| <<<Back 1 day (to 2012/10/31) | 2012/11/01 |
Robin_Watts | Well, that's odd... line 1887 of gdevtsep.c | 10:35.13 |
kens | testing for 8 bits per component, why is that odd ? | 10:35.52 |
Robin_Watts | the if 2 lines later :) | 10:36.13 |
kens | Ah yes, well thatr's clearly nonsense | 10:36.31 |
| I'd have to guess that's some kind of error but exactly waht its supposed to be escapes me | 10:37.02 |
| I can't see any good reason for the test, you're right | 10:37.17 |
Robin_Watts | Well, evidently I changed that code, so it's my fault. ho hum. | 10:37.42 |
| morning mvrhel | 11:29.27 |
| chrisl: Do we have any devices that output 1bpp cmyk tiffs that work? | 12:13.34 |
| 1bpc, I mean, obviously. | 12:13.45 |
chrisl | I assumed tiffsep1 would do, since tiffsep does composite | 12:14.33 |
Robin_Watts | I don't get composite from tiffsep1. | 12:15.14 |
| just CMYK | 12:15.17 |
| Sorry. I mean, I get out(Black).tif, out(Cyan).tif, out(Magenta).tif and out(Yellow).tif, but no out.tif | 12:15.48 |
chrisl | Yes, I got what you meant - maybe we don't have such a device, then - odd | 12:16.16 |
Robin_Watts | You can do -dBitsPerComponent=1 with tiffsep, but the composite image I get is broken. | 12:18.02 |
| In fact all the images I get are broken. | 12:18.34 |
| Now, it's possible that I broke that, of course, but it's not recent. | 12:18.47 |
chrisl | More likely it was when we did away with the compressed colour encoding | 12:19.22 |
| Annoyingly, it looks like we don't have a device of that type :-( | 12:22.46 |
| Oh, does psdcmyk honour BitsPerComponent? | 12:23.57 |
| No, seems not...... | 12:25.11 |
Robin_Watts | I'm gonna dive back in time with git and see when it broke/if it ever worked. | 12:26.33 |
chrisl | The other possibility is that it's never worked right with the libtiff integration...... | 12:30.00 |
Robin_Watts | OK, so you could specify -dBitsPerComponent back in May, and it would give a valid image - but that's because it was ignoring -dBitsPerComponent. | 12:30.29 |
chrisl | Well, that is a valid mode of operation..... | 12:31.22 |
kens | I see Henry's morning ulcer has arrived.... | 12:43.04 |
Robin_Watts | I think my tif viewer doesn't like 1bit cmyk tifs. | 13:13.50 |
| What do people use for viewing tifs? | 13:14.01 |
kens | Many don't what are you using ? | 13:14.04 |
Robin_Watts | Picasa. | 13:14.11 |
kens | I use an old version of Photoshop | 13:14.11 |
| chrisl just managed to write a 7GB PDF file and open it in Ghostscript and Acrobat. 32-bit executable on WIndows | 13:25.07 |
Robin_Watts | nice | 13:55.12 |
kens | Now I'm trying ps2write | 13:55.48 |
| Produces a file slightly under 2GB I may have to add a few pages | 13:57.36 |
| But Distiller seems happy with it | 13:58.08 |
| Hmm, the cluster has just given me 10 errors of the form: | 14:03.53 |
| tests_private/customer_tests/A1TEST_CHART.hp2.pbmraw.600.1 x6 Input_file_missing | 14:03.53 |
| So... WHat's happening there ? | 14:04.10 |
| I'm sure it can't be my code :-) | 14:04.30 |
Robin_Watts | indeed not. | 14:04.38 |
| I suspect an svn update went screwy | 14:04.50 |
| all problems from the same machine? | 14:04.59 |
kens | Not sure, one sec | 14:05.06 |
| No vartious different ones | 14:05.23 |
| Hmm, at least I assume the machine name is the one that failed | 14:06.17 |
paulgardiner | Robin_Watts: if you have a tiff that you think Picassa is displaying wrongly, bung me a copy and I'll convert it to a more commonly format with GIMP. I'd imagine GIMP would get it right. | 14:06.50 |
kens | I don't think GUMP reads 1-bit CMYK TIFF eitehr | 14:07.07 |
