| <<<Back 1 day (to 2013/02/10) | 2013/02/11 |
kens | chrisl ping | 09:08.11 |
| actually, forget that | 09:08.42 |
chrisl | kens: forget what? :-) | 09:19.05 |
kens | THe ping I just put out, but if you are free.... | 09:19.18 |
chrisl | Sure, just getting coffee | 09:19.30 |
kens | I'm having some trouble diagnosing a problem with a type1->CFF conversion | 09:19.36 |
| The problem seems to occur because the first path operation is a rlineto, not a moveto | 09:20.18 |
| type 2 says the first opertation shall be a moveto | 09:20.24 |
| So we special case check this (gdevpsfx.x, line 568 or so) | 09:20.44 |
| and emit a vmoveto or hmoveto | 09:21.10 |
| What we actually emit is "374 0 vmoveto" which I think is invalid | 09:22.33 |
chrisl | vmoveto/hmoveto only take one parameter, don't they? | 09:22.34 |
kens | I don't think kyou are allowed to bundle up hmovetos, what would be the point ? | 09:22.34 |
| By the way, this is bug #693633 if you want to look at the source file | 09:23.02 |
chrisl | It does look like you're not supposed to have multiple consecutive v/hmovetos.... | 09:26.52 |
kens | THat's what I though | 09:28.15 |
| thought | 09:28.18 |
| So, what I need to do is convert the rlineto into a 'moveto and a rlineto where one of the initial relative movements (x or y) is 0 | 09:28.55 |
| And I think that's what is screwing up | 09:29.15 |
| I think its emitting the 'x' relative moevement as one of the arguments to the hmoveto. | 09:29.47 |
| let me see if I can quickly hack this | 09:29.57 |
| Well, I'm getting a corrupted glyph, but at least Acrobat isn't complaining :-) | 09:31.34 |
| SO I'm on the right lines, thanks chrisl | 09:31.43 |
chrisl | NP, I'm wondering if the lack of a moveto "implies" a "0 0 moveto", or "origin_x origin_y moveto" or something | 09:32.44 |
kens | I think in type 1 the implicit origin is 0,0 in type 2 there is no implicit initial point | 09:41.59 |
| The problem (probably) is that this font is converted from TT to type 1, and so is not necessarily of high (or even decent) quality) | 09:42.45 |
chrisl | That's what I'm thinking - if a Type 1 glyph relies on an implicit initial point, when converting to T2 it might need an explicit start point | 09:43.20 |
kens | Type 2 has an implicit start, but all charstrings are required (apparently) to start with a 'movet', not a lineto | 09:44.31 |
| type 1 doesn't seem to have this restriction. In fairness, its possible type 2 interpreters would cope too, the code in gdevpsfx.c is, bluntly, wrong. | 09:45.04 |
| I see now that it is assuming that there will be the glyph 'width' (from hsbw) on the stack, but if we have any hints this will *not* be true. | 09:45.32 |
| I'm guessing that someone added hint processing and missed this little assumption in the code. | 09:45.47 |
| (very possibly me) | 09:46.01 |
| And that fix works perfectly. | 09:46.35 |
| Now to cluster push it. | 09:46.43 |
chrisl | Cool, sorry phone call - and I need to go out for a while in a minute. Sounds like you've found the problem, though | 09:51.51 |
| In fact, now - bbiaw.... | 09:52.05 |
kens | No problem, I understadn the (baroque,impenetrable) code better now, thanks | 09:52.09 |
| chrisl (for when you get back) I am unable to compile on the cluster, and I don't think its my code. I'm getting: | 10:29.39 |
| collect2: ld returned 1 exit status | 10:29.39 |
| make: *** [bin/gs] Error 1 | 10:29.39 |
| "/usr/bin/ld: cannot find -lcupsimage | 10:29.59 |
| "/usr/bin/ld: cannot find -lcups | 10:29.59 |
| "/usr/bin/ld: cannot find -lgssapi_krb5 | 10:29.59 |
| "/usr/bin/ld: cannot find -lgnutls | 10:29.59 |
| "collect2: ld returned 1 exit status | 10:29.59 |
| "make: *** [bin/gs] Error 1 | 10:30.00 |
| THis 'appears' to be only a problem on alex_x6 (I htink) | 10:30.30 |
| SoI've disabled it and am trying again | 10:30.54 |
