really URW's | 14:41.00 |
norbertj | chrisl: yes | 14:41.23 |
chrisl | Okay, henrys and I will need to work out how best to communicate with URW on that - they're going to be sick of hearing from us! | 14:41.58 |
henrys | norbertj: are you switching to urw? | 14:43.24 |
| chrisl: can you make a bug for it? | 14:47.18 |
norbertj | henrys: no, I detected this when building ghostpdl with/without UFST, and compared then. | 14:47.19 |
chrisl | henrys: once I've narrowed things down a bit, yes, I will. | 14:47.43 |
| norbertj: was my reply about your custom font list okay, or do you want me to look at that, too? | 14:48.20 |
henrys | chrisl: I can do that if you want, but if your having fun don't let me atop you ;-) | 14:48.30 |
| s/atop/stop | 14:48.38 |
chrisl | henrys: I'm thrilled to the core..... | 14:49.03 |
norbertj | chrisl: I think I understand it. | 14:49.33 |
chrisl | norbertj: okay that's good. But like I said, just get back to me if you do need me to look into it | 14:50.10 |
norbertj | chrisl: b.t.w. what do you think of the storing of download resources on disk (softfonts, macros, symbolsets) and reloading on startup (i.e. for us very relevant in a printer environment). | 14:50.55 |
chrisl | norbertj: that's not something I'm qualified to give an opinion on, tbh. Maybe henrys is better placed? | 14:51.51 |
henrys | norbertj: there are fill operators in PJL for that. | 14:53.16 |
| s/fill/file | 14:53.47 |
norbertj | henrys: yes, but that is to support loading files to a directory (and that is what we have disabled for security reasons). | 14:54.52 |
henrys | norbertj: it couldn't be password protected? | 14:56.55 |
norbertj | henrys: the functionality I'm thinking of is to just send pcl(5) files with download objects (font, symbolset or macro), make them permanent, we trigger on this permanent-flag, and store the data on disk. On startup the data is readin again, as if it was downloaded at that moment. We still have customers who use this workflow. | 14:57.07 |
| henrys: could be, but company policy is to not open up the printer too much :( | 14:58.08 |
henrys | norbertj: it's been a long time since I looked at that stuff I'll review it. I don't understand how this new scheme would be more secure than using PJL filesystem commands. | 15:01.48 |
norbertj | henrys: with the PJL filesystem commands, you can upload any binary file. And if the destinationpath is not correctly handled, you could upload malware/virusses this way (a bit paranoid but ...). When downloading a softfont, it goes through the pcl checking of being a correct download font , and if not it is discarded. | 15:05.31 |
henrys | norbertj: true good point | 15:06.02 |
norbertj | henrys: I was just wondering if my mods for this could be in ghostpdl-mainstream (inside a define perhaps). But it is not that much work for me to re-apply the mods with each new ghostpdl-version. | 15:07.14 |
| henrys: perhaps other parts could be more relevant. I.e. I have some extensions on the handling of plexity/pagesideselect which was missing in pcl5/xl | 15:08.36 |
henrys | norbertj: a bug and a patch would be easiest for me. | 15:08.42 |
| norbertj: but however you want to communicate the stuff is fine. Do you mind putting it on bugzilla? | 15:09.26 |
norbertj | henrys: I'll collect my mods and provide a patch to you. | 15:09.29 |
| henrys: will do | 15:09.40 |
chrisl | <sigh> the URW fonts aren't even consistent :-( | 15:16.02 |
kens | chrisl chrisl dumb question time | 15:20.32 |
| How did you manage to get the 'UseCIEColor' warning out of pdfwrite ? I can't seem to reproduce that | 15:20.52 |
chrisl | I just ran a PDF file. | 15:21.15 |
kens | The reason is that the pdfwrite device is opened before we process the PDF file, and so UseCIEColor is not set, so I can't trigger the warning | 15:21.33 |
