| <<<Back 1 day (to 2014/04/27) | 2014/04/28 |
chrisl | 4 | 06:38.41 |
kens | chrisl did you see that comment in the logs about the cfax device ? | 07:06.39 |
chrisl | kens: I did, yes | 07:09.35 |
kens | I can't find it built into the Windows binary as far back as 8,71, is it in the Linux build ? | 07:09.54 |
chrisl | Yes, it's in the Linux build | 07:10.19 |
kens | Odd, I wonder why its not in the WIndows build | 07:10.29 |
chrisl | Perhaps I should grab all the devices in default Linux build, and put them in the Windows one - if appropriate | 07:11.42 |
kens | TO be honest I had ratehr assumed they were, obviously I'm wrong..... | 07:12.10 |
chrisl | No, we've had several cases of devices being in the configure builds, but not the Windows one. | 07:12.41 |
| I'll do that later today | 07:13.18 |
kens | How very strange, I can understand not having X devices, but other ones missing seems odd. | 07:13.19 |
| No rush, you've seen the message anyway | 07:13.34 |
chrisl | If the devices were implemented by a Unix user, and no one bother to really review the patches, then it's not that strange | 07:14.06 |
kens | Yes, I guess that's probably it. | 07:14.23 |
chrisl | Actually, I might see if I can make the device lists common to both | 07:15.27 |
kens | What about the X devices ? That's not going to work on Windows | 07:16.03 |
chrisl | Well, the Windows specific devices won't work on Unix, either..... | 07:16.41 |
kens | Yes, that's true too | 07:16.59 |
chrisl | So, I'd have to have "common devices" list, and platform specific devices list(s) | 07:17.37 |
kens | :-( | 07:17.49 |
| AH, looks like the SOT bug deluge is underway | 07:18.26 |
Robin_Watts | not reached here yet. | 07:20.37 |
kens | 4 bugs from Richard Chen | 07:20.53 |
| Or at least I assume they are 4 different bugs | 07:21.03 |
pedro_ | hi folks | 08:39.34 |
Robin_Watts | Morning pedro | 08:39.52 |
| pedro, paulgardiner_lap: So, I have a whole batch of commits on robin/master (sot). | 08:46.51 |
| They are all recommits of stuff that went on the release branch, but didn't make it down to master. | 08:47.17 |
| oh, jogux is here too. | 08:47.43 |
| We need to figure out what we plan to do with them. Do we push them to master now and test with ATS later, or do we wait for ATS? | 08:48.42 |
pedro | we could tag golden before and after the commits and ATS test from specific tags later I guess. | 08:56.14 |
Robin_Watts | pedro: Yeah, really we ought to run through and test every commit we've made since we got the source. | 08:57.09 |
paulgardiner_lap | Couldn't we just get ATS to test the effect of our sequence of commits in one go? Obviously we'd need to analyse further if a problem was uncovered, but if the diffs are all favourable... | 09:02.53 |
Robin_Watts | I find your faith disturbing... :) | 09:03.17 |
paulgardiner_lap | Yeah, what was I thinking?! | 09:04.40 |
Robin_Watts | I have to pop helen to the station. back in a bit. | 09:04.53 |
paulgardiner_lap | Robin_Watts: for the logs - re "Hacky workaround for bug #88674 - copying files with special characters fails", I wonder if we should verify that the problem is present on trunk in case the hack isn't needed. I'd imagin it is, but... | 09:15.15 |
| Maybe suffucient to open bugs to make sure we check at some stage and also don't forget these are temporary hacks | 09:17.36 |
Robin_Watts | paulgardiner_lap: I couldn't see anything on the trunk to address the problem. | 09:59.28 |
paulgardiner_lap | Robin_Watts: "Fixes for the Android library build under Windows" looks odd on master. On the branch, it was adding CPPFLAGS, giving the same value as was used for CFLAGS | 10:01.34 |
Robin_Watts | I will double check, thanks. | 10:02.07 |
| paulgardiner_lap: So, ARC builds... | 10:02.24 |
paulgardiner_lap | sorry, don't understand | 10:02.44 |
Robin_Watts | We've got a customer clamouring at us (well, you :) ) for an ARC lib build of the Picsel stuff ? | 10:03.16 |
| See mail to support on 27/4/14 at 14:35 | 10:03.52 |
| which is a followup to a mail on the 22nd. | 10:04.13 |
