| <<<Back 1 day (to 2015/08/24) | 20150825 |
tor8 | sebras: thanks. I'll update ucdn then. | 09:21.45 |
chrisl | tor8: there was a new font release from URW with more Greek and Cyrillic updates a couple of weeks ago - you'll probably want them for mupdf at some point. In my home directory on casper, file's called: Artifex_Update_Core12_2015_08_11.zip | 09:25.40 |
tor8 | chrisl: Thanks. Any idea what's changed? | 09:33.06 |
| chrisl: in NimbusRomNo9LReg, sfddiff lists changed advance widths for mu and Delta and twenty or so more glyphs | 09:36.21 |
chrisl | tor8: there were metrics fixes in there, too | 09:38.54 |
henrys | Robin_Watts: where is the preliminary work you'd done with the language API? I'm not seeing it in your repos. | 13:30.03 |
chrisl | henrys: http://git.ghostscript.com/?p=user/robin/ghostpdl.git;a=shortlog;h=refs/heads/language | 13:33.46 |
henrys | thanks | 13:35.19 |
| I'm getting sporadic gateway timeouts on casper. Hope it isn't going to fall over again | 13:37.58 |
jogux | still has a suspicion someone is trying to attack the twiki | 13:43.11 |
henrys | the non pdf ps languages were never designed to work with this API. They were designed to work with the graphics libray API which is not well documented but there is a sample client in base/gslib.c | 13:43.16 |
| if we are going to give up on that old api we should probably delete gslib.c | 13:43.52 |
jogux | 227,000 hits on the twiki so far this month, that can't be "normal". | 13:44.20 |
henrys | jogux: there hasn't been much sharing of the twiki outside the company right? | 13:45.20 |
jogux | I think it was in one of the newsletters for mupdfvsghostscript, but otherwise nothing afaicr. | 13:45.39 |
henrys | jogux: specifically pointing folks at it? | 13:45.40 |
chrisl | henrys: this isn't about an API for the graphics library, this is about an equivalent of the gaspi stuff that can be consistent across all the interpreters | 13:46.30 |
henrys | chrisl: I'm sort of defending myself after yesterday's attacks ;-) I didn't mean to abuse the so called "API" | 13:47.25 |
chrisl | henrys: well, the problem is that the graphics lib "API" has always been private, and thus has evolved, not always in the best ways. And certain things are done ways that make for a less than ideal API, but are required for decent performance | 13:48.57 |
henrys | jogux: I'll ask marcosw to have a look he has root on casper. | 13:49.27 |
jogux | http://twiki.ghostscript.com/do/view/TWiki/TWikiDocGraphics has been accessed about 100,000 times from a pool of about 10 IPs in two ranges... wtf. | 13:51.44 |
| (that's just this month...)_ | 13:51.51 |
chrisl | Hmm, am I mistaken, or is the twiki on an http link, rather than https..... | 13:53.07 |
Robin_Watts | henrys: Yesterdays comments were not intended as attacks. | 13:55.34 |
jogux | chrisl: you're correct afaik. | 13:55.55 |
henrys | I'm a sensitive guy you know ;-) | 13:56.10 |
chrisl | henrys: it is a problem that the graphics library was conceived and has evolved mainly to work with Ghostscript, so it generally expects stuff to happen in a "Postscripty" kind of way. | 13:57.50 |
Robin_Watts | One of the earliest quotes I remember from henrys when I joined was "positive reinforcement is for pussies" (or something like that) | 13:58.11 |
henrys | so the best way forward is to get Robin_Watts work merged in with chrisl new layout then I'll pull and fool with PCL and XPS. | 13:58.20 |
| ;-) | 13:58.25 |
chrisl | Er, I think I have a branch with Robin_Watts stuff brought up to date...... let me see | 13:58.54 |
Robin_Watts | yeah, I haven't touched my stuff for ages. I was told to hand it off to chris. | 13:59.18 |
chrisl | Okay, it's not fully up to date, but here it is "post build changes": http://git.ghostscript.com/?p=user/chrisl/ghostpdl.git;a=shortlog;h=refs/heads/language_switch | 14:00.08 |
| I'll see about get it fully up to date either later today or tomorrow | 14:00.34 |
henrys | thanks if I suspect it won't matter it's up to date the pcl problems should be obvious so I can get started after the meeting | 14:01.45 |
chrisl | henrys: the thing I keep stumbling over is that pl_main_aux() is going to be a huge PITA to split into pieces :-( | 14:02.03 |
kens | had completely forgotten it was Tuesday | 14:02.09 |