| GIMP | 14:07.13 |
paulgardiner | Ah ok. That wont help then. | 14:07.31 |
kens | I can read them in Photoshop though so if its a problem I could do that | 14:07.32 |
Robin_Watts | paulgardiner: I have gimp, and it doesn't work. | 14:07.36 |
| I can convert them using imagemagick to png. | 14:07.45 |
paulgardiner | You could generate 8bit but using only 255 and 0. Should that give the same effect? | 14:09.53 |
| ... I'm assuming you are testing the dither not the output rountines. | 14:10.34 |
Robin_Watts | paulgardiner: This isn't test code. | 14:19.31 |
| I need proper 1bpp. | 14:19.54 |
| Aha. I have 1bpp separations working. The composite is still broken though. | 14:29.13 |
| Ah, I see why. | 14:34.49 |
| The code to build a composite from the planes heavily assumes 8bpp. | 14:35.09 |
| So I am tempted to revert to the situation where we only get a composite if we are working in 8bpp. | 14:35.45 |
kens | chrisl ping | 15:16.36 |
chrisl | kens: pong (sorry, been working downstairs......) | 15:26.10 |
kens | Not a problem | 15:26.18 |
| THoguht you would like to know that my tests are done. | 15:26.32 |
chrisl | Cool, so ps2write worked, too? | 15:26.44 |
kens | Tested 32 & 64-bit Windows versions of pdfwrite and ps2write | 15:26.47 |
| 10GB temp file, 7Gb PDF file, 6GB PostScript file, all now work with a 32-0bit version of GS and Acrobat/Distiller | 15:27.14 |
chrisl | Excellent! | 15:27.32 |
kens | I hope so! Of course, its not what you could call *well* tested.... | 15:27.49 |
| But it looks pretty good so far | 15:28.02 |
chrisl | Well, same problem as I had - we don't exactly have a wide range of files available. That's why I decided just to commit it, and wait for "real world" testing | 15:28.36 |
kens | Yes, I've just done the same, it doesn't seem to break anything on the cluster (and I tried a 32 bit cluster test) and it works OK for the very limited set of testing I can reasonably do, so fingers crossed :-) | 15:29.23 |
chrisl | Something nice for the 9.07 release notes, I reckon :-) | 15:30.27 |
kens | Yes, that and linearisation :-) | 15:30.43 |
| And if I'm very lucky, colour management too | 15:30.56 |
chrisl | So now your multigigabyte PDF can be optimised for web viewing, too!! | 15:31.29 |
kens | Err, I'm not sure that works.... | 15:31.39 |
| I figured it was a stupid idea with a file that big.... | 15:31.50 |
| Having said thatm it *ought* to work | 15:32.03 |
| Because I wrote all that stuff with 64-bit data types | 15:32.21 |
chrisl | I definitely *is* a stupid idea - we should run a sweepstake for how long it takes someone to ask about it...... | 15:32.45 |
kens | Not long I bet.... | 15:32.56 |
| For the first time in months I actually feel like I'm making headway against my backlog | 15:35.46 |
chrisl | I've been locked out of gmail IMAP because of a Thunderbird bug - terrific :-( | 15:36.00 |
| kens: I *really* hope no one tries to apply the large file patch - that would be a *bad* idea, I reckon...... | 15:38.17 |
kens | chrisl I think so too | 15:38.27 |
| Even the patch for pdfwrite is quite large | 15:38.35 |
| and has tentacles all over | 15:38.42 |
| But I needed to make it clear it *wouldn't* work on its own | 15:38.53 |
chrisl | I just hope they add up to more than anyone would want to try to "back-patch"...... | 15:40.27 |
kens | I certainly hope that's the case, but you can't stop people.... | 15:40.50 |
chrisl | We can not help them, though | 15:41.09 |
kens | :-) | 15:41.14 |
| If anyone wants the code (and it can't be a common desire, surely) then I will encourage them to take a snapshot | 15:41.42 |