| And I can't take it out of the cluster, I get taken to a page which says it failed | 10:31.51 |
| 'touch /home/regression/cluster/alex_x6.down': Failed! | 10:32.04 |
| If anyone has any ideas, I'd be grateful...... | 10:32.19 |
Robin_Watts | kens: I can take it out. | 10:55.09 |
kens | please robin | 10:55.19 |
Robin_Watts | done. | 10:55.51 |
kens | thanks I'll try again | 10:56.36 |
| still fails on alex_x6 | 11:15.42 |
Robin_Watts | Sorry. I touched alex_x6.done, not alex_x6.down :( | 11:23.13 |
| Try again now. | 11:23.15 |
| Crumbs: http://www.bbc.co.uk/news/world-21411304 | 11:26.53 |
kens | ok | 11:27.01 |
| that looks better Robin_Watts thanks | 11:29.51 |
Robin_Watts | np. | 11:29.56 |
chrisl | I wonder why alex_x6 should suddenly do that..... wonder if it's had a dist-upgrade? | 12:01.46 |
Robin_Watts | marcosw updated a couple of his nodes the other day. Maybe alex followed suit. | 12:17.33 |
chrisl | My experience is that "apt-get dist-upgrade" isn't great at updating the -dev packages.... | 12:18.46 |
kens | I don't know what's wrong exactly, but that node won't build GS | 12:19.00 |
| Off for a haircut, back soon | 12:19.11 |
chrisl | Robin_Watts: are you staying overnight up in London for the Wall in September? | 12:19.54 |
Robin_Watts | chrisl: Probably not. | 12:20.05 |
| but tor8/sebras might be. | 12:20.14 |
chrisl | Are you driving up, then? | 12:20.32 |
Robin_Watts | And there is a possibility that Scott and his 11 year old grandson Austin may come. | 12:20.40 |
| chrisl: I suspect I'll take the train. | 12:20.48 |
chrisl | Robin_Watts: Well, count me in - I'll sort out the logistics nearer the time! | 12:21.11 |
Robin_Watts | Train (Banbury to Marylebone), then Baker Street -> Wembley on the tube. | 12:21.19 |
| And that way we can meet up for food before hand maybe. | 12:21.33 |
| Seated tickets are 82.50 each. I assume that's OK. | 12:21.56 |
| Standing are 66 quid. | 12:22.05 |
chrisl | Yeh, seated is good | 12:22.11 |
Robin_Watts | Or you can get a corporate hospitality package at 395 per person. | 12:22.27 |
| but that includes champagne and blowjobs(*). | 12:22.37 |
| (* May not actually contain blowjobs) | 12:22.44 |
chrisl | Would Miles pay?? ;-) | 12:22.45 |
Robin_Watts | I doubt it,sadly :) | 12:22.57 |
chrisl | Anyway, seated will be fine | 12:23.21 |
Robin_Watts | cool. | 12:23.24 |
| kens, paulgardiner: In case either of you are interested... we are going to see "The Wall" at Wembley stadium on 14th September. | 12:23.59 |
| I spoke to Miles, and he says our staff meeting will be 6/7th September in Chicago, as there is the Print '13 tradeshow from the 8-12th afterwards. | 12:24.48 |
chrisl | I doubt it would be a problem - we've excluded plenty of dates for staff meeting due to folks' other commitments. | 12:26.18 |
Robin_Watts | chrisl: I was prepared to take the risk, but Miles has confirmed it's OK, so that's even better. | 12:26.44 |
chrisl | Hmmm, IIRC, when I was 11 years old, the complete video for The Wall was still banned........ | 12:29.34 |
Robin_Watts | The video or the film? | 12:29.53 |
chrisl | Well, the video of the film | 12:30.07 |
| I think you could get a censored version | 12:30.32 |
Robin_Watts | I have the 25th anniversary DVD release I think. | 12:31.17 |
chrisl | I have the performance Waters did in Berlin on a double LP, but the audio production isn't great | 12:32.28 |
Robin_Watts | The one at the brandenburg gate after the Berlin Wall fell? I have that too. | 12:32.55 |
chrisl | Yes, that's the one | 12:33.11 |
Robin_Watts | The fact that the power failed 4 times during the concert makes the fact they managed to get a recording out at all fairly surprising :) | 12:33.27 |
chrisl | Must have been quite a show to attend | 12:34.00 |
Robin_Watts | Yeah. "Are all these your guitars?" is the role Jerry Hall was born to play ? | 12:34.31 |
| s/?/!/ | 12:34.41 |
