| <<<Back 1 day (to 2015/02/26) | 20150227 |
Robin_Watts | A clean checkout of the mupdf source builds on Win32 for me. | 00:03.01 |
mvrhel_laptop | not sure | 00:24.18 |
| but mutool def has an issue | 00:24.30 |
| let me push my fixes for the gsview stuff first then I will look at why my win32 solution is not all building | 00:24.59 |
| Robin_Watts: see bug 695853 for mutool issue | 00:28.15 |
| Robin_Watts: so I am getting linking errors. js_setcontext in pdf_new_jsimp | 00:29.19 |
| so the mupdf, mupdf-curl, and mujstest projects don't build for me | 00:30.12 |
| likely some mujs stuff? | 00:31.07 |
| and js_pushiterator | 00:32.36 |
| Likely I don't have something set up properly in the 3rd part libs | 00:38.24 |
Robin_Watts | mvrhel_laptop: git submodule update --init | 00:44.34 |
| You're not up to date with the latest mujs changes. | 00:45.10 |
| If you're still having problems after that, let me know and I'll look in the morning. | 00:45.41 |
mvrhel_laptop | ok | 00:47.03 |
Robin_Watts | I'll look at the mutool problem tomorrow too. thanks. | 00:49.46 |
mvrhel_laptop | stupid perl path is wrong with call from git-submodule. I swear I have more perl path issues... | 00:51.21 |
| hmm supposedly it updated but still the same issue | 01:04.16 |
| will come back and look at this a bit laster | 01:04.23 |
| later | 01:04.26 |
| hmm my version of vs does not like | 01:07.32 |
| static __inline int signbit(double x) {union{double d;__int64 i;}u;u.d=x;return u.i>>63;} | 01:07.34 |
| ok def. giving up on this for now. will go work on stefan's issue for a bit | 01:09.14 |
| gosh I do see the changes in the logs so I have pulled them | 01:15.43 |
| not sure why they are not there | 01:15.54 |
| this should not be that difficult. def. stopping for now. may bug you in the morning Robin_Watts | 01:16.43 |
marcosw | bugzilla is back up but I messed up the dns settings so it's going to be up to another hour before the server is accessible. | 02:20.06 |
tor8 | kens: https://www.kickstarter.com/projects/modiphius/the-thunderbirds-co-operative-board-game-by-matt-l/ | 10:45.51 |
kens | Oooh, sounds like fun | 10:46.07 |
| THe people I game with are all the right age too | 10:46.27 |
| Well I've played Pandemic and Forbidden Island, and they were both good, so looks like I'm in on this one, thanks for the pointer tor! | 10:48.39 |
tor8 | kens: I'm tempted, even though I rarely like coop boardgames. but then that's the 12-year old in me that watched lots of thunderbird reruns :) | 10:49.23 |
kens | I enjoy coop games (and also 'mostly' coop, like BSG) | 10:49.52 |
| Hmm the expansion set level is already sold out :-( | 10:51.10 |
tor8 | looks to me like they've got two identical pledge levels at 65 pounds for game+expansion | 10:53.19 |
kens | Yes ans also one in French | 10:53.29 |
tor8 | I tend to enjoy competitive games more, like wargames and the old knizias. not a fan of modern german-style multi-player solitaire games. | 10:53.35 |
kens | We play all kionds. Most recently we've been playing the Hobbit deck-building game, which I tend to do well at for no obvious reason | 10:54.15 |
| But we play co-op games as well and I'm sujre I can con everyone into a game of Thunderbirds for the nostalgia if nothing else :-) | 10:54.45 |
tor8 | Valley of the Kings is the only deck-building game I've tried that I've found to be an actual improvement over Dominion | 10:56.20 |
| but then, I haven't tried all that many... | 10:56.31 |
kens | Yes, I see there's a second 'everything' level so I gues that'll be me | 10:56.33 |
| Well, I don't own that many games, but one of my old school friends has his spare bedroom converted to storage for games :-O | 10:57.14 |
| Hmm, not going to GenConor Essent this year, best go for delivery | 10:57.52 |
tor8 | kens: you've got enough bookshelves, too bad they're all full of books instead ;) | 10:58.55 |
kens | Seriously, Mark has *way* more games than we could sotre here..... | 10:59.20 |