henrys | november 1, candy for breakfast day | 15:44.07 |
kens | Ah, we have a (relatively) tame teenager, and hence no need for that. | 15:44.36 |
Robin_Watts | Well, the bad news is that tiffsep going to 1bpp with my simple FS error diffusion looks much nicer than the ETS code :( | 15:49.34 |
chrisl | Have you tried printing the output? | 15:50.26 |
Robin_Watts | chrisl: no. | 15:50.35 |
henrys | Robin_Watts:I guess you have looked at a ramp of resolutions, is it possible the code has numerical problems at different resolutions? | 15:50.44 |
Robin_Watts | I haven't, I've only looked at 72dpi, but I don't believe resolution should be a factor. | 15:51.14 |
henrys | I am guessing it would have been tested at 720 and 1440 | 15:51.15 |
Robin_Watts | At the moment I'm tidying up changes to tifsep (so it works at 1bpp) and tiffscaled4. | 15:52.06 |
| then I'm going to add some flags to the ETS test code so we can run it with different features disabled. | 15:52.47 |
| That way we can get a true comparison. | 15:52.54 |
henrys | tor8: do you remember anything about the ets screening output? I just remember oohs and aahs at the staff meeting ... Robin_Watts' output wouldn't elicit that response. | 15:53.23 |
Robin_Watts | I disabled everything in the ETS code so that it should drop back to just being simple serpentine FS, and it still looked worse. | 15:53.29 |
| ew and urgh, maybe. | 15:53.43 |
chrisl | Unless the test files at the staff meeting were cherry picked to highlight the problem with FS that ETS is intended to address..... | 15:58.09 |
Robin_Watts | chrisl: OK, so let's suppose I try printing some of these things... | 16:22.58 |
tor8 | henrys: I remember it always needing more tweaking and never being quite there. the oohs and aahs were of the inkjet printouts though. | 16:23.13 |
Robin_Watts | Any suggestions for how I can print this output on my canon inkjet MFP ? | 16:23.38 |
| Is that something that maybe cups can help with ? | 16:25.20 |
chrisl | I guess maybe. You might be able to convert the TIFF to a cups raster, and poke that into the print pipeline. Of course, the person to ask doesn't seem to be here right now! | 16:26.34 |
| The other problem is that a lot of these inkjets actually take RGB data and convert it, which would mess up the output | 16:28.14 |
Robin_Watts | chrisl: My printer is supported by gutenprint, apparently... | 16:30.18 |
henrys | Robin_Watts:could you post something at say 360 dpi? just curious... or maybe I can just download and build it. | 16:30.45 |
Robin_Watts | henrys: Sure. Just a tick. | 16:31.04 |
chrisl | Yeh, what I'm not sure about is whether you can "inject" a cups raster just before the gutenprint filter (to convert to the printer's PDL), and whether that will mess up the color info. | 16:31.52 |
Robin_Watts | henrys: I'm gradually populating http://ghostscript.com/~robin/ETS/ | 16:38.58 |
| The two images there are simple FS from gs and ETS (ETS with everything turned off, so it SHOULD just be simple FS) | 16:39.32 |
| The nasty vertical lines in the dither are still there, and there are the bands of yellow that I don't understand yet. | 16:40.07 |
kens | The yellow ios indeed odd | 16:42.19 |
| Looks like the magenta plate isn't marked and should be | 16:42.31 |
| Some areas nearby look 'greenish' too | 16:43.00 |
| In fact a number of freatures appear to have coloured 'halos' around them | 16:44.07 |
Robin_Watts | I've put a 'nonserpentine' version up there. | 16:44.31 |
kens | One sewc | 16:44.43 |
| sec | 16:44.45 |
