| <<<Back 1 day (to 2013/03/27) | 2013/03/28 |
mvrhel_laptop | have to figure out how to make the clicks work though in this damn xaml flipview object. dinner now though | 00:01.47 |
Robin_Watts | I have images more or less working with mupdfwrite now. I'll try to make it avoid decompressing indexed ones tomorrow. Then I have to do patterns/tiles. | 00:26.52 |
| ray_laptop: (For the logs) I committed the ialloc_init cleanup fixes - thanks for that. | 00:27.48 |
kens | chrisl I thought you were on holiday today.... | 08:47.50 |
chrisl | kens: I'm doing a bit today 'cause I was out quite a lot on Tuesday. I'll be finishing shortly (once I fix my broken build!) | 08:49.34 |
kens | Have a good weekend then | 08:49.49 |
chrisl | Thanks - hoping the weather will cooperate! | 08:50.36 |
kens | Well its very nice at the moment here | 08:50.54 |
chrisl | It is here, too, which I find concerning! | 08:51.45 |
sebras | goodbye hotmail, hello outlook. | 08:52.55 |
chrisl | sebras: run, run like the wind!! Once Outlook gets you, you never escape!! | 08:53.28 |
kens | yeah but.. hotmail ? | 08:53.49 |
kens | shudders | 08:53.58 |
chrisl | kens: that bouncing mail is weird - I'm wondering if the bounce came from somewhere other than GMail. | 08:57.53 |
kens | Which bounce ? Henrys ? | 08:58.11 |
| Marcos said he never got the support mail | 08:58.30 |
| You me and RObin did, but we all pull our mail off the server... | 08:58.42 |
chrisl | Yeh, so if Marcos has GMail forward his artifex mail to another account, it may have been bounced from there, rather than GMail | 08:59.42 |
kens | Hmm yes but I don't htink it obunced for Marcos, he just never got it | 09:00.02 |
chrisl | I thought, from what henrys said, that the customer saw a bounce massage - did I get that wrong? | 09:00.43 |
kens | *Henry* got a spam bounce when he replied to support | 09:01.00 |
| bounced as spam | 09:01.11 |
| Marcos just never saw the customer mail | 09:01.25 |
chrisl | Ah, well, that's just crap, really :-( | 09:01.46 |
kens | At least 3 of us did, but we all pull mail off the server. NB there was a reasonably large attachment which may have contributed... | 09:02.01 |
chrisl | My artifex mail stays on the GMail server, although I work directly with GMail, rather than forwarding elsewhere | 09:03.00 |
| I thought Robin_Watts had his forwarded to his own account | 09:03.20 |
kens | Hmm, well maybe that's no it then | 09:03.21 |
| Mine doesn't forward, I pull it off using Eudora | 09:03.33 |
| When I check mail Eudora polls all my accounts for me | 09:03.47 |
chrisl | Yeh, that's how I have it setup, but working thru IMAP rather than POP3 | 09:04.09 |
kens | OK so broadly similar to my setup. Don't know how Robin_Watts works | 09:04.48 |
chrisl | So, again, I wonder if marcosw is getting his mail forwarded to another server (rather than working directly with GMail), and is getting an extra layer of filtering as a result | 09:05.09 |
kens | Its possible, that might well explain where the mail went, but I got the impression he just used webmail | 09:05.39 |
chrisl | I dunno, might be worth mentioning. If GMail is applying different filtering rules to different accounts, that would be *really* appalling (but not entirely surprising!) | 09:06.33 |
kens | I guess I'll leave that one to Marcos :-) | 09:07.00 |
chrisl | Okay, I *think* I now have the memory allocation ID stuff working - and disabling it working. So I'll check over the changes when get back next week, before committing. | 09:12.11 |
| I was right - "string" allocations were a PITA! | 09:12.24 |
kens | OK have a good weekend chrisl | 09:12.26 |
chrisl | thanks - bye! | 09:12.31 |
Robin_Watts | tor8: http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=9299cd57069a428e8018f4fdf7230efdcfb25e96 | 11:23.28 |
tor8 | "View with MuPDF" perhaps? | 11:23.28 |
Robin_Watts | I wondered if it was convention to keep the name there the same as the exe ? | 11:23.29 |
tor8 | I have no idea | 11:23.29 |
sebras | chrisl: kens: well.. hotmail was there way before gmail, so I registered with them first... | 11:27.29 |
| I still have and use the account... | 11:27.44 |
kens | I paid for my email... | 11:27.45 |
sebras | kens: I pay now too, but I never use it. | 11:28.03 |
| kens: gmail/hotmail solves the problem. however microsoft now decided to convert hotmail.com to outlook.com... | 11:28.26 |
kens | Typically M$ | 11:28.37 |
sebras | kens: hopefully my old address still works for some time... | 11:31.13 |
henrys | paulgardiner: well it went okay - shirataki noodles helped a lot. | 14:23.34 |