paulgardiner | henrys: I'm working on the offics doc stuff now, but earlier I was trying to finish off a couple of jobs both of which went badly. | 15:21.51 |
kens | UseCIEColor doeasn't get set by the PDF interpreter until we try to execute a page | 15:21.53 |
| chrisl so I don't get a warning | 15:22.05 |
chrisl | kens: maybe a PDF with a non-Letter/A4 page size, so it calls setpagedevice? | 15:22.28 |
henrys | jobs? of the iOS business? | 15:22.32 |
paulgardiner | iOS is one of them. | 15:22.46 |
chrisl | kens: I'm guessing, but it might have been Bug689005.pdf or Bug691425.pdf that threw up the warning | 15:23.33 |
kens | chrisl I have a PDF file with a very small page amd it still doesn't trigger for me. In fact that page also (even in my latest code) *does* set UseCIEColor and it *still* doesn't trigger a warning because the device is opened while USECIEColor is still false, before the page is executed | 15:23.33 |
paulgardiner | henrys: Our build script for the library is broken and the release build wont build. I guess that one can wait. | 15:23.51 |
kens | I can't, at the moment, find a way to trigger that warning in the currently checked in code | 15:24.04 |
henrys | paulgardiner: sounds like tor8 should have a look | 15:24.15 |
| paulgardiner: at least I thought he handled that. | 15:24.57 |
chrisl | kens: I really wish my brain had save/restore - it would make task swapping much easier :-( | 15:25.01 |
kens | chrisl sorry, but I'm having genuine difficulty getting that warning to appear at all. | 15:25.21 |
chrisl | kens: this causes it for me, in the current code: ./gs -sDEVICE=pdfwrite -o stuff%03d.pdf ../../../svn-private/ghostpcl/trunk/tests_private/comparefiles/Bug691425.pdf | 15:25.58 |
paulgardiner | henrys: possibly. I think it's just that we are using the wrong sys variable to pick up where to put the .a files. I'm sure I can fix it, but I thought I'd better prioritise the office stuff | 15:26.06 |
kens | OK I'll grab that file, thanks chrisl | 15:26.25 |
kens | wonders if -o has some unexpected side effect | 15:26.41 |
henrys | paulgardiner: yes doing the office stuff is a bit more important. | 15:26.52 |
chrisl | Eek, that would be bad :-( | 15:26.57 |
kens | But not too surprising... | 15:27.13 |
| If I can reproduce it I think I have a decent way to fix it | 15:27.47 |
paulgardiner | henrys: the other job was to reply to Raed Abdallah's enquiry about signatures. | 15:27.49 |
henrys | paulgardiner: oh I didn't even see that one. I think I've been desensitized to him I'll have to look. | 15:29.09 |
kens | chrisl aha! I see you are using %d I bet that's it | 15:29.51 |
| because we close the device and reopen it for every page | 15:30.12 |
paulgardiner | henrys: his example file is a little different to the ones I have looked at before. With a quick hack, I was able to test our signature validation on the contained siganture, but it failed. I get errors reported from openssl. I tranced one to being to do with the placement of a self-signed signature in the chain, but handling that just ran it into another error to do with RSA. | 15:30.17 |
chrisl | kens: crap, yes, that is it - it's only a one page file, so it didn't register that I'd done that | 15:30.53 |
kens | I'm just trying it now, at least it explains why I never saw the problem before | 15:31.19 |
| Yes, that does it | 15:32.43 |
| For every page we close and reopen the device, and it looks like we do it before we've reset the page device dict, so the 'reopen' sees that UseCIEColor has been set | 15:33.28 |
| Lets see if my new code works round that | 15:33.52 |
paulgardiner | henrys: In any case, there are things he wants to do that we don't yet handle, but I was hoping to be able to say they were fairly small additions. For this problem with the verification, however, I can't predict how deep we'll need to delve for a solution. Worse case, it could be something openssl just doesn't have support for (although I'd doubt that). | 15:34.32 |
