| <<<Back 1 day (to 2013/08/04) | 2013/08/05 |
Robbes | hi there, good morning | 07:20.43 |
sebras | Robbes: hi. | 08:49.43 |
tor7 | http://www.dkriesel.com/en/blog/2013/0802_xerox-workcentres_are_switching_written_numbers_when_scanning ... when jbig2 compression goes wrong | 09:10.33 |
kens2 | Wow, that's *really* bad... | 09:11.32 |
sebras | I saw this yesterday and I assumed you guys already knew about it... :) | 09:14.41 |
| on a related note: mirabilos reported yesterday (and I verified) that the testsuite for jbig2dec fails. do you guys run this in your cluster? | 09:15.42 |
kens2 | Only as part of GS/MuPDF | 09:16.06 |
| So no, we don't run any jbig2 test suites | 09:16.25 |
sebras | kens2: well that explains why mirabilos reported the issue and why the tests still fail. I think the jb2-files come from jbig2dec.sf.net, but the tests are actually in the jbig2dec repo at git.ghostscript.com. | 09:19.25 |
Robbes | ah, im back :) | 09:52.19 |
| I was playing around with gs in windows, 9.07 | 09:52.30 |
| i've been creating pdfa2 | 09:52.42 |
| What I saw was that the PDF version got 1.4, but PDF/A-2 should be PDF1.7. Why does it go down to 1.4 ? | 09:53.29 |
kens2 | oversight probably, raise a but report. | 09:56.36 |
| bug* | 09:57.19 |
| I'm not certainthat its a requirement of PDF/A-2 to be 1.7 though, I'd have to check the spoec | 09:57.46 |
Robbes | ok.. i'll fill in a bug report | 10:08.35 |
| another question: if I have a valid PDF/A-1a (created by MS Word), and I let gs create it into a PDF/A-1 or PDF/A-2, it comes out as a PDF/A-*b | 10:09.38 |
| why doesn't it leave the tags in (conformance with -1a and -2a) ? | 10:10.00 |
kens2 | we don't support a | 10:11.17 |
Robbes | yes, but.. now a user should first validate conformance, and then decide to use gs or not :/ | 10:12.38 |
| because it strips out the -1a conformance tags | 10:12.57 |
| but ok,.. nm | 10:13.01 |
| tnx for the answers | 10:13.12 |
kens2 | It does not strip them out | 10:13.14 |
| It creates a totally new PDF file | 10:13.25 |
| Robbes : ping | 12:12.47 |
| chrisl ping | 12:17.33 |
chrisl | kens2: pong | 12:17.55 |
kens2 | chrisl could you try the file and command in bug 694500 pleae ? | 12:18.21 |
| I've been unable to reproduce it, but we've seen differences in our Linuxes before | 12:18.50 |
chrisl | Is that the same image interpolation code that we use for rendering? | 12:19.42 |
kens2 | I expect it is, yes | 12:19.54 |
| I believe at this point it *is* rendering | 12:20.12 |
chrisl | Yes, I meant "normal" rendering, rather than pdfwrite fallback rendering | 12:20.39 |
kens2 | THey are the same, really. | 12:20.51 |
| the image just ends up somewhere different | 12:21.04 |
| I have suggested they try the latest source, but as you can see, no reply yet | 12:21.20 |
chrisl | Remember there was a bunch of seg faults came up in the light of the efficiency improvements Robin implemented | 12:21.28 |
kens2 | Hmm, well I'll give them a couple of days to respond and then close it if they don't say anything | 12:21.57 |
chrisl | kens2: debug exe works fine for me | 12:23.31 |
kens2 | Good enough, then I'll just give them some time to respond, thanks | 12:23.50 |
chrisl | Release binary works, too | 12:24.54 |
Robbes | tnx ken, for pointing out that PDFA-2 can be 1.4 as version | 12:25.50 |
kens2 | AH Robbes, I wanted to be sure you;d seen the thread. Thanks for confirming | 12:26.21 |
Robbes | btw, do you know if ghostscript does anything with JPEG2000 compression? | 12:27.11 |
kens2 | We decompress it, we do not apply it | 12:27.22 |
| JPEG2000 compression is patent encumebred, and cannot be released under GPL | 12:27.38 |
Robbes | oke | 12:27.43 |