Robin_Watts | henrys: I suspect pcl just won't build :) | 14:02.10 |
| There is supposed to be a gx_language defined for pcl with the appropriate entrypoints, but that's not there yet. | 14:03.03 |
henrys | wow can barely see the mountains the last few days and air quality alerts the west is really burning... | 14:06.03 |
jogux | henrys: doesn't sound good... | 14:09.51 |
henrys | chrisl: yes that needs to be completely reconsidered if we are going to do this. | 14:10.18 |
chrisl | henrys: we don't have a choice if we want a vaguely workable language_switch type product | 14:11.00 |
Robin_Watts | The shape of gx_language is probably not ideal. | 14:11.06 |
| It wasn't designed, so much as frankensteined out of the bits I had to hand. | 14:11.28 |
| So it shouldn't be considered cast in stone if we can make changes to it to simplify life. | 14:12.02 |
chrisl | I think it's incomplete, but not bad. We really want something that won't scare away existing gsapi users | 14:12.05 |
henrys | jogux: it is part of the natural process here but it's amazing how widespread it is, this smoke is coming from Idaho and Montana | 14:12.42 |
Robin_Watts | chrisl: I think gsapi remains implemented on top of it. | 14:13.56 |
| I tried to keep gsmain_.... kinda working too, but I think that may not have worked out. We're into crappy memory territory here now :( | 14:14.36 |
kens | fetches coffee, brb | 14:15.22 |
chrisl | Robin_Watts: I'd prefer to keep a clear familial resemblance to gsapi so anyone used to gsapi can easily "move up" to gslang | 14:15.40 |
Robin_Watts | chrisl: yeah. | 14:15.49 |
chrisl | Consistency and all that stuff we usually don't bother with! | 14:16.10 |
Robin_Watts | but if we can keep gsapi implemented on top of it, I would not be averse to bending gx_language towards something that works out easier to implement everywhere. | 14:16.39 |
| The real killer is trying to keep the command line processing equivalent to what we have now. | 14:17.01 |
chrisl | Sure, I'm not suggesting that we strictly and only ape gsapi | 14:17.29 |
| command line processing will never be "nice" because of Ghostscript's special stuff | 14:18.05 |
| I guess we could have a single command line parser with a series of callbacks defined by each interpreter: e.g. set_string_opt(), set_bool_opt(), set_int_opt(). But we'd still hit problems with Ghostscript's "-c" stuff | 14:20.28 |
| henrys: my language_switch branch is now up to date with master - everything except gpdl builds, but I make no claims about them functioning...... | 14:22.29 |
henrys | chrisl: tiger at least? | 14:23.05 |
| for gs? | 14:23.10 |
Robin_Watts | chrisl: the problem is we don't know what interpreter to set the options in, until we get to the file to run. | 14:23.24 |
chrisl | henrys: haven't tried..... | 14:23.51 |
henrys | chrisl: why can't each interpreter just ignore what isn't relevent. Similarly PCL process PJL on the command line. I suppose PDF/PS would ignore that. | 14:24.41 |
chrisl | Robin_Watts: true - we're stuffed then | 14:25.02 |
| henrys: that means always initialising every interpreter - which we specifically wanted to be able to avoid | 14:25.27 |
| henrys: Even tiger segfault gs on the branch...... I'll look at why | 14:26.16 |
Robin_Watts | chrisl: One option is to try to cope with as much as possible at the top level. When we meet -d, -s etc, set those in an interpreter-independent structure. Then when we know which interpreter we are using, pass that independent list of options in, and have the interpreter convert them to its own format. | 14:27.12 |
henrys | without michael and ray I thought it could be a shortened meeting. | 14:28.18 |
chrisl | Robin_Watts: Hmm, two complete passes - and that won't work on the rare occasions we have options that must be in a specific order | 14:28.25 |
Robin_Watts | chrisl: No, we'd still go option by option. | 14:28.45 |
| but options would effectively be buffered up and flushed just before we run a file or command. | 14:29.21 |
| And ordering can be preserved in that buffering if we are careful. | 14:29.44 |
henrys | fredross-perry: how is the demo coming along, we are getting close to Chicago? | 14:29.55 |
marcosw | morning all | 14:30.09 |