norbertj | goinh homr, bye | 15:47.48 |
| going home, bye | 15:47.58 |
chrisl | bye norbertj | 15:48.01 |
kens | bye bye nor | 15:48.06 |
| too slow... | 15:48.14 |
Robin_Watts | oh, the irony :) | 15:49.51 |
| Still no word back from Miles about the flights. | 16:18.25 |
chrisl | henrys: I've opened that bug with the details and the list of fonts. I've assigned it to you, CC'ed me. | 16:21.02 |
| I seem to be spending an inordinate percentage of my life just recently writing down lists of font name and/or font files :-( | 16:21.36 |
henrys | thanks chrisl I'll have a look | 16:22.47 |
chrisl | henrys: I have C706.bin cut down to just the one glyph if you find you need/want that | 16:23.39 |
mvrhel_laptop | Robin_Watts: | 16:24.11 |
Robin_Watts | paulgardiner: If it was me, I'd reply to Raed and say that you've had a look at his file. Say it's not quite in the same format as the ones that we support already, but that you've had a quick attempt. That's resulted in it running into a problem in OpenSSL, so you've stopped looking at it now as you have a high priority task. Tell him that you'll hope to get back to it shortly, and that... | 16:24.21 |
| ...you'll keep him in the loop. Does that seem reasonable? henrys? | 16:24.22 |
mvrhel_laptop | so Miles had replied to me about my flight which left at 7:50 | 16:24.30 |
Robin_Watts | mvrhel_laptop: | 16:24.30 |
mvrhel_laptop | pm | 16:24.31 |
henrys | chrisl: sure attach it to the bug - if it is really isolated I don't worry about QL copyright. | 16:24.37 |
Robin_Watts | mvrhel_laptop: And he thought that was OK? | 16:24.43 |
kens | mvrhel_laptop : what did he say ? | 16:24.54 |
mvrhel_laptop | He said he thought anything after 7:00pm was fine | 16:25.04 |
| he just did not want people looking at watches at lunch time | 16:25.18 |
kens | we shoudl be good then | 16:25.20 |
mvrhel_laptop | I suspect so | 16:25.34 |
| I already bought my ticket | 16:25.39 |
Robin_Watts | ok, so should we go ahead and book then? kens, chrisl, paulgardiner? | 16:25.46 |
mvrhel_laptop | henrys: I am going to be out this morning | 16:26.00 |
kens | IMO yes, go with teh American flights | 16:26.00 |
mvrhel_laptop | have to help out at school field trip | 16:26.16 |
Robin_Watts | 79 on Thursday, 78 on sunday. | 16:26.21 |
kens | Robin_Watts : yesw | 16:26.25 |
chrisl | I'll do it as soon I get get done on the phone to tech support for my mobile :-( | 16:26.33 |
henrys | Robin_Watts, paulgardiner I'd say have marcosw1 create an enhancement and tell Raed updates will be posted. | 16:27.26 |
paulgardiner | Am I right that he's a gs customer looking into also using MuPDF? | 16:29.23 |
henrys | kens:is that latest pcl customer the old pswrite customer or lore | 16:30.32 |
| s/or/of | 16:30.56 |
kens | henrys which one ? | 16:31.03 |
henrys | Sudheesh? | 16:31.15 |
kens | AH, I'm not at all sure | 16:31.23 |
| the engineers there seem to change quite frequently | 16:32.00 |
tor8 | Robin_Watts: booking through AA I can see a real BA flight leaving 10:45, arriving 14:40 as well | 16:32.00 |
henrys | mvrhel_laptop: okay | 16:32.02 |
kens | henrys this is the frist email I see from him | 16:32.10 |
Robin_Watts | tor8: Yes, but getting to heathrow for 10:45 is unpleasant compared to getting there for 13:xx | 16:32.38 |
kens | henrys might be worth checking with them and amke sure its the right group | 16:32.46 |
tor8 | Robin_Watts: tell me about it. no matter what I do it takes me at least 4 hours to get there, and that's cutting it close ;) | 16:33.15 |
| Robin_Watts: the 13:xx flight is a 4-hour layover for me | 16:33.38 |
| which should be just about enough to get from T3 to T5 | 16:33.53 |
| T5 to T3 | 16:33.58 |
Robin_Watts | Allow for an hour through immigration, 30 mins to transit, an hour for security. | 16:34.33 |
| And that doesn't seem so bad. | 16:34.43 |