chrisl | I suspect with the venue, and everything else it would count more as "an experience" than a concert | 12:35.40 |
sebras | Robin_Watts: yes, I suspect we'd stay at least over night. we haven't synced up about it yet though. :( | 12:41.17 |
| oh, and I wonder... should it be Wall or -Wall -Wextra...? | 12:42.36 |
Robin_Watts | would offer a room, but I'm a long way out of london (and I only have 1 spare room). | 12:44.00 |
sebras | I'm trying to contact tor8 to be able to sort out the details tonight. | 12:46.16 |
tor8 | we'll def want to spend the night somewhere | 12:48.07 |
henrys | good morning - sounds like an awesome event at Wembley | 15:03.01 |
Robin_Watts | henrys: come see it! | 15:03.12 |
henrys | is Scott really going? | 15:03.20 |
Robin_Watts | henrys: He's looking into it now. | 15:03.30 |
| but he gets free flights of course. | 15:03.35 |
henrys | that would be a bit more than I'm willing to pay for a concert | 15:03.50 |
| I could have seen the original one when I was a kid and didn't go, I regret it, I did manage to see the The Who and The Clash at Shea Stadium in that same era | 15:07.43 |
| christ I'm old ;-) | 15:13.12 |
Robin_Watts | hehe | 15:13.21 |
kens | I like today's XKCD | 15:23.31 |
henrys | kens:sounds familiar | 15:28.12 |
JakeSays | so i have a type1 font descriptor that defines this charset: /CharSet (/L/A/M/Y/N/C/O/D/P/E/F/R/S/H/T/I/U/space) | 15:29.29 |
| and this line: /R11 9 Tf | 15:29.44 |
| if i understand the spec, Tf sets the current font to object 11, size 9? | 15:30.07 |
Robin_Watts | No. | 15:30.28 |
| It sets the current font to the font object named R11 in the current resources dictionary(ies) | 15:30.53 |
| at size 9. | 15:31.05 |
JakeSays | isnt /R a reference to an object? | 15:31.28 |
Robin_Watts | tor8, paulgardiner: Updated reflow review on robin/master | 15:31.40 |
| About to look at the paul/master reviews now. | 15:31.49 |
henrys | Robin_Watts: nothing from raph I take it. | 15:32.04 |
Robin_Watts | tor8: And the pass by reference stuff still needs review. | 15:32.09 |
| henrys: Nope :( | 15:32.13 |
| JakeSays: No. | 15:32.15 |
JakeSays | bah. no wonder nothing is making sense | 15:32.30 |
Robin_Watts | I think you may be confusing stuff with "11 0 R" (which is a reference to object 11 generation 0). | 15:32.56 |
JakeSays | yeah | 15:33.21 |
| i think you're right | 15:33.26 |
| Robin_Watts: ok so this: /R11 10 0 R indicates R11 is a ref. to 10 0? | 15:36.09 |
tor8 | Robin_Watts: I believe the reference stuff is good to go | 15:37.33 |
Robin_Watts | tor8: There are about 7 reviews on robin/master, of which the reference one is number 3. | 15:41.16 |
| Are all 7 good to go? (The last one being the android fixes) | 15:41.29 |
tor8 | Robin_Watts: and the other stuff on robin/master too | 15:41.34 |
Robin_Watts | Thanks. | 15:41.39 |
| Pauls look fine to me too. | 15:41.45 |
tor8 | I've got two on tor/master for apple fixes. don't push those to origin yet though, I may figure out what's causing crashes and what's broken in xcode eventually | 15:42.26 |
paulgardiner | Robin_Watts: do the characters of a span no longer necessarily share a common style? | 15:43.40 |
Robin_Watts | indeed. | 15:44.18 |
JakeSays | a span as in the spans in the xml output? | 15:48.04 |
Robin_Watts | JakeSays: indeed. | 15:48.40 |
JakeSays | oh hmm. that would mess things up for what i'm currently doing | 15:49.29 |
Robin_Watts | JakeSays: how so? | 15:49.54 |
JakeSays | Robin_Watts: i depend on a span being a single style | 15:50.33 |
| although | 15:50.55 |
| hmm. i'll have to think about it. it may actually help | 15:51.20 |
| Robin_Watts: my app does text extraction to convert printed statements in to their original data | 15:51.59 |
| i do things like "look for a span of text that is in font x, size y, at position z | 15:52.33 |
| and use that as an anchor | 15:52.41 |