| An entire second bedroom shelved | 10:59.37 |
tor8 | kens: over 500 games then? that does sound a bit excessive... given how you'll really want to get them all played eventually. | 11:03.40 |
kens | Mark is a bachelor, so he plays several times a week :-) | 11:03.59 |
tor8 | well nevermind in that case :) | 11:04.21 |
kens | The plus side from our POV is he only brings along the good games when he visits us :-) | 11:04.22 |
tor8 | kens: so I guess we'll meet up at the gate on monday? | 11:07.07 |
kens | Sounds good to me | 11:07.26 |
| Planning to be at the airport at ~10:30 but I'll have to check bag and do the airport security theatre | 11:07.59 |
| So probably easiest to meet at the gate. | 11:08.09 |
kens | thinks; must go and print stuff...... | 11:08.22 |
tor8 | I should be in at around 9:15 but terminal 5 doesn't really have any good meeting spots | 11:08.42 |
kens | I don't think I've ever been there to be honest. | 11:09.00 |
| I do hope the luggage system is all fixed now :-) | 11:09.09 |
tor8 | me too! I suspect I will have to check a bag as well, I doubt everything will fit without risking the bag bursting | 11:09.39 |
kens | Yeah, its a whole week away and ski gear is bulky | 11:09.57 |
| I'm packing my boots too, if they'll fit..... | 11:10.12 |
Robin_Watts | kens: Just take a bigger suitcase :) | 11:11.57 |
kens | Robin_Watts : I'm more worried about the weight, ski boots are heavy | 11:12.16 |
Robin_Watts | ah, weight, yes. | 11:12.22 |
kens | OTOH ski wear is light.... | 11:12.37 |
| Worst case I put my boots back in their bag and check 2 bags, and pay for it | 11:12.59 |
Robin_Watts | or wear the boots :) | 11:13.08 |
| wear them onto the plane, swap for slippers once in your seat. | 11:13.38 |
kens | LOL now that *would* make for an uncomfortable 10 hour flight :D | 11:13.39 |
| I'd have to keep wearing them they're too big to go overhead, unless there's nobody else in the row | 11:14.11 |
Robin_Watts | kens: will they not fit under the seat in front? | 11:14.28 |
kens | Err, no, not a chance | 11:14.35 |
| Solid boots to mid calf, with buckles | 11:14.50 |
| I might get one under each seat, maybe | 11:15.08 |
| Its not a problem, if needs be I'll pay the extra. Its probably not much more than renting a pair and I'd much ratehr wear my own wehen ski-ing | 11:15.58 |
| Good grief, initialising an image enumerator is just too much like hard work. | 11:17.11 |
| tor8 thanks for that pointer, just in time! Tomorrow they are locking the £65 pledge level, so just made it in time | 11:22.50 |
tor8 | kens: not a problem! first thing I thought of when I spotted it last night was you :) | 11:23.40 |
kens | :-D | 11:23.49 |
| Hmm, Call of Cthulhu setting for WW2...... | 11:26.13 |
| "Achtung! Cthulhu" :-) | 11:26.27 |
Robin_Watts | Tentacle Hoch! | 11:27.57 |
kens | At least explosives hsould be easier to come by..... | 11:28.18 |
| Maybe even a prototype nuke, that ciould come in handy | 11:28.28 |
Robin_Watts | tor8: so mvrhel found a problem with the linearisation code last night. | 11:41.25 |
| I'm looking at it now. | 11:41.29 |
| His input file has /Group 239 0 R | 11:41.47 |
| but object 239 is a free one. | 11:42.05 |
kens | Isn't that invalid ? | 11:42.25 |
| It *sounds* like a bad idea..... | 11:42.39 |
Robin_Watts | It certainly screws my code :) | 11:42.59 |
kens | I don't think a PDF file should be using anobject marked as free, that sounds wrong to me | 11:43.17 |
| Because you could (I think) compact the code and discard the object. Of course, renumbering the objects in the PDF file would quickly find that the object *wasn't* free. | 11:44.03 |
Robin_Watts | I think, as a patch, I'm going to make any objects that are supposedly free, but are still referenced, be just 'null'. | 11:48.05 |
kens | Sounds reasonable to me | 11:48.14 |
| Broken file in = broken file out :-) | 11:48.25 |