| Morning tor8 | 10:05.10 |
| tor8: A bug fix on robin/master | 10:05.20 |
tor8 | Robin_Watts: yes, LGTM | 10:05.26 |
Robin_Watts | Ta. | 10:05.29 |
| I wonder if I should rerelease for android cos of that :( | 10:05.45 |
| oh, no, it'll be fine on LE devices. | 10:05.54 |
paulgardiner_lap | Robin_Watts: I have no message for that time and date. My internet was out all of yesterday, although it should have gone to the secondary MTA | 10:07.32 |
| Can you forward it? | 10:07.42 |
Robin_Watts | I will. | 10:07.47 |
paulgardiner_lap | Robin_Watts: other than that, the commits look broadly sane | 10:09.27 |
Robin_Watts | Just testing them a bit now. | 10:11.14 |
| paulgardiner_lap: Did you get my forwarded email? | 10:18.19 |
pedro | Robin_Watts: commit 5937600 in those patches (Enable basic PDF annotations for PSO Enterprise) adds the PDF_ANNONTATIONS_FREEHAND which I think is now unsupported | 10:19.48 |
Robin_Watts | unsupported how? | 10:20.07 |
pedro | I think it has been removed from the allowed build options on master | 10:25.42 |
Robin_Watts | I'm just grepping now. I'll remove that commit unless I can find a compelling reason not to. | 10:26.25 |
paulgardiner_lap | Robin_Watts: I have the forwarded email. Strangely, I don't have the one to which it is a follow up. Do you? | 10:27.47 |
Robin_Watts | yes, I think so. | 10:28.01 |
| I have a mail from you on the 9th. | 10:28.57 |
| Then mails from them on the 22nd and 27th. | 10:29.10 |
| spam filter? | 10:29.17 |
paulgardiner_lap | What time on the 22nd? | 10:29.36 |
Robin_Watts | 04:30 | 10:29.56 |
| I am of a mind to remove the PDF_ENCRYPTION and PDF_ENCRYPTION_AES defines. | 10:33.01 |
| I am fairly sure that the PDF_ENCRYPTION_AES define was only put in place to avoid giving a customer something earlier than they had said they'd deliver it. | 10:33.36 |
| and I can't really see us wanting to support PDF without encryption. | 10:33.52 |
| I concur about the FREEHAND stuff. I can see places where it's set, but never where it's used. | 10:34.30 |
pedro | yup, it has been removed from scripts/core/options.conf | 10:35.08 |
| (the list of allowed defines) | 10:35.20 |
Robin_Watts | It's still in: release-configs/appstore-ios-so-good.txt:-define=PDF_ANNOTATIONS_FREEHAND | 10:36.16 |
| but I think that might be a commit I cherry-picked actually :) | 10:36.30 |
paulgardiner_lap | Robin_Watts: strange, mail logs so no attempt to deliver anything from *, or anything else around that exact time | 10:36.59 |
| Robin_Watts: anyway... problem now is that I have no idea whatsoever how one produces an ARC compatible library. jogux seemed to think it was not a huge job, but I don't really know what it involves | 10:38.09 |
| Robin_Watts: Oops, can you edit the logs. Sorry! | 10:39.34 |
pedro | Robin_Watts: think joseph has a commit on his master to remove that | 10:39.36 |
jogux | yes, I do. | 10:56.21 |
| although I for some reason failed to identify where that define got removed from options.conf. | 10:56.34 |
pedro | too - wonder if it was added on the branch rather than removed from the trunk | 10:57.18 |
jogux | ahh. that would make sense. | 11:00.52 |
| I wonder if SOG should actually be defining it then. | 11:01.18 |
Robin_Watts | Ah, so it was committed with an ifdef to the branch, and to master with no ifdef. | 11:01.34 |
| That would make sense. | 11:01.38 |
jogux | ah. great. :) | 11:01.41 |
Robin_Watts | That's just a theory, but unless I missed a commit it would make sense. | 11:02.26 |
| 27th came just to you/support | 11:02.55 |
| 22nd went to you/support/miles | 11:03.12 |
henrys | paulgardiner_lap: did you get your mail fixed? Just a temporary thing? | 11:56.33 |
henrys | wonders if marcosw should start fielding SOT support, but he might be at about his limit hour wise. | 11:58.03 |
paulgardiner_lap | henrys: I still haven't received a copy of the email I recently copied to support. It doesn't look to be my end because my MTA logs so no attempt to deliver, or didn't when I looked earlier at least | 12:00.01 |