fredross-perry | android is working well, yes. | 14:30.10 |
chrisl | Erm, Robin_Watts, did you run any files through gs with your language branch? | 14:30.22 |
Robin_Watts | chrisl: Quite possibly not. | 14:30.32 |
henrys | fredross-perry: I can download it fine. I just want to know if there are big problems I should be aware of ... | 14:30.36 |
| hi marcosw | 14:31.12 |
Robin_Watts | henrys: I think the only thing missing at the moment is that we want to preserve what page is selected when we start proofing. | 14:31.29 |
fredross-perry | Not that Iâve seen. Itâd be good to know (a) what files you might use in a demo, and (b) if you hate the icons I made. | 14:31.38 |
henrys | marcosw: jogux thinks the twiki might be under attack on casper. Can you look at that? | 14:31.42 |
Robin_Watts | fredross-perry: Your icons look good to me. | 14:32.05 |
henrys | fredross-perry: all of the altona stuff is huge at the show. It all needs to be right... | 14:32.18 |
marcosw | henrys: yes. saw the earlier discussion here. | 14:32.32 |
fredross-perry | Where can I find âthe altona stuffâ? | 14:32.56 |
henrys | ~/tests_private/comparefiles/[Aa]* | 14:33.23 |
fredross-perry | and, do you need any help using the demo? Itâs pretty simple. | 14:33.23 |
| thanks. | 14:33.30 |
henrys | fredross-perry: I don't think so | 14:33.33 |
fredross-perry | swell | 14:34.03 |
henrys | oh you don't have an account on casper? | 14:34.27 |
| fredross-perry? | 14:34.38 |
Robin_Watts | henrys: he does. | 14:34.42 |
fredross-perry | yes | 14:34.42 |
| thanks | 14:34.52 |
henrys | so ssh and locate is a great way to find files on casper... fwiw | 14:35.27 |
Robin_Watts | henrys: No it's not :) | 14:35.40 |
| The testfiles aren't unpacked on casper. | 14:35.47 |
| Cos casper isn't a cluster node. | 14:36.00 |
fredross-perry | where then? | 14:36.02 |
henrys | snag them out of /home/marcosw | 14:36.04 |
Robin_Watts | peeves | 14:36.06 |
fredross-perry | ok iâll look there. | 14:36.14 |
Robin_Watts | tests_private/comparefiles/Altona-Testsuite_p2_S_x3.pdf | 14:36.35 |
| tests_private/comparefiles/Altona.Page_3.2002-09-27.pdf | 14:36.37 |
| tests_private/comparefiles/Altona_Measure_1v1.pdf | 14:36.39 |
| tests_private/comparefiles/Altona_Technical_1v1_x3.pdf | 14:36.40 |
| tests_private/comparefiles/Altona_Visual_bb_1v1_x3.pdf | 14:36.42 |
| tests_private/comparefiles/Altona_Visual_sb_1v1_x3.pdf | 14:36.44 |
henrys | there all on casper too | 14:36.49 |
| s/there/they are/ | 14:37.00 |
Robin_Watts | Those are the Altona files, I believe. You can download them from here: http://ghostscript.com/regression/cgi-bin/svnfetch.cgi | 14:37.03 |
henrys | tor8: wanted to touch base about XFA? | 14:37.31 |
Robin_Watts | henrys: hmm. They never used to be, but yes, I can see these ones in /home/marcosw/cluster_svn/ | 14:37.50 |
henrys | marcosw: what are these "lower case" altona's: /home/marcosw/cluster/tests_private/comparefiles/altona_measure_1v1a.pdf? | 14:38.48 |
| I have henrysx6 locally so I just grab stuff from that most of the time. | 14:39.22 |
| kens: stuff for the meeting? | 14:40.16 |
kens | Not really I don't htink. I'm busy trying to fix some graphics state stuff in PDF, and its going to take a few dyas yet. | 14:40.48 |
Robin_Watts | I take it back. I can't see them on casper. I was looking at peeves :( | 14:40.53 |
tor8 | henrys: getting there, slowly. I spent the past week getting the opengl-based viewer (intended to replace the win32/x11 sample viewer) working on windows. | 14:41.16 |
marcosw | henrys: those are different versions of the Altona test files. I can't recall why we have both versions. | 14:41.17 |
kens | By the waqy, I htnk we have the 1.2 Altona test suite files, not the 2.0 ones. | 14:41.22 |
tor8 | next up --getting it to do forms, and then I can see about XFA | 14:41.30 |
mvrhel_laptop | henrys: just one thing about the gproof code. right now the way it is going it is not color managed properly. I should have that debugged this week and then we just need to add one additional option in mupdf's call to gs in creating the pages | 14:41.32 |
| I had wanted to get it done before I left but ran out of time | 14:41.56 |
Robin_Watts | mvrhel_laptop: So there are changes required in gs? Or in mupdf? Or both? | 14:42.16 |