paulgardiner | henrys: oh ok. thanks for the tip. I need some more ideas. Getting bored with my two eggs some prawns and 1/2kg of veg. | 14:25.32 |
Robin_Watts | paulgardiner: You should speak to Helen. she has loads of really low cal recipies. | 14:26.04 |
paulgardiner | That's a thought. Will do. | 14:27.25 |
henrys | amazing stuff, 0 calories and they are sort of filling for a bit http://en.wikipedia.org/wiki/Shirataki_noodles | 14:27.52 |
paulgardiner | Oh right. So stuff as much as you like, more or less. That is handy. | 14:28.30 |
henrys | paulgardiner: right I put an 8 oz bag of it in miso soup for 35 calories | 14:31.39 |
paulgardiner | Right. 17 portions of that it is! :-) | 14:33.53 |
Robin_Watts | Helen has a low cal chicken noodle soup thing. | 14:34.05 |
| I bet Shirataki noodles are hard to get over here. | 14:34.53 |
paulgardiner | Especially with all the 5/2ers stock piling. | 14:35.58 |
Robin_Watts | amazon sell them. | 14:36.40 |
paulgardiner | Waitrose say "Did you mean "Sharwoods"? | 14:37.27 |
Robin_Watts | http://www.amazon.co.uk/Japan-Centre-Shirataki-Noodles-180g/dp/B003XQKUPS | 14:38.27 |
paulgardiner | Actually that's quite impressive. They're search algorithm hasn't discovered any Shirataki, but it's sussed I'm looking for noodles. | 14:39.01 |
Robin_Watts | And you can get shirataki "rice" too. | 14:39.04 |
henrys | yikes that's expensive | 14:39.13 |
Robin_Watts | Welcome to rip off Britain. | 14:39.31 |
henrys | we eat the miracle noodle brand | 14:39.37 |
paulgardiner | Yeah 7 billion pounds a calorie | 14:39.52 |
Robin_Watts | 3p an ounce: http://www.amazon.co.uk/Miracle-Noodle-Shirataki-Noodles-Fettuccini/dp/B005CD3AZM/ref=pd_sim_grocery_3 | 14:40.27 |
henrys | http://www.miraclenoodle.com/p-293-variety-sampler.aspx?utm_expid=1948735-18&utm_referrer=http%3A%2F%2Fwww.miraclenoodle.com%2Fp-6-fettuccini.aspx#back | 14:44.14 |
Robin_Watts | http://www.skinny-noodles.co.uk/index.html | 14:46.12 |
| In PDF images with an indexed colorspace do we Decode before unindexing? or unindex then decode | 14:54.35 |
| ? | 14:54.38 |
| Unindex, then decode, I believe. | 14:56.31 |
henrys | hmm I wold think decode first | 15:00.10 |
| from the postscript manual: For an Indexed color space, the suggested Decode array ensures that component values that index a color table are passed through unchanged⦠implying decode first | 15:01.56 |
Robin_Watts | mupdf assumes that any decode array with have n*2 entries in it. where n = the number of channels in the base colorspace of the indexed color space. | 15:02.03 |
| I will test when I get back from running an errand. | 15:02.30 |
henrys | better stop yapping and fix my p1 bug | 15:08.24 |
Robin_Watts | Well /Decode [0 1] changes the output. | 15:27.30 |
| As does /Decode [0 1 0 1 0 1] | 15:27.47 |
kens | I htought the Decode applioed toe the main colour space. | 15:28.14 |
| So you should apply the Decode to the Indexed values, then use the result to look up the index table | 15:28.36 |
| Which is why the Decode for an Indexedc space maps 0->255 to 0->255 | 15:29.01 |
| If your Decode array is bigger than required it doens't (normally) trip an error, only if its smaller, so extra values have no effect | 15:29.29 |
Robin_Watts | OK, we have 2bpc coming in, so the default is [0 3], and indeed that seems to work. | 15:31.23 |
| so I think mupdf may be getting it wrong. | 15:31.29 |
| No, I just can't read code. It's OK. | 15:32.24 |
kens | OK goodnight all | 17:17.37 |
Robin_Watts | does the mvrhel_laptop summoning dance | 17:17.46 |
| Morning mvrhel_laptop | 17:51.28 |
mvrhel_laptop | good morning Robin_Watts | 17:51.48 |
Robin_Watts | I just committed some memsqueezing fixes for gs, that I think were in your area. | 17:52.17 |
| If you want to double check my changes, feel free. | 17:52.54 |
mvrhel_laptop | Robin_Watts: ok | 17:57.35 |
Robin_Watts | There is another #ubuntu-touch-meeting about docviewer this evening (same time). I plan on going along to lurk. | 17:59.58 |
mvrhel_laptop | sounds good | 18:02.01 |
Robin_Watts | coo. Interesting valgrindy bugs from marcosw. | 18:39.10 |
mvrhel_laptop | my support box is overflowing..... | 18:42.55 |
| a few of these look like color issues. some look like they are in littleCMS | 18:43.45 |
Robin_Watts | or they are undefined values passed from gs to lcms and only spotted by valgrind in lcms. | 18:46.24 |
| CachedXFORM looks at values coming in to see if they match the previous value it had - so undefined input values will show up there. | 18:47.19 |
mvrhel_laptop | right | 18:48.23 |