| of course, because you have time, each of those will take 5 minutes. | 16:34.53 |
kens | yeah by the time you get tehre we'll be there too | 16:34.59 |
tor8 | so we're set on AA79 on march 6 and AA78 on march 9? | 16:35.25 |
kens | tor8 I believe this is the case | 16:35.40 |
tor8 | in that case I'll just go ahead and buy the tickets now | 16:35.42 |
kens | out of LHR 13:20 | 16:35.46 |
| back at 21:20 from DFW | 16:35.53 |
| I'm just booking mine now | 16:36.05 |
tor8 | kens: it says out of LHR 13:25 here | 16:36.16 |
kens | could be 13:25 | 16:36.24 |
| yes 13:25 pm from LHR | 16:37.38 |
| AA 79 | 16:37.42 |
tor8 | and on sunday, AA78 departure 21:20 | 16:38.10 |
kens | Yes, that's what I'm doing now | 16:38.27 |
| £608 for me | 16:38.41 |
Robin_Watts | Plus 4.50 processing fee ? | 16:39.39 |
kens | hasn't quite got there yet | 16:39.47 |
| just filling in the personal details | 16:39.56 |
Robin_Watts | System Error. This feature is currently unavailable. | 16:40.36 |
kens | O.O What feature was that ? | 16:40.48 |
Robin_Watts | payment :( | 16:40.57 |
tor8 | kens: currency conversion has my fare at GBP 608 too! | 16:41.00 |
kens | Oh dear..... | 16:41.02 |
paulgardiner | So we are definitly going with these flights? I'm at the payment stage too | 16:41.13 |
| Fee is only for credit cards | 16:42.21 |
kens | Hmm I see the extra legroom is expensive | 16:42.51 |
chrisl | I didn't get a processing fee - odd..... | 16:42.58 |
kens | and no extra legroom on the way out | 16:43.15 |
tor8 | kens: where do you see the legroom options? | 16:43.24 |
| after you pay? | 16:43.29 |
kens | orange seats tor | 16:43.32 |
| when I select my seat before paying | 16:43.43 |
tor8 | kens: I haven't seen a seat selection option anywhere... | 16:44.07 |
kens | Oh, it comes up here | 16:44.15 |
tor8 | kens: which site are you booking on? | 16:44.23 |
kens | But there are no extra legroom seats on the way out | 16:44.25 |
| american | 16:44.29 |
| american.co.uk | 16:44.38 |
tor8 | odd. probably because mine is a stupid connection. | 16:44.41 |
kens | Robin_Watts : says the return flight has bad reviews | 16:45.07 |
| So I'll buy the legroom on that one | 16:45.15 |
tor8 | http://americanairlines.eu.amadeus.com is where I'm redirected to | 16:45.37 |
kens | hehe noscript blocks that one for me | 16:46.00 |
| and then redirects to aa.com | 16:46.10 |
henrys | chrisl: your job prints neither currency or Euro on the HP sigh. | 16:47.10 |
kens | norbert indicated it was hte 'swedish' compatibility didn't he ? Maybe US printers are different ? | 16:47.39 |
henrys | chrisl: I can't even find the unicode symbol it is a circle with 4 lines radiating out like a boat steering wheel | 16:47.49 |
chrisl | henrys: you could try the whole QL job, instead | 16:47.59 |
| henrys: how old is your HP? | 16:48.54 |
henrys | chrisl: 5 years | 16:49.22 |
paulgardiner | Do we have an address for the hotel? | 16:49.41 |
chrisl | Should have the relevant glyphs, then...... | 16:49.43 |
kens | paulgardiner : you mean we aren't all camping in Scott's field ? | 16:50.00 |
tor8 | kens: okay, I might be able to pick up legroom seats now. which chair are you in? | 16:50.12 |
kens | err not sure, maybe 16J | 16:50.24 |
paulgardiner | kens: an address for the field will do. | 16:50.25 |
kens | Scott's address is on the contact PDF :-) | 16:50.50 |
chrisl | I put Scott's address - I figured that would work as a "clearing house" | 16:50.58 |
kens | yeah £4.50 for credit card | 16:51.36 |
chrisl | I wonder why I wasn't charged for the credit card | 16:52.02 |
tor8 | kens: yikes, $136 for extra legroom is pretty steep | 16:52.11 |
kens | tor8 yes that's what I said :-( | 16:52.21 |
| gouging I call it | 16:52.30 |
tor8 | yeah. but considering they even charge for "preferred" seats, I think we're pretty much hosed if we don't splurge :( | 16:53.20 |