mvrhel_laptop | both | 14:42.22 |
henrys | mvrhel_laptop: so altona's are going to be a big problem with what we have now. | 14:42.28 |
marcosw | looks like the lower case a versions were added in 2008, the older versions predate the repository. | 14:42.39 |
Robin_Watts | Can you give us a rundown of what's missing? Is it stuff that we can do in your absence? | 14:42.59 |
mvrhel_laptop | someone familiar with the file will think the colors look off | 14:43.17 |
henrys | marcosw: I recall version info on the output | 14:43.40 |
paulgardiner | henrys: bug 696123 is proving harder to fix than I'd envisaged. Getting somewhere with it, but not done yet. Has left me no time for NUI | 14:43.52 |
mvrhel_laptop | Robin_Watts: yes. in the mupdf code you need to add -sPostRenderProfile=srgb.icc to the gs call | 14:44.20 |
kens | The Altons files seem to be available here (v1.2 and v2.0): | 14:44.20 |
| http://www.eci.org/en/downloads | 14:44.20 |
mvrhel_laptop | that will result in the icc code being used | 14:44.35 |
| however, you will get garbage now if you do that | 14:44.45 |
henrys | paulgardiner: darn | 14:44.46 |
Robin_Watts | mvrhel_laptop: OK. That's the only mupdf change? | 14:44.51 |
mvrhel_laptop | ys | 14:44.56 |
Robin_Watts | OK, so I'm not worried about our ability to get that done in time :) | 14:45.12 |
mvrhel_laptop | no that is the easy one :) | 14:45.27 |
| but lets wait on it | 14:45.34 |
henrys | paulgardiner: does it make sense to split your time or do you think that would be a lot less productive than focusing on one? | 14:45.41 |
mvrhel_laptop | until I get gdevgrpf.c working | 14:45.43 |
| gdevgprf.c | 14:45.53 |
paulgardiner | I'm hopeless when I'm swapping | 14:45.59 |
Robin_Watts | Yes. Is there any prep work we can do in gdevprf.c to make it easier for you when you get back ? | 14:46.00 |
mvrhel_laptop | with that option | 14:46.01 |
paulgardiner | I may be close, but it's difficult to be sure | 14:46.22 |
mvrhel_laptop | Robin_Watts: I dont think so, it should not take me too long to figure it out | 14:46.26 |
| I am going to work on it this week | 14:46.46 |
| possibly today | 14:46.51 |
henrys | paulgardiner: let's talk more at the skype meeting and see what's best to do next. I can talk to the customer if need be. | 14:46.51 |
mvrhel_laptop | everyone is still sleeping so I may fool with it now | 14:47.09 |
henrys | sleeping I thought you were on the east coast? | 14:47.44 |
| marcosw: can you add the latest altona's from kens link? | 14:49.22 |
fredross-perry | let me know if/when thereâs something I should use to update android. | 14:49.37 |
mvrhel_laptop | no west coast | 14:50.00 |
| At the beach in Oregon | 14:50.10 |
| much easier trip | 14:50.25 |
| but water is colder | 14:50.46 |
henrys | mvrhel_laptop: clear of the fires? I can't see a damn thing here and our air quality is in the "danger zone" | 14:50.46 |
mvrhel_laptop | it was bad back home | 14:50.53 |
| here it is clear | 14:50.57 |
| stephanie's race in september is located right where the wildfires are in Washington | 14:51.18 |
jogux | henrys: for my 2 cents, I'd probably say Paul should stay focussed, get it done, then make a clean break back to nui, unless it's going to be weeks... | 14:51.22 |
mvrhel_laptop | it is likely it will be cancelled | 14:51.24 |
henrys | mvrhel_laptop: yeah sometimes they don't cancel though and I think they should. | 14:52.07 |
Robin_Watts | I'm with jogux. | 14:52.12 |
mvrhel_laptop | henrys: that is what she is worried about. apparently they are gong to make a decision in a bit over a week | 14:52.54 |
henrys | jogux, Robin_Watts if that's okay with paulgardiner it's fine by me. | 14:53.03 |
| mvrhel_laptop: plenty of other races ... | 14:53.26 |
paulgardiner | Suits me. If it drags on, we can revisit the decision | 14:53.28 |
henrys | marcosw: any customer issues for the meeting? | 14:54.16 |
| that's all I had for the meeting. | 14:56.04 |
fredross-perry | I grabbed the Altona files. The first one I tried, Altona_Measure_1v1.pdf, never finished proofing on my nexus 7. | 14:59.02 |
| at 300 dpi. | 14:59.21 |
kens | The media is quite large for those files, and they are very complex | 14:59.37 |
| http://www.eci.org/en/downloadsMight require a lot of memory | 14:59.44 |
