| <<<Back 1 day (to 2014/05/22) | 2014/05/23 |
andremedeiros | Hi everyone! | 10:25.23 |
| Quick question: what's the easiest way to solve the undefined symbols problem when linking to mupdf? | 10:25.59 |
| I've got a build on OS X that will throw a fit everytime I try to -lmupdf | 10:26.14 |
Robin_Watts | What undefined symbols problem? | 10:30.12 |
andremedeiros | if I don't link to the build folder | 10:37.46 |
| mupdf.a doesn't contain symbols for jpeg, openjpeg, freetype, etc | 10:37.54 |
kens | Of course not, since those are the third part libraries | 10:38.10 |
Robin_Watts | So if you miss out part of the code, it complains that it's missing part of the code? | 10:38.22 |
andremedeiros | I thought it statically linked them :-( | 10:45.21 |
Robin_Watts | They aren't linked into the mupdf lib itself. | 10:46.23 |
| So where you specify -lmupdf you should also specify -lfreetype -ljpeg etc. | 10:46.42 |
andremedeiros | does make install copy those artifacts ot the destination too? | 10:47.58 |
Robin_Watts | make install copies binaries, not libraries, AIUI. | 10:48.22 |
| tor8: 3 reviews on robin/master, 074cd54, 3eba9a3 and 20bbb27 | 10:52.20 |
andremedeiros | Robin_Watts: gotcha | 11:37.51 |
| makes sense now | 11:37.55 |
| I'll just have a "download and compile mupdf" step as a dependency and it now works :-) | 11:38.07 |
| another Q, is it possible for me to link to the mupdf library and traverse through a list device's nodes without having to copy paste the structs? | 11:47.17 |
tor8 | andremedeiros: copy paste? | 11:48.11 |
andremedeiros | I guess ;-) | 11:48.17 |
tor8 | what do you mean? #include <mupdf/fitz.h> should give you the device struct definitions | 11:48.41 |
| the display list struct is opaque by design, you're not supposed to go poking into its internals | 11:50.21 |
| if you need to traverse through it, create a fz_device and fill it with callback functions and fz_run_display_list on it | 11:51.49 |
| andremedeiros: does that answer your question? | 11:52.34 |
| Robin_Watts: a change of CTM should be detected when adding more text, so that commit doesn't make sense | 11:54.03 |
| 20bbb37 and 074cd54 LGTM, but "Flush pending text on a change of CTM" looks like it's hiding a more serious problem | 11:55.30 |
Robin_Watts | tor8: Why? | 11:55.55 |
| The problem in the test file is: | 11:56.29 |
| BT <set the ctm> "some text" TJ <set the ctm> ET | 11:57.05 |
| Un-flushed text doesn't have a matrix stored with it. | 11:59.57 |
| hence whenever we change the matrix, we should flush the text. | 12:00.13 |
tor8 | I'm reading the text showing code again to refresh my memory | 12:00.20 |
| and yes, it looks like we don't store the fully baked TRM * CTM in the text object | 12:00.32 |
| which means your patch is good | 12:01.00 |
Robin_Watts | cool. ta. | 12:01.18 |
tor8 | I remember checking a matrix for if we should flush, but not all the details | 12:01.41 |
| Robin_Watts: I've briefly looked at the pdfium stuff. they seem to be shipping a multiple master substitute font, if the name FoxitSansMM isn't a lie | 12:02.38 |
| also, it looks like they've scrubbed the code clean of all blank lines and comments... | 12:03.02 |
| the set of CMaps they ship is a pretty small subset | 12:03.45 |
| Robin_Watts: two commits on tor/master | 14:29.38 |
rayjj | kens: darn. I didn't notice that the question about the PDF page split and merge was from HCL. I explained how to do this, then saw your email :-( | 14:35.34 |
Robin_Watts | tor8: OK. | 14:35.50 |
kens | rayjj, oops.... | 14:36.40 |
Robin_Watts | rayjj: The correct answer is "use mutool clean" :) | 14:36.50 |
rayjj | henrys: should I add something to the bug reminding them to make sure and be AGPL compliant, or to contract for support ? | 14:37.14 |
| henrys: or just let it go ? | 14:37.22 |
kens | Miles mailed therm this mornign | 14:37.40 |
rayjj | Robin_Watts: I know, but if somebody is already using gs, then I didn't want to have to explain all about mupdf and mutool | 14:38.06 |