| Thehalos are still there for me | 16:45.12 |
Robin_Watts | The grey background looks better (more even) in the ETS nonserpentine thing than it does in the vanilla FS. | 16:45.25 |
| I would not be suprised if I'd broken something in doing the serpentining, cos I just did it as a quick hack. | 16:45.57 |
| But the halos are indeed still there. | 16:46.06 |
kens | I'd have to agree that teh non-serpentine is nicer gray | 16:46.13 |
| The tiffsep hjas hte same halo effect | 16:46.47 |
| But I'm not sure what screening that is | 16:47.02 |
Robin_Watts | kens: Where are you seeing the halo effect? | 16:47.03 |
| The tiffsep image is my simple FS. | 16:47.21 |
kens | Look at the yellow tooth in the tongue | 16:47.21 |
| It has a pake shado on the tongue | 16:47.36 |
| I'm not seeing that in a GS to RGB | 16:47.53 |
| (screen | 16:47.57 |
| If I could remember the incantation I'd try a CMYK to screen | 16:48.17 |
Robin_Watts | Let me upload a png made from the cmyk source that this is based on (given by pamcmyk32) | 16:49.21 |
kens | That would be good | 16:49.28 |
chrisl | png is RGB, yes? | 16:50.13 |
Robin_Watts | chrisl: In all cases, I'm working in pams, but pams are too large to upload, so I convert to png. | 16:50.51 |
| The source is there now. | 16:51.08 |
chrisl | Yeh, I'm just wondering if we're seeing artifacts of the conversion to RGB | 16:51.21 |
Robin_Watts | In all the different screenings, I see the magenta bleed into the whiskers, sometimes quite badly. That's wrong. | 16:51.53 |
kens | Yes | 16:52.22 |
Robin_Watts | Ah,no, sorry. | 16:52.23 |
| In all the *ETS* screenings, I see the magenta bleed into the whiskers. | 16:52.38 |
kens | Yes, I think this is the problem all over | 16:52.55 |
Robin_Watts | The tiffsep one has no horrible effect on the tongue either. | 16:53.06 |
kens | THe magenta (at least) is 'mis-registered' | 16:53.09 |
| I'm looking at a CMYK composite tiff and these effects are not present there | 16:53.26 |
| (8 bpc) | 16:53.33 |
Robin_Watts | kens: from gs? | 16:53.45 |
kens | Yes | 16:53.50 |
| using tiiger.eps fropm ghostpdl/gs/examples | 16:54.03 |
| I could try a 1-bit and recombine in Photoshop | 16:54.27 |
| what deos 1-bit seps ? | 16:54.37 |
Robin_Watts | I'm uploading a tif version of the tiger_360_source now. | 16:54.52 |
kens | OK | 16:54.58 |
Robin_Watts | I hope that's CMYK - can you try that in PS please? | 16:55.06 |
| gawd, it's huge. Let me try to compress it. | 16:55.25 |
kens | I see only 2MB | 16:55.53 |
| Is it bigger than that ? | 16:56.09 |
Robin_Watts | I stopped uploading - it was going to take 20 mins. | 16:56.26 |
kens | OMG | 16:56.30 |
| Well my compositge was default resolution, not sure what that is for tiff32nc | 16:56.59 |
chrisl | 72dpi I'd think | 16:57.11 |
kens | COUld be, its fairly rough | 16:57.20 |
Robin_Watts | OK, uploading now. | 16:57.22 |
kens | 360 dpi comes out at 49MB | 16:57.59 |
Robin_Watts | ok, it's there. It's lzw so 2.5Meg | 16:58.22 |
kens | one moment | 16:58.29 |
| Loks the same as mine | 16:59.03 |
| No artefacts | 16:59.08 |
| whiskers are all white | 16:59.17 |
Robin_Watts | Is it CMYK in that .tif? | 16:59.27 |
kens | Photshop says yes | 16:59.36 |
Robin_Watts | excellent. | 16:59.45 |
| Right, so there IS a problem with ETS. Or at least in how I'm driving it. | 17:00.00 |
| At least I have something to aim for now. Thanks all. | 17:00.09 |
kens | Well the 'mis-registration' is pretty bad | 17:00.14 |
Robin_Watts | kens: yeah, I really want to understand that. | 17:00.26 |