fredross-perry | I lied. The one I tried was Altona-Testsuite_p2_S_x3.pdf. | 15:01.01 |
| I also tried Altona.Page_3.2002-09-27.pdf at 72 dpi. Much better. | 15:01.24 |
mvrhel_laptop | ok the icc code is working now | 15:01.30 |
| let me clean up some spacing stuff and put in on my repos for review | 15:01.46 |
fredross-perry | itâs got cmyk, plus âMyRedâ and âMyBlueâ, both of which appear to have nothing in them. | 15:01.58 |
mvrhel_laptop | actually let me check the altona file first | 15:02.04 |
| to make sure that looks ok | 15:02.09 |
kens | fredross-perry : MyRed and MyBlue shopuld not be empty | 15:03.20 |
fredross-perry | When I do Altona-Testsuite_p2_S_x3.pdf at 72dpi, that works fine. | 15:03.24 |
| ok | 15:03.35 |
marcosw | henrys: no, you saw the email regarding pcl->pdf->pdf/a | 15:03.48 |
kens | Do you have access to Acrobat Pro ? | 15:03.51 |
fredross-perry | Altona.Page_3.2002-09-27.pdf appears to have âOrangeâ, which also does not appear. | 15:04.17 |
kens | Let me check that oen | 15:04.27 |
fredross-perry | I donât have Acrobat Pro, no. | 15:04.27 |
kens | Oh, that's unfortunate. | 15:04.36 |
henrys | marcosw: I did, he didn't specifically say he wanted us to write a script, do you think we should? | 15:04.44 |
fredross-perry | maybe mvrhel_laptop can see if his gsview has the same results. | 15:05.01 |
mvrhel_laptop | which file | 15:05.10 |
| it is possible there is an issue with spots | 15:05.20 |
marcosw | henrys: I told him we would in my email asking him if that was an acceptable solution (ken or you had suggested that in the internal email thread). | 15:05.32 |
kens | For Altona.Page_3.2002-09-27.pdf the red/blue lines behind each of the tests are in MyRed and MyBlue | 15:05.57 |
| In the p2_S_x3.pdf file the picture at top right should be in Orabge | 15:06.45 |
| The one labelled 'Black + Spot color "Orange" | 15:07.00 |
mvrhel_laptop | hold on | 15:07.10 |
marcosw | jogux found the issue with the twiki 'attack'. it was a web crawler, so not apparently not an intentional dos attack | 15:07.15 |
jogux | marcosw: bet that's what took out casper twice though :( | 15:07.57 |
fredross-perry | it looks to me like MyRed and MyBlue are getting rendered into the other four colors. When I switch off Magenta, for example, all the MyRed stuff disappears. | 15:08.47 |
mvrhel_laptop | let me look at Altona_Visual_bb_1v1_x3.pdf | 15:09.11 |
kens | I imagien that MyRed has a alternate space of DeviceCMYK and mostly consists of magenta | 15:09.17 |
| mvrhel_laptop : that one also has the Orange spot colour (and overprintign) in the top right image | 15:09.57 |
mvrhel_laptop | yes | 15:10.04 |
marcosw | jogux: yes, their robot seems overly aggressive. | 15:10.10 |
kens | The crop and register marks in these files are in /All and so should also appear on every plate | 15:10.41 |
fredross-perry | In the p2_S_x3.pdf file the picture at top right is shown as grey (well, sepia), even before proofing. | 15:10.47 |
marcosw | jogux: btw, we are still getting hits from ahrefs.com, did you add an entry to the robots.txt file? | 15:10.49 |
mvrhel_laptop | looks good on gsview | 15:11.20 |
kens | fredross_perry it should be orange'ish | 15:11.23 |
mvrhel_laptop | this is with the icc flow though | 15:11.26 |
fredross-perry | kens: when you say âred/blue lines behind each of the testsâ, are you talking about lines, or blocks? | 15:11.27 |
mvrhel_laptop | has overprint and the orange | 15:11.33 |
| let me do the non icc flow | 15:11.46 |
fredross-perry | kens: itâs really very sepia. And not what I see in Preview. | 15:11.47 |
kens | THey are blocks of colour, whch are composed of vertical lines ofr MyRed and MyBlue | 15:11.48 |
mvrhel_laptop | hlod on | 15:11.50 |
fredross-perry | hloding. | 15:11.59 |
mvrhel_laptop | ha | 15:12.06 |
| natives are up... | 15:12.14 |
kens | But are they restless ? | 15:12.29 |
| fredross-perry if you look at (for example) test A,9 Its a gray capital T on a background. The background is alternating vertical stripes of blue'ish and redd'ish colour, those are MyRed and MyBlue | 15:13.48 |
| In fact those blockunderlie *all* the tests, but some of the tests are opaque and cover the background up | 15:14.15 |