kens | Yeah, and since Robin's seatguru said the seats were crampled..... | 16:54.15 |
Robin_Watts | finally booked. | 16:55.05 |
kens | is still working on it | 16:55.17 |
| filling in passport & stuff | 16:55.25 |
| Well I *think* I've finished.... | 16:57.52 |
| Robin_Watts : did your ticket go from 'pending' to ticketed ? Mine is still stuck at pending and there's not obviously anything else I need to do | 16:59.52 |
Robin_Watts | My printout says "Ticket Pending" | 17:00.20 |
kens | Oh OK thenits not just me I feel happier | 17:00.30 |
Robin_Watts | I've tried for 33C out and 36D back. | 17:02.26 |
kens | 38G out 16J back | 17:02.35 |
chrisl | 37B out and 35D back | 17:03.18 |
| Haven't had a confirmation e-mail though, I hate that...... | 17:04.00 |
kens | Me neither, definitely don't like the lack of response | 17:04.20 |
chrisl | I vaguely remember it was the same last time - it did come eventually, I think | 17:04.41 |
tor8 | kens: it took a couple of minutes for mine to arrive as well | 17:04.44 |
kens | Been more thn a few minutes now..... | 17:05.01 |
chrisl | I mean like next day.... | 17:05.25 |
tor8 | got booking confirmation at 17.47, then e-ticket confirmation at 17.53 | 17:05.36 |
Robin_Watts | I had terrible trouble booking mine. It didn't like my card. I had to phone them, but changing the card worked. | 17:05.36 |
kens | It 'seemed' OK with my card..... | 17:06.01 |
tor8 | kens: have you got the 6-letter code? | 17:06.04 |
kens | But still pending | 17:06.05 |
| tor8 I have a record locato | 17:06.14 |
tor8 | kens: that's what I used to then log in to aa.com and pick my seats without being redirected | 17:06.27 |
Robin_Watts | kens: I have two :) | 17:06.27 |
kens | tor8 I picked my seats before I had the record locator :-) | 17:06.43 |
tor8 | kens: yeah, it wouldn't let me do that since it had me book through amadeus because of the extra leg on BA | 17:07.04 |
kens | OK well I guess you are in luck and have a reservation, the rest of us presumably will get one 'eventually' | 17:07.30 |
tor8 | kens: check your spam filter. the confirmation mail I got for the seat changes was all HTML, no plain/text part | 17:08.15 |
kens | :-( | 17:08.29 |
| Nope nothing there yet | 17:08.47 |
tor8 | I had to save it and open it up in firefox in order to see what it said... | 17:08.48 |
| :( | 17:08.52 |
kens | status is still 'pending' I gues I won;t expect an email till that changes | 17:09.25 |
n0ldor | hi | 17:19.28 |
| can you help me? ^^ | 17:19.40 |
ray_laptop | n0ldor: with what ? | 17:20.17 |
kens | aha ray_laptop couldn't stand the tension and blinked first :-) | 17:20.34 |
n0ldor | i am trying to export a PDF to images and i download a script but i don't know how to export all pages | 17:20.58 |
| i paste | 17:21.02 |
ray_laptop | kens: nah, I just got bored with gdb | 17:21.02 |
n0ldor | system('/usr/bin/gs -q -sDEVICE=jpeg -dBATCH -dNOPAUSE -dFirstPage='.$_POST['pagenumber'].' -dLastPage='.$_POST['pagenumber'].' -r250x250 -sOutputFile=./'.$_POST['pagenumber'].'.jpg ./'.$_FILES["file"]["name"].'');.'.jpg ./'.$_FILES["file"]["name"].''); | 17:21.30 |
kens | n0ldor : well I'd suggest you drop FirstPage and LastPage without that GS will process all the pages | 17:22.08 |
ray_laptop | n0ldor: and use -%d so that each page has a separate file | 17:22.38 |
kens | and substitute %d in the output filename so that they don't all end up in the same file | 17:22.41 |
ray_laptop | ha. I typed (a bit) faster | 17:22.54 |
kens | I typed more letters :-) | 17:23.01 |
| But I admit I don't really grok the script | 17:23.43 |
n0ldor | ok thanks i go to try | 17:23.57 |
ray_laptop | n0ldor: so try: system('/usr/bin/gs -q -sDEVICE=jpeg -dBATCH -dNOPAUSE -r250x250 -sOutputFile=./'.$_POST-%d.'.jpg ./'.$_FILES["file"]["name"].â);.'.jpg ./'.$_FILES["file"]["name"].â); | 17:24.00 |
| kens: me neither -- just guessing | 17:24.23 |