kens2 | That is, the code can be releases, but you need a licence to use it..... | 12:28.15 |
chrisl | kens2: 9.06 exe seg faults...... so I would suggest whatever it was has been fixed | 12:28.32 |
kens2 | chrisl ah, that's interesting. I'll still give them some time to try it themselves and then close it if I don't hear. | 12:34.49 |
kens2 | coffees | 12:35.04 |
vtorri | tor7, hey | 13:07.53 |
henrys | holiday isn't it? | 13:19.46 |
kens2 | where is it holiday ? | 13:20.03 |
henrys | oh I though it was a bank holiday | 13:20.21 |
kens2 | Not here | 13:20.26 |
henrys | thought | 13:20.27 |
kens2 | There's a Scottish one today | 13:21.10 |
| Hmm, lots of 'ceonnection timed out' error son cluster | 13:22.15 |
henrys | I have a UK holiday calendar template on Google and it has it as a "UK" holiday | 13:23.17 |
kens2 | Hmm, well we don't take the UK ones any more than the US ones anyway | 13:23.39 |
| But I think its only in Scotland today anyway | 13:24.15 |
| Of course Scotland is still part of the UK | 13:25.06 |
henrys | hi marcosw I gave the 801 customer bug back to you but if you want me to talk to them let me know. | 15:11.06 |
| oops missed hm | 15:13.13 |
| mvrhel_laptop: just to verify my understanding if I want to see what will roundtrip I should use the output profile (the one with the extra blue plane in this case) as the input and srgb as the output. Do I have the right? Is there more to it? | 15:15.47 |
mvrhel_laptop | henrys: that is it | 15:17.39 |
henrys | also I'm booking chicago soon, did you decide? | 15:19.05 |
mvrhel_laptop | henrys: brb | 15:19.52 |
| henrys: sorry I misunderstood your earlier question | 15:32.38 |
| for roundtripping | 15:32.46 |
| but yes. you would first start with an RGB value. Do the conversion from sRGB to CMYK+B. Then take the CMYK+B value and do conversion back to sRGB | 15:35.06 |
henrys | yes that is basically what I did - there seems to be a constant factor pushing the values up in, there I assume that is the white point | 15:36.14 |
mvrhel_laptop | henrys: so with respect to chicago, I am fine either way. If you are want to do the whole show that is fine but I am equally fine with doing the tail end if you would like. It really is not a problem for me | 15:36.41 |
| henrys: try doing absolute colorimetry as the rendering intent | 15:37.02 |
| or colorimetric | 15:37.11 |
| there may be some funny business going on with the perceptual rendering intents | 15:37.45 |
henrys | mvrhel_laptop: Okay I'll do the entire show. | 15:38.15 |
mvrhel_laptop | henrys: ok. sounds good. I am going to change my ticket today then | 15:38.48 |
| henrys: I have to run my son off to computer camp | 15:39.19 |
henrys | mvrhel_laptop: okay that should be easy enough to fix in my script. | 15:39.21 |
mvrhel_laptop | and then daughter off to soccer | 15:39.33 |
henrys | okay | 15:39.39 |
mvrhel_laptop | I will be back in a bit | 15:39.44 |
henrys | marcosw:you around? | 15:49.09 |
| see the logs about 801 | 15:49.22 |
rayjj | peeves is back up. We had no power last night from a few minutes past 10 | 15:52.00 |
| seems to have come up quickly and clean | 15:52.16 |
henrys | rayjj:wow quite an outage for your area. | 15:52.47 |
marcosw | henrys: I am around and will look at the logs | 16:00.26 |
kens | Night all | 16:46.39 |
marcosw_ | henrys: I just re-read your comments on bug 694435 and I think it would be better if you emailed the customer directly. They are going to reply and that will require a follow-up discussion and I'll just end up forwarding emails between you and them. | 20:45.41 |
henrys | okay | 20:46.00 |
mvrhel_laptop | bbiaw | 22:01.25 |
| Forward 1 day (to 2013/08/06)>>> | |