Robin_Watts | Broken file in = Equivalently broken file out is as good as we can offer I think. | 11:48.46 |
| currently the code asserts. | 11:48.49 |
kens | I have no problem with GIGO, but its amazing how many people think we shoudl fix the file...... | 11:49.24 |
Robin_Watts | Well, sadly this code is often what people use to fix broken files :) | 11:49.50 |
| 'fix'. | 11:50.03 |
kens | Yeah, people use GS for that too :-( | 11:50.16 |
| Ah, now today's xkcd cartoon makes sense: | 12:16.01 |
| http://www.bbc.co.uk/news/uk-scotland-highlands-islands-31656935 | 12:16.01 |
Robin_Watts | tor8: There are various places in pdf-object.c where we say return; /* Can't warn :( */ | 13:13.19 |
| presumably that was because we didn't have a ctx before? We do now... | 13:13.29 |
kens | <sigh> gx_image_enum_common_init doesn't set the structure's memory pointer...... | 13:18.09 |
Robin_Watts | kens: Can you extend it so it does so? | 13:21.25 |
kens | I could, but everywhere else just initialises the memory poitner after the 'init' | 13:21.49 |
Robin_Watts | It sounds like it's broken as it is. | 13:22.07 |
kens | I'd have to change the arguments to the fuinction to do the init, and that means I'd then have to go back and change all theplaces that call it | 13:22.14 |
| Ro there are so many things that are 'broken' in my opinion :-( | 13:22.40 |
Robin_Watts | Your choices are to leave it broken for someone else to trip over in future, or to improve it. | 13:22.43 |
tor8 | Robin_Watts: true. we could add fz_warn's to those spots. | 13:23.08 |
kens | I'm leaving it, at least for now. I'm trying to get this code working, and then address all the *other* broken things I notcied, and have documented in the code..... | 13:23.14 |
Robin_Watts | kens: So you're opting for "it was broken when I found it". | 13:23.44 |
kens | Yeah pretty much | 13:23.51 |
tor8 | Robin_Watts: yeah. a X 0 R reference to a free object should be 'null' | 13:23.56 |
| as in, we already resolve it to the 'null' object IIRC | 13:24.11 |
kens | If I dfon't I'll be the next couple of years working through all the things I've found that I would like to fix | 13:24.13 |
Robin_Watts | Which is valid, but will limit the sympathy you get from people when you trip over the next broken thing :) | 13:24.22 |
tor8 | so changing the reference to a null when cleaning/linearizing would be the right thing to do | 13:24.24 |
kens | Robin_Watts : Like I said, there are so many, many broken things | 13:24.39 |
Robin_Watts | tor8: yeah, I'm updating the sweeping code to spot references to free things and to bake in nulls. | 13:25.16 |
kens | I could start with the names of methods. We have "text_begin" but "begin_image". Consistency ? We've never heard of that | 13:25.34 |
Robin_Watts | kens: I agree. but then that's in the device interface, and if we change that ray will have a fit :) | 13:27.00 |
kens | Yeah I know :-( | 13:27.10 |
| Most of what I'm referring to is in the device interface. It needs to be swept awat and replaced wholesale. But that really would be a mammoth task | 13:27.40 |
Robin_Watts | ok, I have just swept away all but 3 objects in this file. I think I cocked up :) | 13:28.06 |
kens | Ooops :-) | 13:28.17 |
| But its much cleaner now I guess ? :-D | 13:28.25 |
| Oh good, that *increased* the number of seg faults. I understand two thirds of them, I guess its time to try out Lnux for the other third. | 13:39.49 |
henrys | kens: can you make a list of your observations and put them in a "misc" bug report. When you're done there may be a few things you've found that will be easy to fix. | 14:15.35 |
| kens: I need a question mark in there somewhere | 14:15.50 |
| chrisl: do you cluster test with the ufst or just wait for marcosw ufst test? I have a change I want to make sure doesn't affect the ufst build for pcl. | 14:18.41 |
kens | henrys, I have a list of stuff in comments in the C file I'm working on | 14:18.49 |
| I'll add this last one as well. | 14:19.05 |