| is that perl or something else even yuckier ? | 17:24.54 |
kens | No clue, that's why I don't grok it | 17:25.10 |
ray_laptop | oh, well. Back to gdb fun :-( | 17:25.34 |
n0ldor | hmmm it show to me an error | 17:25.48 |
| Parse error: syntax error, unexpected '%' in /home/casinopontevedra/.ftp-users/web/pdftoimage/helper_class.php on line 7 | 17:25.59 |
kens | THen you need to esapce the '%' in %d so that you parser doesn't get cross about it | 17:26.26 |
| Can't help you there, I know nothing about PHP | 17:26.38 |
n0ldor | ok i go to test more | 17:27.24 |
| another thing, how can i configure the quality of the jpeg? | 17:27.56 |
ray_laptop | n0ldor: then try system('/usr/bin/gs -q -sDEVICE=jpeg -dBATCH -dNOPAUSE -r250x250 -sOutputFile=./out%d.jpg ./'.$_FILES["file"]["name"].â);.'.jpg ./'.$_FILES["file"]["name"].â); | 17:28.14 |
| n0ldor: you can get better rendering quality by adding -dTextAlphaBits=4 -dGraphicsAlphaBits=4 | 17:29.00 |
kens | woudl say 'don't use JPEG' | 17:29.18 |
ray_laptop | kens: there is that. n0ldor: PNG is a *MUCH* better choice for images | 17:29.51 |
| the corresponding gs device is 'png16m', so use: -sDEVICE=png16m | 17:30.32 |
n0ldor | ok Thanks! i go to test again hehehe | 17:31.42 |
Robin_Watts | Aha. Reservation says "ticketed" | 17:33.43 |
n0ldor | perfect!! i show you how works | 17:33.43 |
| system('/usr/bin/gs -q -sDEVICE=png16m -dBATCH -dNOPAUSE -r250x250 -sOutputFile=./out%d.png ./' . $_FILES["file"]["name"]); | 17:33.45 |
ray_laptop | n0ldor: still, using -dTextAlphaBits=4 -dGraphicsAlphaBits=4 will improve the image quality | 17:37.00 |
kens | ah my status is now ticketed | 17:37.35 |
ray_laptop | n0ldor: system('/usr/bin/gs -q -sDEVICE=png16m -dTextAlphaBits=4 -dGraphicsAlphaBits=4 -dBATCH -dNOPAUSE -r250x250 -sOutputFile=./out%d.png ./' . $_FILES["file"]["name"]); | 17:38.01 |
n0ldor | ok I go to test the change with this | 17:39.03 |
Robin_Watts | tor8: ping ? | 17:39.03 |
n0ldor | yes! perfect | 17:41.02 |
Robin_Watts | ray_laptop: system('/usr/bin/gs -q -sDEVICE=png16m -dDownScaleFactor=3 -dBATCH -dNOPAUSE -r750 -sOutputFile=./out%d.png ./' . $_FILES["file"]["name"]); | 17:41.05 |
kens | OK got to go hopefully email will be here when I get back tonight. Goodnight all | 17:41.06 |
n0ldor | thanks a lot!! | 17:41.14 |
Robin_Watts | n0ldor: Try my version and compare. | 17:41.22 |
n0ldor | ok i try the 2 options | 17:41.42 |
ray_laptop | Robin_Watts: right -- that'll post process anti-alias the entire page, including images. Thanks | 17:43.15 |
n0ldor | the quality of the Robin_Watts it's better but the size it's the double | 17:44.00 |
| :P | 17:44.02 |
Robin_Watts | file size? | 17:44.14 |
n0ldor | yes | 17:44.18 |
| file size sorry | 17:44.24 |
| ;) | 17:44.27 |
Robin_Watts | That's the price you pay for more quality :( | 17:44.37 |
n0ldor | hehehe yes | 17:44.43 |
Robin_Watts | You could pngcrush it ? | 17:44.43 |
n0ldor | with this file size i can work good, no problem | 17:45.16 |
| thanks a lot, you save my life hehe | 17:46.12 |
Robin_Watts | bah. the mupdf projects have been updated to VS2012 :( | 17:46.30 |
ray_laptop | that was probably mvrhel (just guessing) | 17:49.25 |
Robin_Watts | It was. | 17:49.37 |
ray_laptop | I know it wasn't me | 17:49.38 |
| :-) | 17:49.40 |
Robin_Watts | But I suspect I reviewed the commit, so I can't escape the blame :) | 17:50.02 |
ray_laptop | but I've almost updated gs projects to VS2008 inadvertently | 17:50.14 |
| I have to make sure and have a separate copy and use that with VS2008 so the up-conversion from 2005 doesn't change the originals | 17:51.09 |
| I guess I could just use 2005 all the time, but... | 17:51.32 |
Robin_Watts | I use VS2005 all the time, partly because of this, and partly because you can disable (but not remove) a breakpoint in 2005. | 17:51.57 |
| In 2008 and above you can only remove it, AFAICT. | 17:52.08 |