Robin_Watts | The only real difference is that the style pointer is now a per char thing. | 15:54.06 |
| (and the bbox for chars is calculated rather than being stored in each one) | 15:54.39 |
| spans and lines have bboxes. | 15:54.48 |
JakeSays | do you know what the relationship of the values in a /CharSet is to glyph ids? | 15:56.41 |
kens | What do you mean by glyph ID ? TYpe 1 fonts don;t have glyph IDs, they have anmes | 16:00.26 |
| names* | 16:00.36 |
JakeSays | kens: ah well that explains why mapping gids to that charset array isn't making sense | 16:01.37 |
kens | CharSet is only valid for type 1 fonts | 16:01.59 |
JakeSays | yeah i have a type 1 font | 16:02.12 |
kens | OK so type 1 fonts are encoded. | 16:02.23 |
paulgardiner | Robin_Watts: you also have a change to strings.xml that you might not have intended for this commit. | 16:02.24 |
Robin_Watts | I think I've removed that too. | 16:02.36 |
kens | The 'text' is actually an index into the ENcoding array | 16:02.39 |
| THe Encoding array gives you the glyph name corresponding to that index | 16:02.55 |
paulgardiner | Robin_Watts: I thought you might of, but... | 16:02.56 |
kens | The CharSet just tells you which glyphs in a font are used. | 16:03.15 |
JakeSays | ah ok | 16:04.52 |
kens | Err, actually it tells you which glyphs are 'available' I think | 16:05.20 |
| ie which ones are present in the font | 16:06.00 |
JakeSays | kens: the encoding array being /Encoding? | 16:06.35 |
| or defined in /Encoding | 16:06.39 |
kens | "/Encoding" is a name. | 16:07.40 |
| THe font dictionary contains a key/vlaue pair, the key is the name "/Encoding" its value is an array | 16:08.01 |
JakeSays | right | 16:08.08 |
| thats what i meant | 16:08.12 |
| the array defined by /Encoding | 16:08.23 |
kens | Well its not 'defined' by /ENcoding | 16:08.53 |
JakeSays | in this particular pdf, i have /Encoding /WinAnsiEncoding | 16:09.26 |
| i'm assuming that means the encoding array is the standard win ansi one | 16:09.46 |
kens | The font dicitonary contains a key with that name, and the array (which defines 'the encoding') is the value associated with that key | 16:09.51 |
| WinAnsiEncoding is defined in the PDF spec | 16:10.04 |
| A PDF interpreter should contain a copy of the WinAnsiEncoding array which it canuse | 16:10.36 |
JakeSays | so whats the proper terminology? when i say defined by /Encoding, i mean the /Encoding dictionary entry's value indicates or contains the encoding. | 16:11.13 |
kens | I woudl normally say the encoding array (not /Encoding) | 16:12.36 |
JakeSays | how do you determine the encoding array? | 16:13.28 |
| actually, i have a feeling for what i'm doing none of this should really matter to me. | 16:14.38 |
| i suspect i had a bad pdf that has sent me down a rat hole | 16:14.51 |
kens | THis is not uncommon | 16:15.19 |
Robin_Watts | tor8: I've updated the commit on my reflow branch with pauls suggestions. | 16:50.47 |
henrys | sorry I was rearranging stuff resulting in henrysx6 and Mac Pro going down they should be back soon. | 17:06.12 |
mvrhel_laptop | henrys; I am going to out this week wed through friday. I will send out an email | 17:22.22 |
| off to whistler for a few days | 17:22.38 |
henrys | oh cool have fun | 17:22.48 |
mvrhel_laptop | got myself a helmet cam. I think my son is going to be wearing it most of the time though | 17:23.27 |
henrys | mvrhel_laptop: I just bought a plane ticket to canada how is that just flying over the border seems to double the airfare? | 17:24.34 |
mvrhel_laptop | yes. it is weird. | 17:25.06 |
| where and when are you going? | 17:25.10 |
| henrys^^ | 17:25.17 |
henrys | mvrhel_laptop: penticton, august | 17:25.35 |
| good snow at whistler - we finally got some hear it has been fairly dry | 17:26.08 |
| ? | 17:26.13 |