henrys | kens: oh comments are fine too. | 14:19.07 |
| with something like NB or something so we can find them ? | 14:19.28 |
kens | Well these are things I meant possibly to address at some point, but I'll put everything there, it'll be easier to find them all at once | 14:19.53 |
chrisl | henrys: I just leave it until marcosw's tests run. I prefer not to risk their code appearing accidentally where it shouldn't | 14:20.23 |
| henrys: I do have scripts I use locally to run a lesser number of tests before pushing | 14:20.55 |
henrys | chrisl: probably sensible. | 14:20.58 |
| well the problem with your trip will not be too little snow... I hope the driving conditions can hold, so far everything seems okay. | 14:23.33 |
kens | It seems to be frequent relatively light snow, so hopefully the ploughs and gritters can keep up | 14:24.01 |
henrys | kens: yeah copper is 3-5 inches sat-tues... then tapering weds. | 14:26.56 |
kens | And sunny friday and saturday, oh well..... | 14:27.11 |
| At least it looks like it'll start nice for teh folks going to Vail | 14:27.56 |
| Ah, they've upghraded Thursday to sun as well, and only -3 top daytime temperature | 14:28.21 |
| OTOH wednesday is now a top of -12, that's quite chilly | 14:28.53 |
| My freezer is only at -18 | 14:29.09 |
henrys | kens: you might be surprised, the low humidity here makes it feel a bit warmer. I really notice a difference between cold here and where you are. | 14:43.38 |
kens | Oh yeah I'mwell used to that from the Alps | 14:43.50 |
henrys | kens: oh right | 14:44.10 |
kens | It *always* feels colder here for the same temperature because of humidity, and often wind chill as well, here in the UK | 14:44.13 |
| But -12C is pretty cold :-) | 14:44.32 |
| I do find it amusing that we get on the plane in Salzburg, in the sun admittedly, at 0 or below in a T-shirt, then get off the flight in the UK and immediatley find a coat to wear | 14:45.58 |
Robin_Watts | 1 commit on robin/master then tor8. | 14:49.33 |
tor8 | Robin_Watts: um, if (markobj()) returns an object, insert a new_null? | 14:53.37 |
| shouldn't that be if (!markobj())? | 14:53.48 |
| oh wait, it's me not capable of reading diff output... | 14:54.42 |
Robin_Watts | phew. was contexting switching back there :) | 14:55.21 |
tor8 | Robin_Watts: so I've got the pdf_csi rework into a mostly working shape now | 14:56.24 |
| for op-run. missing is the handling of OCGs, which I've still to understand. the op-buffer stuff needs testing and the op-filter is still #if 0'd out | 14:56.56 |
| if you want to have something to bash on while I'm on the plane, I could leave the op-filter for you :) | 14:57.15 |
Robin_Watts | tor8: I've started work on my proposed PDF_NAME_ changes in front of the telly. | 14:57.46 |
| I plan to do that on the plane :) | 14:57.54 |
tor8 | right. I'm not sure I like that, but I guess it's worth benchmarking at least. | 14:58.11 |
Robin_Watts | The changes are actually working out really nicely. | 14:58.26 |
tor8 | are you doing the table of static global pdf_obj's and == comparisons? | 14:58.41 |
Robin_Watts | For things like pdf_is_int etc, we typically do: | 14:58.43 |
| if (!obj || obj->kind != PDF_INT) etc. | 14:59.01 |
| That translates to if (opj < PDF_NAME_LIMIT || obj->kind != PDF_INT) | 14:59.25 |
tor8 | oh, you're using the obj pointer as an integer instead? | 15:00.08 |
Robin_Watts | the 'nastiness' is entirely contained within pdf-object.c, and is invisible above the interface. | 15:00.13 |
| tor8: yes. | 15:00.21 |
| and the nastiness is way less than I thought it would be. | 15:00.39 |
tor8 | Robin_Watts: it does mean nastiness is exposed through the debugger though | 15:00.52 |
| "gdb$ print obj" won't work nicely | 15:01.04 |
Robin_Watts | It means the storage requirement for every name drops from namelength + structure size + malloc overhead to 0. | 15:01.29 |
| tor8: That is true. | 15:01.40 |
| but you can use "call pdf_printobj(obj)" | 15:02.01 |