ray_laptop | Robin_Watts: I do that all the time in 2008 | 17:52.13 |
Robin_Watts | hmm. Maybe I'm being dim then. | 17:52.40 |
ray_laptop | if you right-click on the little red BP symbol, it's on the list. Also you can uncheck it from the "Breakpoints" window | 17:53.26 |
| now the thing is, that the red ball goes away with 2008, so you can only see it on the "Breakpoints" window | 17:54.29 |
| so that's the only place you can re-enable it | 17:55.21 |
| Just one of the many bugs in 2008 | 17:55.47 |
Robin_Watts | ah, that's what I was missing. | 17:57.54 |
| I like to right click the ball and choose 'disable' and have it shown as an empty circle. | 17:58.09 |
ray_laptop | Robin_Watts: yes I agree, that's nicer | 17:58.42 |
| but often the place I want to re-enable isn't in that file's window anymore anyway, so using the Breakpoints window is easier then | 17:59.29 |
Robin_Watts | paulgardiner: ping ? | 18:11.37 |
paulgardiner | Robin_Watts: pong | 18:12.16 |
Robin_Watts | care to nod at this for me please? | 18:12.26 |
| http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=5106b010ec248bf254effcb3445a16dc7c4b015b | 18:12.27 |
| My git checkout is still going. I expect at least another 3 hours :( | 18:12.54 |
paulgardiner | Robin_Watts: looks okay. Why was it needed? | 18:14.19 |
Robin_Watts | cos mvrhel (ably assisted by me) accidentally commited the change to VS2012 format. | 18:14.48 |
paulgardiner | Hang on. It refers to different project files. | 18:15.08 |
| Ah right | 18:15.17 |
| LGTM then | 18:15.25 |
Robin_Watts | Thanks. | 18:15.30 |
ray_laptop | Robin_Watts: maybe we need a git hook for us Windoze users to reject changing the project files (so we have to use -f to actually change them) | 18:29.04 |
Robin_Watts | nah, it happens rarely enough that it's not a huge problem. | 18:30.11 |
| car booked for staff meeting. Miles said to get a reasonable sized one, so Tor should escape having to go in the boot. | 18:33.14 |
ray_laptop | well, I guess I understand why the pdf14 compositor device is left in the chain. There are way too many places that have hooks to the device | 18:39.58 |
| Robin_Watts: boots in TX are something on your feet. They call it a trunk ;-) and Tor is long enough that even folded up he probably wouldn't fit. Just strap him to the top like they do with things they kill | 18:41.16 |
| nobody there will probably look twice ;-) | 18:41.47 |
Robin_Watts | If anyone asks, we'll tell them he's a democrat. | 18:42.41 |
ray_laptop | now to see if I can 'disable' the pdf14 accumulator compositor properly :-( | 18:42.47 |
| *was* a democrat, now he's a good democrat | 18:43.54 |
| (the only kind that's good in TX) | 18:43.55 |
Robin_Watts | oh ass. mvrhels project change lost us all the dependencies. | 19:28.54 |
| tor8, paulgardiner, sebras: ping ? | 19:43.07 |
sebras | Robin_Watts: pong. | 20:12.33 |
Robin_Watts | sebras: Can you nod at: | 20:12.49 |
| http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=0c0cc6897fd8fa967a24b08d3981148411755eac | 20:13.05 |
| please ? | 20:13.07 |
sebras | Robin_Watts: I checked platform/win32/mupdf.sln in 9a83c56~1 vs your 0c0cc68 and I see no differences, to me that means that this is a revert patch. though I know next to nothing about VS and its sln-files... | 20:18.53 |
Robin_Watts | It is exactly a revert patch. | 20:19.12 |
| mvrhel updated the solution to 2012. I cocked up in putting it back to 2005 in the last commit as I dropped all the dependencies. | 20:20.09 |
| This completes the reversion and gets us something that actually builds. | 20:20.20 |
| so do I take it you're happy enough for me to commit it? | 20:21.37 |
Robin_Watts | heads for food. bbl. | 20:25.21 |
sebras | Robin_Watts: I am, despite my limited knowledge about VS. :) | 20:26.44 |
zeniko | tor8: Robin_Watts: may I limit the samples array argument to xps_parse_color to FZ_MAX_COLORS (for fixing bug 694957)? | 20:54.19 |