henrys | rayjj: I'd just leave it. | 14:38.53 |
rayjj | Robin_Watts: and I wasn't sure how good mutool clean is with handling links when selecting pages | 14:38.56 |
| henrys: OK. thanks | 14:39.02 |
| at least the documentation on doing it with gs is now in the bug tracker :-) | 14:39.33 |
henrys | rayjj, kens:we probably should start steering folks away from gs and pdfwrite for that sort of job. | 14:43.49 |
| hcl aside | 14:45.08 |
kens | They could just as easily use pdftk | 14:45.42 |
Robin_Watts | henrys: FYI: We just committed the font changes and the required test script changes for SOT. | 14:45.42 |
henrys | Robin_Watts: wow, I'll have to read the commit for that one. | 14:46.45 |
Robin_Watts | henrys: Turned out to be far fewer script changes than we'd feared. | 14:47.14 |
| (pedro, marcosw and I slogged through all the bitmap changes, and made a list of the scripts that needed fixing) | 14:47.41 |
henrys | Robin_Watts: so the scripts are "TODOS" | 14:48.08 |
Robin_Watts | no, we then fixed the scripts. | 14:48.20 |
rayjj | Robin_Watts: that's probably good news if there are future 'fixes' the the layout, right ? | 14:48.20 |
| s/the the/to the/ | 14:48.56 |
Robin_Watts | We spotted some scripts that weren't really testing what they were supposed to be, or were looking odd anyway, and those are TODOs, but hopefully we haven't damaged the coverage with the current commit. | 14:49.07 |
rayjj | back to "fun" with transparency :-( | 14:51.59 |
kens | For me, named destinations.... | 14:52.18 |
henrys | Robin_Watts: I just see changes to font map file, the rest is not version controlled? | 14:55.27 |
Robin_Watts | henrys: What commit are you looking at? | 14:59.57 |
henrys | 44b18c10 | 15:00.31 |
Robin_Watts | 12bb887 | 15:00.46 |
henrys | oh that is out of date | 15:00.55 |
| it said update fonts ;-) | 15:01.05 |
Robin_Watts | I see not 44b18c10 | 15:01.16 |
| Are you looking at sot.git or epage.git ? | 15:01.25 |
henrys | epage | 15:01.33 |
Robin_Watts | We moved from epage.git to sot.git | 15:01.43 |
| sot.git contains both epage and resources. | 15:01.50 |
| You need to recheckout. | 15:01.58 |
henrys | Robin_Watts: will do | 15:02.02 |
| I can blow away epage completely then? | 15:02.32 |
Robin_Watts | yes. | 15:02.36 |
henrys | Robin_Watts: the coordinate changes look painful | 15:22.59 |
Robin_Watts | henrys: indeed. | 15:30.53 |
| but overall it was much less work than I'd expected | 15:31.09 |
| I was fearing that it would be man-weeks of work. | 15:31.29 |
marcosw | I'm adding a couple of 20 Amp circuits to my garage and so have turned off power to some of the cluster nodes, they'll be back up in a couple of hours (unless I electrocute myself, in which case it may take a bit longer). | 15:56.38 |
Robin_Watts | marcosw: Take care. | 15:57.11 |
pedro_pc | marcos: yeah, rubber suit time ;) | 15:58.02 |
Robin_Watts | Wearing rubber soled shoes doesn't help if you touch live/earth with different hands. | 15:59.37 |
marcosw | I'll turn off power to the entire house, wear rubber soled shoes, and work with one hand behind my back. | 15:59.43 |
Robin_Watts | shocks across the body are the things to avoid. | 15:59.49 |
marcosw | The one hand behind the back trick is something I've seen "real" electricians do. | 16:00.01 |
Robin_Watts | yeah. | 16:00.07 |
henrys | marcosw: I don't suppose you'd consider a licensed insured electrician and a bill for miles. | 16:00.57 |
Robin_Watts | In the UK, by law, all "new" installations have to be done by certified electricians. | 16:01.42 |
| repairs can be done by anyone, but householders can't put in new circuits themselves. | 16:01.58 |
| (or at least, they have to be signed off by a qualified sparks) | 16:02.25 |
kens | night all | 16:58.26 |
marcosw | I'm turning off power to my house now, so will take down the rest of the cluster nodes here. | 17:06.23 |
Robin_Watts | ooh, so HCL are working for a real customer. | 17:16.38 |
chrisl | I think we still leave it to Miles to sort out...... | 17:17.03 |