tor8 | preallocated static global objects would also mean the storage requirement drops to 0 | 15:02.03 |
Robin_Watts | tor8: at a cost of slowing down every comparison operation. | 15:02.47 |
tor8 | but globals in DLLs are not nice on windows | 15:02.55 |
Robin_Watts | Avoiding the strcmp's is an aim of this. | 15:03.19 |
tor8 | Robin_Watts: really? !strcmp(pdf_to_name(obj), "Resources") would become (obj == PDF_NAME_Resources) | 15:03.20 |
Robin_Watts | tor8: will ponder on that some more. | 15:03.49 |
| we can try both ways just by changing the one file. | 15:04.06 |
| the changes throughout the rest of the code should be identical. | 15:04.18 |
tor8 | PDF_NAME_Resources being a macro that casts the enum to a pdf_obj, vs a global variable, I take it? | 15:04.44 |
Robin_Watts | tor8: yes. | 15:05.01 |
tor8 | yeah, then trying both ways would be trivial | 15:05.44 |
Robin_Watts | It also occurred to me that null/true/false would fit into this scheme too. | 15:06.14 |
tor8 | I wonder if you can do un-stringifying in the preprocessor | 15:06.17 |
| so we could have a macro pdf_is_name(obj, "Resources") that would expand to obj == PDF_NAME_Resources | 15:06.37 |
| or well, we could do the opposite, PDF_IS_NAME(obj, Resources) that does the stringification if we need to | 15:07.03 |
Robin_Watts | Yeah. | 15:07.10 |
tor8 | and otherwise just (obj == PDF_NAME # name) | 15:07.18 |
Robin_Watts | Then, if I have time while I'm away, I want to look at turning all the path processing into a callback mechanism, and extending the path representation so it can be more compact. | 15:13.09 |
| mvrhel_laptop: Morning. | 15:20.13 |
| tor8: So, were you happy with that commit ? | 15:20.21 |
| mvrhel_laptop: The pdf clean thing was a long standing issue with the code. It fails if you give it a PDF file that references an object that doesn't exist. | 15:25.46 |
| I have a fix for that now (just waiting for tor to review). | 15:25.56 |
tor8 | Robin_Watts: yes. | 15:30.34 |
Robin_Watts | Ta. | 15:30.43 |
mvrhel_laptop | Robin_Watts: ok cool! | 15:30.58 |
Robin_Watts | mvrhel_laptop: You were having git problems? | 15:31.09 |
mvrhel_laptop | Robin_Watts: yes. not sure why the submodule is not fast forwarding it seems | 15:31.28 |
| I look at the logs and there are commits in there for this month | 15:31.41 |
Robin_Watts | mvrhel_laptop: What does git diff show at the top level ? | 15:31.50 |
mvrhel_laptop | for the submodule? | 15:31.57 |
| or the whole project? | 15:32.04 |
Robin_Watts | The whole project. | 15:32.09 |
| (i.e. does it mention thirdparty/mujs in the diff?) | 15:32.33 |
mvrhel_laptop | yes. for mujs, curl and openjpeg | 15:33.43 |
| not sure why | 15:33.45 |
| not like I have made any changes there | 15:33.51 |
Robin_Watts | ok, and git submodule update --init doesn't make them go away ? | 15:34.22 |
mvrhel_laptop | correct | 15:34.28 |
Robin_Watts | OK. cd thirdparty/mujs && git stash && git drop | 15:34.54 |
| OK. cd thirdparty/mujs && git stash && git stash drop | 15:35.01 |
| That should throw away any local changes to the thirdparty. | 15:35.27 |
| Does that make git diff at the top level stop showing differences in the submodule ? | 15:36.37 |
mvrhel_laptop | no. and it says there are no local changes to stash | 15:38.04 |
| but it shows up on the diff still | 15:38.10 |
Robin_Watts | What does the diff say? | 15:38.19 |
mvrhel_laptop | oh hold on | 15:38.33 |
| it shows master as being at 9e41afffa65f4c42aa7e68ddb2ffcacaadaa32ed | 15:39.42 |
| but I am at c1ad1ba1e482e7d01743e3f4f9517572bebf99ac | 15:39.52 |
| and it shows me as being way ahead of master | 15:40.07 |
| it is really odd | 15:40.13 |
| oh no its not odd | 15:40.42 |
Robin_Watts | c1ad1ba is where you should be. | 15:40.52 |
mvrhel_laptop | let me just do a hard reset to there | 15:40.58 |
tor8 | mvrhel_laptop: where 'master' is in the submodule is irrelevant, it's only where HEAD is that matters | 15:41.24 |