Robin_Watts | zeniko: looking... | 20:58.03 |
tor8 | ray_laptop: Robin_Watts: trying to give me more excuses to miss my flight? ;) | 20:58.39 |
| zeniko: hm. that's a buffer overflow waiting to happen... | 20:59.03 |
| in the code as it is now | 20:59.14 |
Robin_Watts | tor8: 2 commits on robin/master | 20:59.20 |
tor8 | paulgardiner: ah, you're working today? there's a small patch on tor/master for you. | 21:00.15 |
Robin_Watts | I suspect paulgardiner has stopped for the day, but he said he'd be working tomorrow. | 21:01.03 |
| zeniko: Looks sensible to me. | 21:01.31 |
tor8 | zeniko: I looked through the transfer function patch. I'm not entirely happy about the device interface to it, since the call happens at the end of the transparency group. | 21:02.48 |
| somehow I would be more comfortable having the transfer function set up when we push the group rather than at the pop | 21:03.33 |
zeniko | tor8: so rather pass it as an additional argument and store it in the draw stack? | 21:04.18 |
tor8 | zeniko: if that would work, I think I'd prefer it. | 21:04.46 |
| Robin_Watts: the project file commit LGTM | 21:05.57 |
| Robin_Watts: so if I read this correctly, we keep the old objects that have been loaded around in the xref table so that we don't free them from underneath fz_dict_gets | 21:09.08 |
| but we make no such guarantee for the trailer | 21:09.08 |
| do I understand the source of the problem correctly? | 21:09.39 |
zeniko | for review: http://git.ghostscript.com/?p=user/zeniko/mupdf.git;a=commitdiff;h=60dabde18d7fe12b19da8b509bdfee9cc886aafc | 21:10.21 |
tor8 | if that's the case, I wonder if it would be easier to either (a) make a shallow copy of the trailer dictionary and work off that, or (b) disconnect the trailer dictionary temporarily | 21:10.55 |
| or (c) fz_keep/drop it while guarding | 21:11.35 |
| zeniko: LGTM. | 21:12.52 |
Robin_Watts | tor8: yes. | 21:16.39 |
| I've got to run for food, sorry. | 21:16.40 |
| I'll think about, but I'm not sure I can see how a copy of the trailer would work. | 21:16.40 |
zeniko | tor8: thanks, pushed. | 21:16.40 |
mvrhel_laptop | Robin_Watts: so I sent you an email about the car | 22:32.07 |
| maybe a minivan would be a good idea if you all can wait a bit for me | 22:32.30 |
| henrys: when are you arriving? | 22:32.44 |
| or marcosw1 | 22:32.51 |
| or ray | 22:32.58 |
henrys | thursday night also I haven't booked yet | 22:33.13 |
| mvrhel_laptop: I'll try and get a time close to yours so we'll have some options | 22:34.05 |
mvrhel_laptop | ok sounds good | 22:34.10 |
henrys | mvrhel_laptop: are you coming back sunday - maybe just sent me your itinerary. | 22:36.31 |
mvrhel_laptop | reading the logs. Robin_Watts sorry about accidently updating the win32 project | 22:36.50 |
| I usually make sure that I uncheck that one | 22:37.00 |
| henrys: ok will do | 22:37.16 |
| I am heading back Sunday yes | 22:37.22 |
| henrys: ok sent it to you | 22:41.31 |
henrys | okay I'll book as closely as I can mvrhel_laptop | 22:41.54 |
mvrhel_laptop | sounds good | 22:42.06 |
| Robin_Watts: you can disable breakpoints in all the VS2008, 2010, 2012, 2013 | 22:44.59 |
| versions | 22:45.10 |
| anyway enough log reading. back to real work | 22:45.19 |
| henrys: I am going to play around a bit with the v4 windows driver stuff. it is supposed to be pretty easy to get something working | 22:46.04 |
henrys | mvrhel_laptop: okay sounds like fun (for some people ;-)) | 22:46.47 |
mvrhel_laptop | see if I can throw together something interesting | 22:47.03 |
Robin_Watts | mvrhel_laptop: (For the logs). I have booked a car, but we can always upgrade it if we need more seats. | 23:55.27 |
| mvrhel_laptop: I have booked a car, but we can always upgrade it if we need more seats. | 23:58.51 |
| Forward 1 day (to 2014/01/17)>>> | |