chrisl | BTW, gimp can view CMYK tiffs now, it just converts to RGB "en-route" | 17:00.29 |
kens | Oh that's new | 17:00.37 |
| So it should be able to handle the file Robin sent me | 17:00.50 |
Robin_Watts | chrisl: so it converts to RGB as it loads? That's not much help :( | 17:01.43 |
kens | Would 4 8-bit spes from this file help you ? | 17:02.02 |
| seps | 17:02.08 |
Robin_Watts | kens: No thanks. | 17:02.15 |
chrisl | Yes, it's not great. But it's better than it used to be when it assumed it was RGBA..... | 17:02.24 |
Robin_Watts | I can get the data out in any form I'd like by tweaking the file saving. | 17:02.44 |
| but thanks for the offer. | 17:03.02 |
kens | Robin_Watts : do you see what I mean by hte 'halo' around objects ? | 17:04.06 |
Robin_Watts | kens: can you point me at such an object? | 17:04.40 |
kens | A lot of the black stripes in the top of the tigers head have 'yellow' halos round them | 17:05.17 |
Robin_Watts | Right, yes, I can certainly see the yellow bands. | 17:05.34 |
kens | Not the birght yellow but just more yellow than orange | 17:05.36 |
Robin_Watts | Yes. | 17:06.12 |
kens | Aha, that's fine then | 17:06.22 |
Robin_Watts | AHAHA! | 17:06.55 |
kens | ? | 17:07.06 |
Robin_Watts | There was some code I hadn't switched off. | 17:07.22 |
kens | Oh, interesting | 17:07.30 |
Robin_Watts | Now I've turned that off, I get no banding/halos. | 17:07.33 |
| and we're back to simple FS. | 17:07.40 |
kens | THat sounds like a distinct improvement | 17:07.51 |
Robin_Watts | Let me update the pictures online. | 17:07.53 |
kens | Yes, that looks better Robin_Watts | 17:11.26 |
| Seems to be no 'lines' either | 17:12.27 |
Robin_Watts | ok, both the serpentine and the non-serpentine have been updated. | 17:12.39 |
chrisl | That is *much* better | 17:12.39 |
Robin_Watts | and I prefer the non-serpentine background. | 17:12.51 |
kens | I'm inclined to asgree | 17:13.26 |
Robin_Watts | and I think I prefer the ETS output now to the gs one. | 17:14.25 |
kens | tiffsep is the gs one ? | 17:14.52 |
Robin_Watts | indeed. | 17:15.00 |
kens | Hmm the GS one has funny wavy vertical bnanding on my screen | 17:15.48 |
| The ETS one is nicer | 17:16.01 |
Robin_Watts | kens: Yes. | 17:16.06 |
kens | I woudl say the gradients are betgter too | 17:16.30 |
Robin_Watts | Such 'wormy' patterns are what ETS is intended to solve - they are a standard artifact of FS dithering. | 17:16.32 |
| Which leads me to wonder if there is still some ETS magic that I haven't disabled yet ;) | 17:16.50 |
kens | Well the two are definitely not the same.... | 17:17.08 |
Robin_Watts | I certainly wouldn't expect exact bit for bit matches! | 17:17.29 |
| I think ETS keeps the errors to a higher accuracy than I do, maybe. | 17:17.47 |
kens | No, but the background is quite different | 17:17.52 |
| THe GS output is more 'grainy' in some places, the ETS output seems smoother | 17:18.23 |
Robin_Watts | yeah. | 17:18.29 |
kens | Anyway I have to go, Melanie's riding lesson | 17:20.15 |
| Goodnight all | 17:20.36 |
henrys | so you've disabled all the ets stuff and we should be left with FS? | 17:21.01 |
Robin_Watts | Yes. | 17:21.08 |
| I think. | 17:21.21 |
| Ah, yes, he's working in at least 16bits of error, and I'm working in 8. Dunno if that explains the difference between the gs and ets with everything turned off, but I'm prepared to grasp at that straw for now. | 17:23.03 |
henrys | would it be difficult to update your stuff to 16 bit - it would give us a sane starting point at least to know we are back at FS | 17:26.48 |
Robin_Watts | not sure I can really see a sane way of doing that. | 17:28.52 |