| submodules generally operate in detached HEAD mode | 15:41.55 |
mvrhel_laptop | ok well origin/Head and origin/master are both at c1ad1ba | 15:41.57 |
chrisl | Does git diff actually list any differences for the submodules? | 15:42.08 |
mvrhel_laptop | no | 15:42.14 |
tor8 | where HEAD is checked out without but not connected to an actual branch | 15:42.14 |
| origin/HEAD is only where the remote (i.e. source) repository is at | 15:42.36 |
Robin_Watts | chrisl: git diff will only show that a submodule is in the wrong place (or that it's dirty) | 15:42.40 |
tor8 | not your checkout | 15:42.43 |
chrisl | Robin_Watts: I thought mvrhel_laptop said git was reporting changes in the submodules..... | 15:43.26 |
mvrhel_laptop | git diff does report diff | 15:43.41 |
| but not in the normal way | 15:43.57 |
Robin_Watts | chrisl: git diff can say something like: thirdparty/mujs\n-2343274923843298742309842304987230497\n+3297423498723409237482374032\n | 15:44.13 |
mvrhel_laptop | right | 15:44.17 |
Robin_Watts | mvrhel_laptop: So is top level git diff happy with mujs now? | 15:44.51 |
mvrhel_laptop | no | 15:44.57 |
Robin_Watts | Ok, bandwidth is cheap, right? | 15:45.05 |
| delete thirdparty/mujs | 15:45.12 |
| then do: git submodule update --init | 15:45.19 |
mvrhel_laptop | yes. I feel like removing the whole thing and reloading | 15:45.20 |
chrisl | Right - my reason for asking is that I've seen an issue with msys-git before where it wouldn't update because there was changes, but git diff reported no differences...... | 15:45.34 |
| The only solution we could find was to delete and re-checkout that portion of the tree | 15:46.32 |
mvrhel_laptop | ok now I am very confused | 15:49.33 |
| I deleted and did the update init | 15:49.42 |
| and it still says there is a diff | 15:50.06 |
| wtf | 15:50.09 |
Robin_Watts | mvrhel_laptop: Do you have any other changes in your tree? | 15:50.27 |
mvrhel_laptop | yes | 15:50.32 |
| all the api fixes for gsview | 15:50.42 |
| let me get those up on the repos for you to review | 15:50.50 |
Robin_Watts | commit those so they are safe. | 15:51.01 |
| then git reset --hard | 15:51.13 |
mvrhel_laptop | ok it is all pushed up to my repos on casper | 15:55.00 |
tor8 | Robin_Watts: page sanitisation doesn't quite work on pdfref17.pdf page 1... the resulting resources dictionary is empty | 15:56.46 |
| and I'm too stupid to run the correct executable again... | 15:57.09 |
mvrhel_laptop | after reset hard the only diffs shown are thirdparty/curl, thirdparty/mujs and thirdparty/openjpeg | 15:57.11 |
Robin_Watts | mvrhel_laptop: We need to drop the mupdf.sln changes, I think. They don't do anything except change the format? | 15:59.26 |
| Otherwise all looks plausible to me. | 16:01.17 |
mvrhel_laptop | oh did I accidently include that | 16:03.34 |
| in the win32 solution | 16:03.41 |
| sorry | 16:03.43 |
Robin_Watts | np. | 16:03.46 |
mvrhel_laptop | thanks for catching that | 16:03.46 |
Robin_Watts | I can't see any reason for the submodule wierdness though. | 16:04.29 |
| mvrhel_laptop: I'm cloning your repo now. | 16:07.51 |
mvrhel_laptop | Robin_Watts: oh that will be interseting | 16:08.09 |
| Robin_Watts: ok. I removed the changes to the win32 project and fixed the commit. I am going to push to golden then | 16:13.35 |
Robin_Watts | just a mo. | 16:13.49 |
mvrhel_laptop | ok | 16:13.53 |
Robin_Watts | I can't see that that would be bad, but let me finish cloning your repo here first :) | 16:14.13 |
| Cloning the thirdparty stuff now. ironicly, curl is slow... | 16:14.35 |
mvrhel_laptop | Robin_Watts: do you think the 3rd party changes will even be there? I am not sure if they get pushed | 16:17.08 |
| since they are not committed | 16:17.17 |
| actually I guess they may just be behind | 16:17.40 |
Robin_Watts | mvrhel_laptop: There are commits in the history that say what the thirdparty positions should be. | 16:18.06 |
mvrhel_laptop | ok | 16:18.10 |