| I wondered whether I'd lost my reverse DNS mapping, which can make some MTA refuse to deliver, but it seems okay. My internet was out all of yesterday, but I can't see anything wrong today. | 12:00.58 |
Robin_Watts | paulgardiner_lap: Did you get the mail from Will (Arabic issue) this morning? | 12:01.00 |
paulgardiner_lap | Yep. Got that | 12:01.20 |
Robin_Watts | From Richard (and my reply) about PPTX components. | 12:01.22 |
| From Marcos about Misc problems in HPGL? | 12:01.42 |
paulgardiner_lap | Got that too | 12:01.45 |
Robin_Watts | From Biro about gx_color_index ? | 12:01.51 |
| From rex about MuPDF layers? | 12:02.21 |
| That's everything I've got from 26th onwards. | 12:02.28 |
paulgardiner_lap | All those are here | 12:03.14 |
Robin_Watts | Well, the missing mails should have been sent to you twice, once directly, and once from support. | 12:06.12 |
| And given how other support emails are getting through, it seems odd. | 12:06.32 |
paulgardiner_lap | Just received your latest | 12:06.56 |
Robin_Watts | I have a bloke here fixing an electrical fault on the cooker, so if I drop off the net, you'll know it's the power tripped (Computer is on UPS, monitors and router are not) | 12:07.42 |
henrys | paulgardiner_lap: I can't see your email on gmail but I see you are using just under a gig for it, that's odd since I thought your email was just forwarded. | 12:11.29 |
paulgardiner_lap | henrys: interesting. Maybe I didn't set it up to delete after forwarding | 12:18.10 |
| ... if that makes sense | 12:18.18 |
| I'll take a look | 12:18.25 |
| I forgot that gmail foward was part of the path. | 12:18.44 |
| henrys: good call. Found the messages in the gmail spam folder | 12:26.03 |
henrys | paulgardiner_lap: you can mark them and tell gmail they are not spam, I guess you did that. | 12:32.20 |
paulgardiner_lap | yeah, just doing that now. | 12:33.35 |
| I wonder if it has white listing | 12:33.43 |
| henrys: an issue has come up on Skype where we could do with your advice when you have a moment | 12:34.58 |
henrys | paulgardiner_lap: okay give me a minute | 12:35.15 |
norbertj | henrys: do you know whether async-rendering (i.e. interpreting page 2, while rendering page1) is working? | 13:18.56 |
| henrys: I ask because we still have (on a few files) some performance degradation w.r.t. 9.06, and I was wondering whether async-rendering could be a solution. | 13:20.12 |
henrys | norbertj: as far as I know, but it is new, ray is the person to talk to | 13:20.15 |
norbertj | henrys: you know when ray will be online? | 13:20.45 |
henrys | norbertj: if they are regression in our code can you report them in a bug? | 13:20.59 |
| norbertj: should be here shortly | 13:21.13 |
Robin_Watts | norbertj: Are they text heavy files? | 13:45.10 |
| chrisl and ray found some speedups for text files last week, I think (making the glyph cache usage more persistent, I think). | 13:46.22 |
chrisl | Not for PCL | 13:46.31 |
Robin_Watts | oh, ok, sorry. | 13:46.48 |
chrisl | I guess I could look at doing a similar thing for PCL, though..... would have to think about that | 13:47.16 |
norbertj | Robin_Watts: if I look at the output it most text (with a bit of graphics). Kopia seems to position each char. | 13:47.42 |
henrys | chrisl: pcl shouldn't purge the cache per page does it? | 13:48.02 |
norbertj | Robin_Watts: I uploaded 3 testfiles on peeves. In my home directory 20140428-pclperf | 13:48.19 |
chrisl | henrys: no idea - that's not an area I've looked at in PCL | 13:48.38 |
henrys | norbertj: just create a bug and I'll get to it soon. | 13:49.05 |
| norbertj: I have 0 customer bugs so you're at the front of the queue ;-) | 13:49.39 |
chrisl | Seems likely to be text rendering, though :-( | 13:50.31 |
henrys | or I can create the bug - if you tell me what release is the baseline for the regression | 13:50.40 |
norbertj | henrys: see 695190 | 13:54.01 |
henrys | thanks norbertj | 13:54.34 |
Ziai | do u know if jhabjan is comming any time soon? :) | 14:01.24 |
henrys | chrisl: well let's not have both of us working on it. Shall I look first? | 14:02.20 |