henrys | Robin_Watts: yeah I think we knew that ;-) It does ring a bell at least | 17:17.21 |
Robin_Watts | I thought the procedure was that we had (theoretically at least) a single point of contact for a support controact and everything should come from them. | 17:17.28 |
| Otherwise we could end up supporting HCL even when they weren't working for a customer. | 17:18.20 |
henrys | notes we did the HCL drill over support email Sept. 2012 | 17:24.32 |
Robin_Watts | Indeed, so we did. | 17:25.35 |
| marcosw: success? | 17:32.29 |
marcosw | Robin_Watts: yup, all done. | 17:32.48 |
Robin_Watts | cool. | 17:32.53 |
marcosw | henrys: I did consider getting an electrician but I'm too Ron Swanson to do that :-) | 17:33.08 |
Robin_Watts | marcosw should convert all the nodes to water cooling. Think of the cash you'd save never needing to heat hot water for your house again :) | 17:33.59 |
marcosw | Robin_Watts: but I could only take showers during a cluster run... | 18:05.12 |
henrys | mvrhel_laptop: hello | 19:25.43 |
mvrhel_laptop | henrys: good morning | 19:25.55 |
henrys | mvrhel_laptop: miles certainly has me confused this today | 19:25.56 |
| s/this// | 19:26.06 |
mvrhel_laptop | henrys: I saw the emails about the unsupported customer stuff | 19:26.17 |
| is that what you mean? | 19:26.26 |
henrys | mvrhel_laptop: yeah | 19:26.33 |
mvrhel_laptop | oh I see that he just sent them a bit ago | 19:27.51 |
henrys | mvrhel_laptop: well one of us doesn't understand what's going on ;-) | 19:30.25 |
mvrhel_laptop | henrys: well that includes me. Are either of them a customer? | 19:31.22 |
| hold on, looking up customer numbers | 19:31.38 |
henrys | hcl is not a customer the other is. The "other" hires hcl to fiddle with gs. | 19:32.38 |
mvrhel_laptop | right. and the other is full support according to the list | 19:33.12 |
henrys | mvrhel_laptop: anyway I'm sure this will straighten itself out, hows the trip otherwise? | 19:33.14 |
mvrhel_laptop | henrys: I think the visit here has gone well. Jeong Kim was a great help getting us to the meetings and helping out during the meetings | 19:34.03 |
| Scott has come down with food poisoning though | 19:34.17 |
| I am hoping he is better today | 19:34.34 |
henrys | ugh | 19:34.50 |
| jeong kim - haven't seen him in ages. Is he still doing the pdf editor? | 19:35.17 |
mvrhel_laptop | No. He is doing some cloud based stuff these days | 19:35.36 |
| He seems like a pretty sharp guy | 19:35.47 |
henrys | mvrhel_laptop: did you guys eat something unusual? | 19:36.25 |
mvrhel_laptop | No. I suspect Scott had something on the airplane. He only ate here at the Hyatt. He and Jeong Kim had the same thing. Korean BBQ | 19:37.03 |
henrys | cleared up miles thinks it's another group at the company | 19:58.17 |
mvrhel_laptop | henrys; yes just talke with him about it | 21:25.24 |
| scott is better | 21:25.31 |
| I think we are doing breakfast then off to the DMZ | 21:25.40 |
henrys | mvrhel_laptop: well you wouldn't want to miss that right up there with Disney | 22:13.11 |
rayjj | henrys: I emailed that I had already replied to the bug report (support request) from HCL before I noticed that it was HCL, so maybe someone could point out to the real customer that they got prompt support | 23:43.12 |
| (due to my lack of attention to who submitted the bug) | 23:43.33 |
| mvrhel (for the logs) I am mired in transparency problems with knockout background, but am getting closer to understanding. If you have any knockout files that work, that would help. | 23:44.49 |
| mvrhel: other than that, I hope you are having a good time, and don't eat what Scott eats (or Miles, either, for that matter) :-) | 23:45.38 |
| hi, marcosw. Power all sortef out (as opposed to shorted out) ? | 23:55.25 |
| marcosw: please see the logs re tha HCL question about pdf merging. I think I inadvertently told them what they need, so IF there customer is actually OUR customer I didn't mess up -- I did a good thing | 23:57.04 |
| Forward 1 day (to 2014/05/24)>>> | |