Robin_Watts | I can't see how your commit can alter them, but let's double check. | 16:18.30 |
| mvrhel_laptop: If you have no differences in your thirdparty dir (at least none that you know about), I'd just delete the entire thirdparty dir. | 16:22.27 |
mvrhel_laptop | ok | 16:22.36 |
Robin_Watts | Then from the top level do: git submodule update --init to bring 'em back in again. | 16:22.42 |
mvrhel_laptop | yeah! no diffs | 16:23.27 |
| lets see if win32 builds.. | 16:24.12 |
Robin_Watts | mvrhel_laptop: Have you got cable or something? ! | 16:24.44 |
mvrhel_laptop | internet does seem zippy today | 16:25.03 |
| and there is the def for js_getcontext now | 16:25.44 |
Robin_Watts | OK, I can't see any problems here. | 16:25.53 |
| so, I reckon your commit is good to go (unless you want to test it a bit first?) | 16:26.31 |
mvrhel_laptop | and it all builds for me now. | 16:26.32 |
| Robin_Watts: I did quite a bit of testing last night | 16:26.46 |
| and found two issues | 16:26.57 |
Robin_Watts | The clean one, and ? | 16:27.07 |
mvrhel_laptop | one I fixed as it was my goof in stomping a context during multiple threads | 16:27.19 |
Robin_Watts | oh yes. | 16:27.26 |
mvrhel_laptop | it all seems to be working well. I ran it through the paces of text search, rapid scrolling, saving files | 16:27.59 |
Robin_Watts | mvrhel_laptop: Go ahead and push, then I'll rebase and push my clean fix on top. | 16:28.08 |
mvrhel_laptop | ok thanks | 16:28.14 |
| ok all done | 16:29.21 |
| I have to run to my daughters school this morning. | 16:29.31 |
| thanks for helping me get this sorted out Robin_Watts | 16:29.39 |
Robin_Watts | np. | 16:29.43 |
mvrhel_laptop | gosh it keeps coming from stefan.... | 19:45.42 |
| back to work on his stuff | 19:45.49 |
Robin_Watts | I haven't seen a new email from him... was it not to support? | 20:02.41 |
mvrhel_laptop | Robin_Watts: he did cc support | 20:19.25 |
| found the issue with his previous email | 20:19.40 |
| this additive problem though I need to think about | 20:19.55 |
| I wonder if he would be willing to share his device | 20:20.38 |
Robin_Watts | <sigh>his stuff has been going into my spam bin. | 20:22.00 |
mvrhel_laptop | I hate when that happens | 20:22.07 |
| I wonder if his issues with the additive devices will show up with the psdrgb device | 20:29.18 |
| I will check that out | 20:29.23 |
| and yes it crashes | 20:31.22 |
| perhaps psdrgb should be added to a weekly testing | 20:33.43 |
| I am going to do a bisect to see when it broke | 20:34.03 |
| weird. bisect failed | 20:41.00 |
| wonder why | 20:42.01 |
| oh this has been broken for a while | 21:01.19 |
pipitas | Quick 'mutool' question⦠Is `mutool clean -d` supposed to be able to work with incrementally updated PDFs, preserving the previous updates in the output? I suppose not. I suppose it is a "feature" to clean these out, and only return the file with the last document version as the only one⦠| 21:51.37 |
| (I just want to be sure.) | 21:52.00 |
mvrhel_laptop | I can't quite figure this out | 22:30.24 |
| something is getting screwed up during gx_default_put_params | 22:30.48 |
| I really dislike trying to figure out put_params issues | 22:31.23 |
| error is occurring during param_commit | 22:32.05 |
| ok. I am at a lost as to what is going on. There is def something odd in the psd_write_header code but I can't even get that far | 22:48.12 |
| opening a bug and hoping that someone who is better at put_params can quickly see the issue | 22:48.42 |
| oh interesting | 22:50.24 |
| it has to be a deviceN spot color issue with the rgb device | 22:50.49 |
| tiger runs just fine but the altona file with the spots fails | 22:51.11 |
| actually, I take that back. tiger comes out blank | 22:52.00 |
| just device is really screwed | 22:52.09 |
| s/just/this/ | 22:52.21 |
Robin_Watts | pipitas: clean -d will rewrite the entire pdf. | 23:47.33 |
| Forward 1 day (to 2015/02/28)>>> | |