chrisl | henrys: if you are free, then I'd appreciate it - I'm kind of in the middle of something | 14:02.47 |
henrys | chrisl: that's fine | 14:02.57 |
norbertj | henrys: beware of cmp_02_08 (it's 24718 pages), Arrowhead (3580 pages), Kopia (1368 pages). Could be that there is a xml-ticket (text page) in front. | 14:12.33 |
Adam___ | Hi! Anybody can help me with command line options, please? | 15:14.38 |
chrisl | Adam___: you might as well just ask - specify for what (gs or mupdf)...... | 15:16.52 |
Adam___ | gs converting pdf to tiff file, it generally works but i can not convert from color space to antoher space using icc profiles | 15:18.12 |
kens | sounds like a question for mvrhel_laptop | 15:19.37 |
chrisl | Adam___: you mean you want to use a specific icc profile? | 15:20.03 |
Adam___ | yes, i want to assign my icc profile for source pdf and covvert it to another icc profile using absolute colorimetric intenet for proofing purposes, all objects are CMYK | 15:21.20 |
| target file should be TIFF | 15:21.44 |
chrisl | Adam___: I think what you need is covered in this: http://www.ghostscript.com/doc/9.14/GS9_Color_Management.pdf | 15:24.11 |
| Specifically from page 5 | 15:25.08 |
Adam___ | yes, i have read this document and tried many switches, but it seems to be not working, all the time i get good TIFF file but without conversion, all colors are the same like in source PDF file | 15:25.35 |
kens | This sounds too nebulous to me. I thinkyou shoudl report a bug, attach the source file and give the command line you are using. Then describe carefully where on the output TIFF you see a problem, and why you think the colour is incorrect. | 15:27.14 |
Adam___ | moreover all the time gs is embedding "Artifex CMYK SWOP Profile" | 15:27.23 |
chrisl | That suggests it isn't picking up your preferred output profile | 15:28.34 |
Adam___ | Yes it looks like this, but ia have no error message | 15:29.10 |
kens | If you (for example) mistype the switch, you waon't get an error | 15:29.42 |
Adam___ | I tried this, typing wrong profile name and no error occured | 15:30.19 |
chrisl | Hmm, there shouldn't be spaces in the command line examples - that's not great..... | 15:30.19 |
Adam___ | Do you think that path to Windows icc profiles folder can be a problem? "C:\Windows\System32\spool\drivers\color" | 15:31.26 |
kens | Rather than randomly poking stuff, report a bug | 15:31.55 |
chrisl | Ghostscript won't look there by default | 15:32.02 |
Adam___ | This is my command line: | 15:36.02 |
| C:\gs\gs914\bin\gswin64c.exe -dBATCH -dNOPAUSE -sDEVICE=tiff32nc -r300 -dTextAlphaBits=4 -dGraphicsAlphaBits=4 -sOutputFile=%1.tif %1 -sICCProfilesDir=C:\Windows\System32\spool\drivers\color -dOverrideICC=true -sDefaultCMYKProfile=ISOcoated_v2_eci.icc -sOutputICCProfile=PSO_LWC_Improved_eci.icc | 15:36.04 |
| I have set path to icc profiles folder | 15:36.18 |
chrisl | Adam___: maybe try using absolute paths to the profile files, rather than using -sICCProfilesDir ? | 15:36.55 |
Adam___ | ok, i will try | 15:37.18 |
chrisl | I *do* get an error if the profile isn't found..... | 15:38.33 |
| Well, actually, a warning | 15:38.45 |
Adam___ | Oh! It is interesting, because i have no even warning | 15:40.58 |
chrisl | It's possible it's because I'm running a debug build | 15:41.22 |
Adam___ | Can i switch on debug mode? | 15:41.38 |
chrisl | If you rebuild gs from source...... but no, it happens in a release build, too | 15:42.00 |
Adam___ | I can't rebuild, sorry :) | 15:42.24 |
chrisl | I tried using the full path and file name for "-sOutputICCProfile=....." and it did change the device profile - I won't promise it used the one I specified, but..... | 15:44.10 |
Adam___ | Does name of the profile should be exactly name of ICC profile file with extension? | 15:45.32 |
chrisl | Yes | 15:45.52 |
Adam___ | Do you use Windows wersion? | 15:46.32 |
chrisl | No, I'm running linux | 15:46.40 |
Adam___ | May it make a difference? | 15:46.56 |
chrisl | It's feasible | 15:47.05 |
| Adam___: what version of gs are you using? | 15:47.27 |