| henrys: I believe we're now at the point where the ETS code with everything turned off, looks slightly nicer than than the plain FS code from gs. | 17:32.10 |
| So this isn't a bad place to start work from. | 17:32.17 |
| If I can get it so that as I turn stuff back on, I always see an improvement, we'll be golden. | 17:32.46 |
henrys | makes sense | 17:34.50 |
Robin_Watts | I would love to see any old versions of this code we may have. | 17:36.54 |
| Does anyone have anything other than this last version? | 17:37.12 |
henrys | rayjj would have it if we have it, I doubt we do. | 17:39.21 |
| you definitely have the classic FS worms under the tiger's tongue in the ets output | 17:41.03 |
Robin_Watts | I think I may see at least part of what is wrong here. | 18:09.26 |
| The potted concept for ETS is that for a given greyscale, you can predict how far each dot should be on average from other dots. | 18:10.09 |
| And so you use the difference between the *actual* distance and the expected distance to bias the threshold function. | 18:10.36 |
| What Raph does is to add this biasing into the error term, which seems to do the same thing, but actually doesn't. | 18:11.11 |
| The difference in the error term gets propagated onto subsequent pixels. | 18:11.32 |
| That propagated error should NOT include the biases that have been added. | 18:11.58 |
| henrys: OK. tiger_ets_360dpi.png online | 18:30.50 |
| That's with all the ETS magic turned back on, but with the errors not being propagated. | 18:31.11 |
henrys | those vertical lines under the tongue don't seem right to me, but ... | 18:33.39 |
| but look better than the worms | 18:34.01 |
Robin_Watts | Where are the vertical lines? | 18:34.18 |
henrys | vertical white zig-zag under the tongue | 18:35.04 |
Robin_Watts | I can't see that. | 18:35.45 |
| The background looks worse with ETS on than off. | 18:35.57 |
henrys | are you zooming in? | 18:36.32 |
Gigs- | I've got some older "linearized" PDFs with standard table xrefs... the top xref generation number is "6", does that mean anything special? | 18:39.12 |
Robin_Watts | henrys: 1:1 | 18:39.48 |
| Gigs: In linearised PDFs *all* generation numbers must be 0. | 18:40.14 |
Gigs- | thats what I thought too, my program was reporting them as incremental PDFs with broken startxref offsets, but it turns out it's some sick person's idea of a linearized pdf | 18:40.57 |
| the startxref is "0" as well, even though there's another object before the top xref :( | 18:41.12 |
| I tried that mutool Robin... it doesn't seem to touch that kind of thing unfortunately | 18:41.34 |
| mutool clean doesn't seem to be very agressive about fixing xref problems | 18:41.51 |
Robin_Watts | Gigs: It does if you give it the right options. | 18:42.37 |
Gigs- | oh really | 18:42.50 |
Robin_Watts | By default, no object renumbering is done. | 18:42.53 |
| You said you wanted minimum changes. | 18:43.05 |
henrys | jesus is the screen grabbing stuff screwed up in mountain lion something is wrong | 18:43.24 |
Robin_Watts | If you do -g then it will do more stuff. | 18:43.32 |
Gigs- | Robin_Watts: cool, thanks | 18:43.38 |
Robin_Watts | henrys: I open both images in browser tabs and flip between them. | 18:43.54 |
| That way I can 'blink test' it. | 18:44.08 |
| Decorator has just left. I need to go help clean up. bbiab. | 18:44.24 |
henrys | I was just trying to chop out the artifact I see and mail it to you. | 18:44.37 |
Gigs- | Robin_Watts: I did want minimum changes before, but now that I've got it down to the ... degenerate cases... to abuse a math term, I think we can go with more radical surgery | 18:44.42 |