jogux | marcosw: yeah, I setup robots.txt, it may take a few days before they notice it | 15:16.16 |
chrisl | You should be able to see what the separations should look like using Ghostscript and tiffsep...... | 15:16.30 |
fredross-perry | kens: yes I see. When proofed, I see the grey capital T and the underlying stripes. When I turn off Magenta, all the stripes behind the capital T in A9 disappear. | 15:16.43 |
kens | OK then I suspect they are being rendered in the alternate space, not the spot colour | 15:17.10 |
mvrhel_laptop | ok so the orange is there and so is the overprinting in the non icc case. colors look way to saturated of course | 15:19.02 |
| so I am going to get these things cleaned up and pushed on my repos | 15:19.45 |
marcosw | should I block the ip addresses? | 15:22.01 |
jogux | marcosw: feel free :) | 15:22.23 |
| marcosw: I was tempted to suggest doing so | 15:22.32 |
Robin_Watts | henrys, marcosw: One thing that cropped up last week... we had a bloke register with bugzilla and start to spam it. | 15:23.17 |
marcosw | it's odd their robot is so broken, they seem like a real company (they have offices in San Francisco, maybe I'll apply for a job :-) ) | 15:23.42 |
Robin_Watts | There is no one outside the PDT/MST timezones with the ability to block a bugzilla user. | 15:23.52 |
marcosw | Robin_Watts: would you like admin access on bugzilla? | 15:24.11 |
Robin_Watts | marcosw: I wouldn't object. But chrisl or kens would be equally worthy candidates. | 15:24.32 |
kens | Anyone will do. In ths case it was fine, but he could have spammed a load of bugs before we could stop him | 15:24.56 |
chrisl | I'd like two off us, at least, so we have coverage during holidays | 15:25.09 |
marcosw | henrys: is it okay if I give all staff admin access on bugzilla? | 15:25.34 |
henrys | marcosw: it wasn't easy to delete so I passed it off to you ;-) something about fouling up the database. So if you do give this to someone can you give them some instructions? | 15:25.38 |
jogux | it's easy to disable users at least | 15:26.06 |
Robin_Watts | henrys: You can never delete users from bugzilla. | 15:26.12 |
| All you can do is to disable them, and I had the admin permissions to see his account, but I couldn't disable it. | 15:26.36 |
marcosw | henrys: yeah, you can't delete users. just put text in the "Disable text" field. | 15:26.38 |
| "If non-empty, then it will not be possible to log in using this account, and this text should explain why." | 15:26.50 |
Robin_Watts | marcosw: I couldn't see any Disable Text field. | 15:27.05 |
| (and I did look!) | 15:27.12 |
marcosw | maybe because you don't have admin priv? | 15:27.34 |
Robin_Watts | marcosw: That would explain it :) | 15:27.41 |
marcosw | btw, i'm pretty sure that deleting the user would have been okay in this case. it would have made all of their bugs unsearchable, but that would have been a good thing. | 15:28.26 |
henrys | marcosw: okay I wonder why bugzilla even has the option to delete users? | 15:28.31 |
Robin_Watts | henrys: Does it? | 15:28.44 |
marcosw | if a user is created and doesn't create or edit any bugs they can be safely deleted. not sure how useful that is. | 15:29.03 |
henrys | marcosw: if this spamming becomes a serious problem we are going to have to figure out how to get rid of folks I would think... | 15:30.18 |
marcosw | i've added 'editusers' to the permissions of the artifex staff. | 15:33.11 |
| henrys: as I said, we can delete users if we don't mind the bugs and comments they entered disappearing as well. which I suppose would be the case for users that are spamming. | 15:39.11 |
kens | In ths case I'd have been happy for his bug to vanish too | 15:39.45 |
marcosw | jogux: we are now dropping all packets from the ahref robot ip addresses. this does have the problem that they won't be able to access robots.txt... | 15:40.55 |
henrys | marcosw: I thought that was the case but the error messages you get when trying to delete a user sound like their might be other effects, but probably not. | 15:40.59 |
jogux | marcosw: cool, thanks. I don't think we care they can't access it. :-) | 15:41.27 |
marcosw | henrys: if you deleted a real user (i.e. an ex-employee) it would be a problem. | 15:41.46 |