mvrhel_laptop | I think they have had a lot of snow there henrys. Is Penticon going to be a race. I have not been over in that part of BC I bet it is beautiful | 17:27.16 |
henrys | yes challenge penticton - it's an ironman like event | 17:27.50 |
mvrhel_laptop | nice | 17:27.56 |
henrys | actually the race is put on by challenge who is an ironman competitor | 17:28.21 |
| anyway I could have flown to spokane 100 or so miles south for half the price. crazy | 17:28.49 |
mvrhel_laptop | 5 -6 hour drive from here. just looking at it | 17:28.54 |
| bellingham might have been ok to fly into | 17:29.16 |
| but either way it is a drive | 17:29.44 |
| are you flying right into Penticton? | 17:30.01 |
henrys | yeah I just paid it's my first time doing this kind of thing and I didn't want to fool with transportation | 17:30.28 |
mvrhel_laptop | right | 17:30.33 |
henrys | can't wait to see the helmet cam flicks! | 17:31.12 |
mvrhel_laptop | hehe. I will post a few someplace | 17:31.50 |
| I really want to use the thing windsurfing this summer | 17:32.07 |
henrys | I guess they are fairly weather-proof | 17:32.27 |
mvrhel_laptop | yes. it comes is a waterproof case | 17:32.43 |
| and will do video at various resolutions or timelapse photos, or photo bursts | 17:33.02 |
henrys | I ride my bike fast down canyons here and I wanted to try it out too. Lots of folks use them here. | 17:33.17 |
mvrhel_laptop | yes. bike riding would be good to use too. just need to edit out the part of me pushing my bike uphill ;^) | 17:33.46 |
| or breaking something.... | 17:34.20 |
Robin_Watts | mvrhel_laptop: Hero? | 17:40.11 |
mvrhel_laptop | Hero 2. Yes Robin_Watts | 17:40.24 |
Robin_Watts | nice. It's the one that Stuff magazine consistently recommends. | 17:42.42 |
henrys | I was going to mention at the meeting we should announce days off To: Joann and CC:staff. I think that's easy and consistent | 17:55.48 |
| bbiab | 18:04.37 |
Robin_Watts | http://ghostscript.com/~robin/MuPDF-8.apk <- New version of mupdf. Better paragraph detection in reflow mode. | 18:17.50 |
JakeSays | kens: bah. yeah. that one bad batch of pdfs consumed two weeks of my life :( | 18:49.52 |
marcosw | Except for the one UFST with PS/PDF issue that Chris points out none of our customer are using, there are were no issues found in 9.07rc1 so I recommend we ship it. There are some performance regressions, but some of these are due to increase in output correctness and other I can't duplicate. There are some ones that should be addressed, but I don't think we need to hold up the release. | 18:54.38 |
henrys | marcsow:did you have a chance to look at some of the extreme pcl slowdowns? Was that an error or something - all the files I looked at were very simple, shouldn't have changed | 18:56.36 |
marcosw | alexcher: pong | 18:56.49 |
| henrys: many of them I can't duplicate. I don't understand why the occurred on the performance regression machines and am continuing to investigate. | 18:57.35 |
| I plan on opening a bug for the ones I can duplicate. | 18:57.53 |
henrys | marcosw:okay thank you | 18:57.56 |
alexcher | marcosw: I think I've fixed alex_x6. | 19:00.22 |
marcosw | alexcher: you had to reinstall the cups dev libraries after a linux update? | 19:00.47 |
mvrhel_laptop | henrys: resent my email to joann with a cc to staff as you suggested | 19:00.57 |
| much better that way. last time I was in canada scott called me and wanted to chit chat.... | 19:01.21 |
henrys | mvrhel_laptop: oh that's why you copied him in explicitly | 19:01.48 |
mvrhel_laptop | hehe yes | 19:01.53 |
alexcher | marcosw: I've tried to install 32-bit libraries and broke 64-bit ones. | 19:02.16 |
mvrhel_laptop | in the past I have just sent an email to joann and one to tech | 19:02.17 |
marcosw | alexcher: sounds reasonable :-) | 19:02.33 |
chrisl | henrys: there may be some inefficiencies in the PCL/FAPI integration - I was more focussed on getting the correct output at the time...... | 19:04.48 |