mvrhel_laptop | Robin_Watts: oh good. reading the logs. you have the code just running FS with serpentine | 18:45.19 |
| as a starting point | 18:45.28 |
henrys | I put black ovals around 3 of the zigzag lines http://www.ghostscript.com//~henrys/vlines.png | 18:50.57 |
mvrhel_laptop | yes. those lines are not good | 18:52.44 |
henrys | they are actually visible zoomed out they look like little tears | 18:53.20 |
mvrhel_laptop | is that from FS or ETS? | 18:53.49 |
henrys | ets there used to be worms there when it was fs | 18:54.21 |
mvrhel_laptop | ah | 18:54.35 |
| ok | 18:54.37 |
| worms would be preferable to vertical lines | 18:55.10 |
henrys | yes I agree | 18:56.54 |
| amazing how our visual system bears down on that. | 18:57.45 |
mvrhel_laptop | right | 18:58.23 |
| here is something for you. one company here was selling a tool that lets you write c# code for iOS and android. barf | 18:59.40 |
henrys | anyway robin_watts has been making really good progress with at least understanding things. | 18:59.49 |
mvrhel_laptop | henrys: yes it looks like it | 18:59.59 |
| talk is starting now on building mixe language apps | 19:00.36 |
henrys | I don't think that c# idea is going to get far but who knows. | 19:01.49 |
mvrhel_laptop | i doubt it too | 19:01.59 |
henrys | I've got to get back to this pcl bug for norbert, I'll be away from IRC for a bit. | 19:02.47 |
mvrhel_laptop | Robin_Watts: are you there? | 19:28.23 |
Robin_Watts | mvrhel_laptop: Pin | 21:26.39 |
| ping | 21:26.41 |
mvrhel_laptop | oh hi Robin_Watts | 21:26.45 |
Robin_Watts | sorry, have been manically cleaning up after painter in the lounge. | 21:27.14 |
mvrhel_laptop | I was just having issues with mupdf. I did not realize I needed to do the Submodule stuff with the third party libs | 21:27.15 |
| it still says on the website that you need to grab them as a zip file | 21:27.30 |
| also, are the v8 projects in the solution not supposed to build? | 21:27.57 |
| they fail due to v8.h not being found | 21:28.11 |
Robin_Watts | mvrhel_laptop: Yes, the website needs to be updated to talk about git submodules - but what is there are the instructions for the last release, which was zip based. | 21:29.11 |
| The v8 projects will not build unless you have the v8 thirdparty installed. | 21:29.26 |
| You can download it from http://ghostscript.com/~robin/ | 21:29.36 |
mvrhel_laptop | gotcha. it would be good to have the website updated when things like that change | 21:29.41 |
| oh another thirdparty secret | 21:30.13 |
Robin_Watts | yeah, we ought to have 'how to build a snapshot' and 'how to build from git'. | 21:30.21 |
| yeah, it'll all be tidied when we do a release. | 21:30.39 |
| tor has control of the website, and he must not be distracted from the viewer. <FX: Cracks whip> | 21:31.13 |
mvrhel_laptop | gotcha | 21:31.27 |
Robin_Watts | What file are you two seeing the vertical lines in? | 21:31.31 |
mvrhel_laptop | dont know. the one henry posted | 21:31.46 |
Robin_Watts | http://ghostscript.com/~robin/ETS/tiger_ets_360dpi.png <- That one is the latest from ETS. | 21:31.49 |
mvrhel_laptop | I dont see them in that one | 21:32.23 |
| got to go | 21:33.01 |
Robin_Watts | cu. | 21:33.07 |
| Right. THe only place we get vertical lines is in the serpentine FS only ETS. | 21:33.22 |
| I'm grudgingly agreeing with Raph that serpentine is a bad idea. | 21:33.38 |
| but the ETS code as supplied to me seems badly broken :( | 21:34.02 |
| Forward 1 day (to 2012/11/02)>>> | |