jogux | I think when I've deleted bugzilla users before I've then had to manually clean up things in the database or the sanity check fails | 15:42.03 |
kens | Just disablign them is enough at least for us dumb users | 15:42.49 |
henrys | I'm sure folks won't delete when they see the warnings | 15:43.05 |
marcosw | jogux: I believe iptables are reset at boot, so it's not a permanent solution. presumably when that happens they'll see the robots.txt file and stop on their own. | 15:44.31 |
jogux | nods. sounds good. | 15:44.43 |
marcosw | jogux: btw, I noticed warning related to the twiki in error.log: setlib.cfg: defined(@array) is deprecated at setlib.cfg line 55. | 15:45.01 |
| any reason I can't fix that? | 15:45.07 |
jogux | marcosw: pass. Robin set this all up :-) | 15:45.20 |
marcosw | jogux: right | 15:45.29 |
| Robin_Watts: see question about twiki ^^ | 15:45.44 |
jogux | marcosw: I'd be tempted to try upgrading to the latest first to see if that fixes it | 15:45.56 |
henrys | that reminds me I keep meaning to look at twiki notification so folks can sign up for particular page changes. | 15:46.47 |
marcosw | it looks like the latest release is 6.0.1 and we are running 6.0.0 | 15:48.06 |
chrisl | henrys: my language_switch branch now has Ghostscript running and tigering - I haven't tested beyond that. | 15:48.15 |
marcosw | the twiki upgrade procedure is to "install the new version, install any extensions, copy the content , copy the users, apply customizations, apply preferences, switch-over". that's insane. | 15:51.17 |
chrisl | I guess that's why nobody packages it for a distro | 15:52.30 |
| There are quite a few wiki systems in the Ubuntu repo, so it's a shame we didn't use one of those, and then, presumably, it would be (almost) automatic to keep it patched | 15:56.12 |
jogux | we use foswiki, which I don't think is in official debian, but there are .debs available. | 15:58.23 |
henrys | chrisl: but if your using LTS all the packages are ancient. I have to locally install stuff all the time. Maybe I need to switch off LTS | 15:59.27 |
chrisl | henrys: the LTS versions get security updates | 16:00.08 |
| That's the point of LTS | 16:00.22 |
henrys | chrisl: in theory gs is at 9.10 on LTS, we've made a lot of security related fixes since then but I can't imagine they are available. | 16:04.17 |
| available to 9.10 that is | 16:04.40 |
jogux | henrys: sounds like they patched one recently http://www.ubuntu.com/usn/usn-2697-1/ | 16:06.55 |
chrisl | henrys: on LTS, the "9.10" currently has 24 patches on top of our 9.10 release | 16:07.25 |
Robin_Watts | marcosw: I set up twiki initially, because Picsel used to use it, and I had the idea that we could just import data from their twiki. | 16:07.35 |
marcosw | Robin_Watts: that makes sense | 16:08.10 |
Robin_Watts | Then I kinda stuck with it, cos I'm lazy and it seemed to work, and there didn't seem to be any obvious benefits of other wikis that I could find in my few minutes of googling. | 16:08.14 |
henrys | chrisl: wow I guess they do keep up... | 16:08.34 |
jogux | marcosw: robin: I don't believe it would be particularly difficult to switch if we wanted to. The raw data is .txt files in markup format or whatever it's called. | 16:08.36 |
Robin_Watts | If someone else wants to handle the switch, then I won't object. | 16:08.57 |
marcosw | the setlib.cfg warnings have not been fixed in 6.0.1, so I'm just going to fix them in 6.0.0 | 16:12.29 |
Robin_Watts | marcosw: Should we not be upgrading to 6.0.1 and then fixing them there? | 16:14.04 |
| AIUI the diffs between 6.0.0 and 6.0.1 are tiny small/ | 16:14.16 |
marcosw | I suppose I could do that. I'm wasn't sure if there 6.0.0 -> 6.0.1 changes were worth the effort. | 16:14.49 |
| they seem to have a plugin to make backup/restore/update easier. | 16:15.59 |
Robin_Watts | 6.0.0 -> 6.0.1. is a security fix that avoids people pwning the box. Though when we checked, we didn't seem to be vulnerable. | 16:16.06 |
marcosw | okay, I'll added upgrading the twiki to my list. I'm out of town Thursday to Sunday, so probably won't get to it before next week. | 16:17.12 |
jogux | you'd vaguely hope a x.y.1 type release could just be a 'dump it over the top' or something :) | 16:18.08 |
marcosw | i'm just happy they have a plugin to make it easier. | 16:18.38 |