| marcosw: given the isolated nature of the change needed to fix the UFST regression, I wouldn't be averse to pulling it in - it only effects outline glyphs. | 19:06.12 |
henrys | chrisl: well marcosw can't reproduce any of it. | 19:06.22 |
marcosw | henrys: I can reproduce some of the performance issues. | 19:07.01 |
| chrisl: you point out that no one is using UFST with PS/PDF, so I don't see any point. | 19:07.29 |
chrisl | marcosw: fair enough - I do think I should fix it soon, though..... | 19:08.21 |
henrys | if you did anything that would disable caching I can imagine the sort of slowdown I'm seeing, but we'll wait for marcosw bug reports, easy to see with a profile | 19:08.41 |
marcosw | chrisl: if no other reason that it makes the weekly regression testing easeri. I agree. | 19:08.45 |
chrisl | marcosw: for the performance regressions in PCL, you can isolate it to the FAPI revamp using the -dDisableFAPI option (for now.....) | 19:09.27 |
| marcosw, henrys: so we go for release tomorrow? | 19:10.13 |
marcosw | chrisl: yes. | 19:10.27 |
henrys | I vote yea | 19:10.34 |
chrisl | Cool :-) | 19:10.42 |
| henrys: BTW, I'm taking Wednesday (mostly) off - I sent the mail to Joann before I saw your comment.... I'll mail tech in a moment | 19:11.37 |
henrys | so should we hold off until thursday for the release? | 19:12.19 |
| release and run ;-) | 19:12.32 |
chrisl | I can do that - I will be around for a while Weds morning, but I don't about later in the day | 19:13.34 |
henrys | I'd rather you be available a few days after the release so thursday is good for me, but I don't feel strongly about it. | 19:14.52 |
chrisl | It'll be a little easier for me - I can update the docs and stuff tomorrow, and actually release it to the world on Thursday | 19:15.32 |
| So Thursday it is! | 19:15.55 |
mvrhel_laptop | bbiaw | 20:01.39 |
ray_laptop | we had a power outage all around the area until recently. As a result my home, office and my mother-in-laws house were all down. It looks like peeves came back up, but didn't recognize the monitor, so all I can get is 1280x1024 :-( | 21:04.12 |
| trying a reboot of it now | 21:04.23 |
| sure is TAKING a while to shut down peeves :-( How long does it wait after "asking all remaining processes to terminate" ? | 21:06.59 |
| OK. it shut down. coming back up... | 21:08.40 |
Robin_Watts | ray_laptop: Test mupdf while you wait :) | 21:08.58 |
| http://ghostscript.com/~robin/MuPDF-8.apk <- New version of mupdf. Better paragraph detection in reflow mode. | 21:09.03 |
ray_laptop | Robin_Watts: don't have my tablet here (just my phone). Besides, my laptop is still up :-) | 21:09.44 |
Robin_Watts | well, when you do get back to tablet range, I'd appreciate it if you could give it a quick test. | 21:10.16 |
henrys | so this reflow is enabled with the lines with a bend on the right yes? | 21:10.21 |
Robin_Watts | Yeah. | 21:10.29 |
| They are supposed to be arrows bending downwards. | 21:10.54 |
ray_laptop | ahh... that's better. Have a full res screen now | 21:11.21 |
henrys | okay so pages are resizing when I do that, I flow then go to the next page and the background white part of the page is a different size sometimes shrinking down to a couple lines | 21:11.31 |
Robin_Watts | yup. | 21:11.38 |
| Each page is still kept as a page, but the html page produced is only large enough for the text on it. | 21:12.01 |
| You can zoom as usual. | 21:12.09 |
henrys | ugh | 21:12.30 |
Robin_Watts | ugh? | 21:12.42 |
ray_laptop | henrys: were you looking for 'continuous' reflow (across pages ??) | 21:13.22 |
| hurray -- even the regression monitor sees peeves as 'up' | 21:13.59 |
Robin_Watts | ray_laptop: The roadmap we did makes it clear that that's a separate item. | 21:14.01 |
ray_laptop | Robin_Watts: sorry. | 21:14.14 |