Robin_Watts | TetrahedralInterp16 only works on values that come from lcms though, I think. | 18:48.55 |
| I'll look into the lcms ones if you want. | 18:49.14 |
mvrhel_laptop | Robin_Watts: that is fine. I am not going to be looking at these in the near future | 18:52.34 |
Robin_Watts | gsicc_open_search has an unchecked alloc in it. | 18:56.50 |
| I can check it and return NULL, but then does that function returning NULL mean something? | 18:57.19 |
| I think NULL means 'not found'. | 18:58.02 |
| mvrhel_laptop: How much of an upset would it be to change gsicc_open_search to return an error code, and have the pointer return as a param? | 18:58.39 |
| (I'll make the change, just want to know whether there are docs that need updating etc) | 18:58.59 |
| oh, it's a static, so no problems. sorry! | 18:59.46 |
mvrhel_laptop | Robin_Watts: ok | 19:00.01 |
henrys | don't we get compiler warnings when assigning an int to a short? I've spent an hour or so chasing what should be a warning. sigh | 19:16.44 |
Robin_Watts | henrys: On some compilers. | 19:16.58 |
| mvrhel_laptop: Do you want to see what I've done to your code before I commit it? | 19:19.04 |
mvrhel_laptop | Robin_Watts: it depends upon what you have done | 19:19.26 |
| ;) | 19:19.30 |
Robin_Watts | http://git.ghostscript.com/?p=user/robin/ghostpdl.git/.git;a=commitdiff;h=045235372355a640bdf637a6065bfc8597e00b8e | 19:19.51 |
| :) | 19:19.53 |
mvrhel_laptop | ok this is sticking the stream in as a parameter. looks fine | 19:22.08 |
Robin_Watts | that and a couple of other checks elsewhere. Thanks. | 19:22.29 |
mvrhel_laptop | lunch time . bbiab | 19:24.48 |
ray_laptop | mvrhel_laptop: In case you are interested, there is a problem that while playing back a pattern-clist that has transparency, if an error happens in the clist playback, we may have executed some compositor actions (begin_transparency_mask in this case) and the ctx->stack is not correct | 19:26.19 |
| mvrhel_laptop: this is a weakness in the pdf14_tile_pattern_fill but NOT the underlying cause. | 19:27.13 |
Robin_Watts | ray_laptop: That sounds "interesting" in the same sense that illnesses or car problems are "interesting". | 19:27.48 |
| You know it's going to cause you pain of some form... | 19:27.59 |
ray_laptop | mvrhel_laptop: so, if you want to look at how to make the recovery more robust, fine, but not really needed. I am now focusing on what caused the error in the first place | 19:28.33 |
| Robin_Watts: hehe | 19:28.58 |
| Robin_Watts: or the curse I've heard attributed to ancient Chinese: May you live in interesting times | 19:29.33 |
Robin_Watts | yeah. | 19:29.39 |
| I have an "interesting" car problem at the moment. I'd really rather have a boring car at times. | 19:30.25 |
ray_laptop | the boring parts of the history books might be less stressful | 19:30.29 |
Robin_Watts | wonders how many more of these valgrind bugs there are to come. | 19:46.01 |
ray_laptop | Robin_Watts: yeah, my email has been flooded :-( | 19:47.47 |
sebras | Robin_Watts: mupdf or gs? | 19:49.23 |
| oh... wow. marcosw has been busy. | 19:50.37 |
henrys | Ah -Wconversion on gcc, who wants 10,000 more warnings? | 19:51.42 |
Robin_Watts | sebras: Indeed. | 19:57.15 |
| mvrhel_laptop: gsicc_set_device_profile... | 19:57.25 |
| at line 1654 it calls gsicc_get_profile_handle_buffer - should that returning NULL be an error? | 19:58.01 |
| yes, it's an error, cos lots of functions after that fail if it's null. | 20:04.09 |
| 1564, sorry | 20:05.33 |
mvrhel_laptop | ok back from lunch | 20:05.42 |
| Robin_Watts: yes, that should really be treated as an error | 20:10.23 |
henrys | that's a lot of valgrind reports ;-( | 20:39.42 |
ray_laptop | 92 new ones (4 old ones still have "grind" in the summary) | 20:52.17 |
| bug 693732 through bug 693823 -- Is marcosw finally done ? | 20:53.08 |
| should make it easy to meet my bug fix quota each week ;-) | 20:54.07 |
Robin_Watts | mvrhel_laptop: gssicc_alloc_link_entry... | 21:05.57 |
| (gsicc_cache.c line 661) | 21:06.15 |
mvrhel_laptop | Robin_Watts: hold on | 21:06.22 |
Robin_Watts | If the allocation fails, we don't check for it. | 21:06.25 |
| I can add a check, but what should it do? Can I safely return false there? | 21:06.43 |
| (unlock and return false) | 21:06.56 |
mvrhel_laptop | Robin_Watts: yes that makes sense to do | 21:07.38 |
Robin_Watts | Do callers know that 'false' may mean "failed to allocate one" ? | 21:09.13 |
| no, at least one will SEGV. | 21:10.14 |
| That'll do me for the night. | 21:17.48 |
| Forward 1 day (to 2013/03/29)>>> | |