Robin_Watts | wordpress rocks in this regard. | 16:18.54 |
chrisl | wordpress is a full cms, though, more involved than a wiki | 16:20.25 |
Robin_Watts | chrisl: Indeed. I set up a wordpress site at the weekend, dead simple. | 16:20.58 |
chrisl | That's the type of system I'd originally wanted our whole ghostscript.com site to use | 16:21.52 |
Robin_Watts | chrisl: wordpress seems well set up for blogs (posts with comments). Not so sure how it copes with a proper structured site. | 16:24.49 |
henrys | chrisl: so release after meeting, right? | 16:25.26 |
| after chicago meeting | 16:25.48 |
chrisl | henrys: it'll be towards the end of the month | 16:25.54 |
henrys | right | 16:26.00 |
chrisl | Hopefully an rc just before or just after the meeting | 16:26.17 |
| Robin_Watts: This looks okay: http://www.spectator.co.uk/ | 16:26.45 |
Robin_Watts | The spectator is done using wordpress?! | 16:27.19 |
chrisl | Apparently so | 16:27.30 |
marcosw | mvrhel_laptop: did you see the customer email asking about "setting transfer functions w/o halftone"? do you want to answer it directly or should I open a bug? | 16:29.53 |
chrisl | marcosw: I suspect the answer is a simple "no" | 16:31.51 |
mvrhel_laptop | marcosw: rayjj may be able to answer that better than me | 16:56.01 |
Robin_Watts | chrisl: Ah, I meant to ask... we're looking at replacement coffee machines. | 16:56.13 |
mvrhel_laptop | he was fooling with transfer functions and halftones just a bit ago | 16:56.20 |
Robin_Watts | ISTR that you have one? Are you nespresso or tacimo? | 16:56.30 |
| any recommendations either way? | 16:56.43 |
mvrhel_laptop | bbiaw | 16:56.58 |
chrisl | Robin_Watts: I use a Philips Senseo type machine | 16:57.07 |
Robin_Watts | gawd, another option. | 16:58.06 |
chrisl | TBH, I'd have thought just a filter machine would suit you | 16:58.36 |
jogux | We have a nespresso that yukko seems very happy with | 16:59.45 |
Robin_Watts | chrisl: I have a Krups filter/espresso/foam machine. | 17:00.19 |
| But the boss has decreed that it's too big for the new kitchen. | 17:00.32 |
| (And supposedly makes 'funny tasting coffee'. Which is interesting as she doesn't drink coffee.) | 17:01.07 |
chrisl | But it's not like either of you drink coffee...... | 17:01.27 |
Robin_Watts | chrisl: Indeed. The fact we don't drink coffee means the beans lie around for longer than we might like. | 17:01.53 |
| so possibly a pod based system would be a good idea. | 17:02.07 |
| Also, it needs to be easy enough for the idiots-in-law to use. | 17:02.21 |
chrisl | Robin_Watts: in your position I'd probably look into the tassimo - you can get tea, and chocolate pods for those | 17:02.56 |
| So you *might* get more use out of it | 17:03.13 |
Robin_Watts | chrisl: I was put off the tassimo by some of the reviews. the nespresso seemed a more reliable thing. | 17:03.25 |
| And you can get chocolate pods for the nespresso | 17:03.37 |
| (third party ones) | 17:03.41 |
| We have a boiling tap thing for tea. | 17:03.54 |
| (or at least we will have in a few weeks) | 17:04.08 |
chrisl | Really? Never seen them. I wouldn't recommend the Senseo for you because it's only coffee, and it still (effectively) uses ground coffee | 17:04.32 |
Robin_Watts | http://www.amazon.co.uk/capsules-MOCAVA-CHOCOLATE-compatible-Nespresso/dp/B00KFA1QFS/ref=sr_1_1?ie=UTF8&qid=1440522302&sr=8-1&keywords=nespresso+chocolate+pods | 17:05.12 |
chrisl | Ah interesting! Although Helen might not thank you for making hot chocolate easier to get hold of! | 17:06.03 |
| In your position, then, I'd probably go nespresso - from my limited experience of both, that seems to make the nicer coffee | 17:06.54 |
jogux | Robin_Watts: interesting, though like chrisl I'm not sure ordering them would be a good idea :) | 17:07.59 |
Robin_Watts | A single capsule would seem to be 50 calories and no fat. That's not a huge amount. | 17:26.18 |
jogux | not terrible. would be nice to try /one/. not suire about 120 :) | 17:42.16 |
Robin_Watts | jogux: yeah. | 17:42.35 |
| I found some which have (presumably powdered) milk in the capsules, and those are higher, obviously. | 17:43.20 |
sebras | tor8: sounds like a good plan. | 19:20.28 |
| Forward 1 day (to 2015/08/26)>>> | |