henrys | it just looks odd | 21:14.19 |
Robin_Watts | ray_laptop: No worries. | 21:14.22 |
| henrys: "Looks odd" how? | 21:14.30 |
ray_laptop | I just didn't understand henry's "ugh" | 21:14.32 |
Robin_Watts | Looks odd as it zooms ? | 21:14.37 |
| ray_laptop: Me either :) | 21:14.40 |
| or the extracted text looks odd ? | 21:14.49 |
| Looking odd as it zooms - well, that's androids WebView for you. | 21:15.10 |
| The extracted text looking odd - what page are you looking at of what document ? | 21:15.56 |
henrys | I have a document the entire background is white page 1 - next page the white is a small narrow strip with 3 or 4 lines inside | 21:16.05 |
Robin_Watts | And how would you rather it was laid out ? | 21:16.38 |
henrys | I didn't expect the background to be effected at all just the text | 21:18.05 |
Robin_Watts | You expected the original page sizes to be kept? | 21:18.27 |
| How does that work when you zoom? | 21:18.37 |
kens | OK I'm off, night all | 21:18.52 |
Robin_Watts | We're extracting the text, and throwing everything else away currently. | 21:19.01 |
| The background is just "a blank html page big enough to hold the text". | 21:19.21 |
henrys | okay I think that most people would be surprised if the page size changes dramatically each page, maybe I'm wrong | 21:22.38 |
Robin_Watts | We could probably enforce a minimum page height for each page. | 21:23.32 |
henrys | What do you think about a static page size, if reflow goes over the page - a scrollbar pops up or is that very difficult to do? | 21:25.54 |
Robin_Watts | henrys: It's a webview. | 21:26.29 |
| All we are doing is showing a webpage. | 21:26.39 |
| and we've got hooks in there that set the 'zoom' factor of the webpage. | 21:26.56 |
| and we fix the width of the webpage. | 21:27.05 |
| so, we could probably set the webpage to have a minimum height (of the size of the screen, maybe?) | 21:27.43 |
| Then as you zoom, you'd start to be able to move up and down. | 21:27.57 |
henrys | it may be just disturbing to me. Let's see what others say I don't want to waste time if I'm the only one concerned. | 21:28.48 |
Robin_Watts | Unless we want to change our view mode fundamentally, we want to keep the 1:1 relationship between reflown and non-reflown text. | 21:31.05 |
| s/text./pages./ | 21:31.31 |
| henrys: Well, it would be good to gather opinions from ray/michael/anyone else with an android device. | 21:36.44 |
| Certainly I find our reflow behaves more like I expect than Repligo's or Adobe Readers. | 21:37.10 |
| OK, there is a min-height css thing. | 21:48.55 |
henrys | adobe does not reflow the entire document? | 21:51.48 |
Robin_Watts | Adobe works on a page by page basis too. | 21:52.05 |
henrys | of course they will support reflow tags | 21:52.47 |
Robin_Watts | but if it reflows a source page into 2 pages, the second will be half full, I believe. | 21:52.54 |
| So, take pdf_reference17.pdf for example | 21:53.39 |
| if I look at page 26/1310, it reflows that onto 3 screenfuls. | 21:54.03 |
| which I get to by sweeping left, so they appear as 3 reflown pages. | 21:54.27 |
| The last of these reflown pages just has a single line on. | 21:54.43 |
| well, 2 lines. | 21:54.48 |
| wheras, if it was truly reflown you'd expect the content on page 27 to continue immediately. | 21:55.15 |
| Our keeps the 1:1 page correspondence, so is nicer (IMHO), and you can page through by just tapping the bottom (or right) edge of the screen | 21:56.22 |
henrys | Robin_Watts: and why not one page with a scroll bar? | 21:59.26 |
| bbiam | 22:02.13 |
henrys_ | looks like we are pretty close to getting fiber in my town ⦠yeah! | 22:54.42 |
sebras | henrys: google fiber? how fast is that? | 23:32.42 |
henrys | sebras:unclear they are talking 200 mbs I get 20 now so that would be a nice improvement for me | 23:34.15 |
| Forward 1 day (to 2013/02/12)>>> | |