Adam___ | today i downloaded 9.14 | 15:47.57 |
chrisl | Okay, well, that's the extent of my knowledge on this subject exhausted - as kens said, please open a bug, and that will get to the right engineer to look at it | 15:48.44 |
Adam___ | thank you very much for help | 15:49.10 |
chrisl | Drat, I just noticed something..... and he's gone | 15:52.42 |
kens | well if he opens a bug report you can put it there | 15:53.01 |
chrisl | He was specifying the ICC options *after* the input file - I bet that doesn't work | 15:53.20 |
kens | No, it won''t | 15:53.43 |
chrisl | So, I bet that's the problem.... | 15:53.56 |
kens | All options (usually) have to precede the input file | 15:53.57 |
| I couldn't even see the input file i his command line, too complicated | 15:54.35 |
chrisl | It's from a batch file - the "%1" was the input file | 15:55.02 |
kens | OK I see it now | 15:55.29 |
| Which is why I wish people woudn't quote the command line from a batch file | 15:55.59 |
chrisl | Yeh, me too | 15:56.17 |
| Aw crap, I'm late - byte | 15:56.57 |
kens | Goodnight all | 16:08.36 |
ray_laptop | mvrhel_laptop: I have a transparency question from cust 532. Can you chat for a minute ? | 17:57.08 |
mvrhel_laptop | sure | 17:57.15 |
ray_laptop | mvrhel_laptop: in the HEAD code it's in gdevp14.c line 4756+ | 17:57.35 |
mvrhel_laptop | ray_laptop: ok hold on a sec let me get gs opened and get there | 17:57.55 |
ray_laptop | mvrhel_laptop: thanks | 17:58.04 |
| they are getting the "wrong" tag, and the hack they put in was that if alphag_off was 0, to just set dst_ptr[tag_off] = curr_tag | 17:59.20 |
mvrhel_laptop | ok I am there | 17:59.40 |
ray_laptop | mvrhel_laptop: I think the problem is that when we "blend" the tags (OR them), the tag doesn't come out the way they want. | 17:59.59 |
| mvrhel_laptop: the thing I've found is that the only time the OR is happening, the src_alpha == 0 | 18:00.33 |
| mvrhel_laptop: and that seems to be something we could use, such as: if (dst[num_comp] == 255 || src_alpha == 0) { | 18:01.30 |
mvrhel_laptop | ray_laptop: right I agree. | 18:04.10 |
| we should also have the case with src_alpha == 0 | 18:04.26 |
| in which case we maintain the current tag | 18:04.35 |
| which should be equivalent to their hack | 18:05.33 |
| this needs to be done in two spots | 18:05.43 |
| someone added a num_comp additive case I see | 18:06.29 |
| had to be a brit from the spelling of the word optimise in the comment.... | 18:06.49 |
ray_laptop | mvrhel_laptop: sorry. phone call. reading to catch up | 18:07.29 |
| mvrhel_laptop: OK. so if src_alpha == 0, then we want to maintain GS_UNTOUCHED ? | 18:08.30 |
mvrhel_laptop | hmm good question. | 18:08.52 |
| I would say yes | 18:08.58 |
ray_laptop | thinks so, but wanted confirmation | 18:09.11 |
| so, probably above that change: if (tag_off) { to: if (tag_off && src_alpha > 0) { | 18:09.58 |
| the customer's idea of making it conditional on alphag_off didn't make ANY sense to me | 18:10.50 |
ray_laptop | is fuzzy on alphag and shapeg and how they should relate to tags (if at all) | 18:13.10 |
mvrhel_laptop | ray_laptop: yes f (tag_off && src_alpha > 0) makes sense | 18:13.23 |
ray_laptop | mvrhel_laptop: thanks | 18:13.34 |
mvrhel_laptop | ray_laptop: alphag and shapeg are very confusing. everytime I get involved in that one, there at least a day start up to review | 18:14.07 |
| I was hoping your question was not about that | 18:14.16 |
ray_laptop | mvrhel_laptop: no, I think we'll just ignore that for now (until the come up with a file that that above approach doesn't fix) | 18:15.49 |
mvrhel_laptop | ray_laptop: sounds good | 18:16.14 |
ray_laptop | mvrhel_laptop: thanks for the consultation. I sent the description and patch off to cust 532 | 18:39.14 |
mvrhel_laptop | ok. np | 18:39.24 |
ray_laptop | two problems for cust 532 addressed that came in Friday. Pretty good response, if I do say so myself (IMNSHO) :-) | 18:40.48 |
mvrhel_laptop | bbiaw lunch | 18:54.51 |
Robin_Watts | mvrhel_laptop: probably me. | 18:58.09 |
| Forward 1 day (to 2014/04/29)>>> | |