| | <<<Back 1 day (to 2014/06/24) | 2014/06/25 |
| kens | Great, another customer who thinks that a simple font substitution investigation is helped by supplying 3 pages of text | 07:18.41 |
| chrisl | My guess is that they haven't accounted for the files being in romfs | 07:19.11 |
| kens | You could be correct, I haven't looked yet, got another twit to deal with overnight as well. | 07:19.33 |
| | Well when I run the file here, the fonts they complain about Airal-BoldMT and TimesNewRomanPSMT are embedded in the PDF file. In fact they are the *only* fonts embedded in the PDF file..... | 07:21.23 |
| chrisl | Oh, well maybe they think font substitution will forcibly replace fonts *in* the file - doh! | 07:22.41 |
| kens | It seems like that's what they expect...... | 07:22.53 |
| | Certainly thre's a CIDFont named Arial (with a subset prefix) embedded in the file, still looking to see which others are there | 07:23.24 |
| chrisl | I can imagine why they think it's a good idea, but they're going to be disappointed....... | 07:23.36 |
| kens | THey certainly are. | 07:23.42 |
| | Looks like two different embedded CIDFonts called Arial | 07:23.59 |
| | Ah and there's an Arial-BoldMT as well | 07:24.36 |
| | And a TimesNewRoman | 07:25.11 |
| | Given that this is the WIndows PostScritp driver output, I'm not surprised to find the fonts embedded. | 07:25.34 |
| | Ah and I see that pdf_info.ps does actually already print the value of the Rotate key for each page.... | 07:30.44 |
| pedro_mac | hi folks | 08:12.44 |
| kens | Morning | 08:12.51 |
| Robin_Watts | morning | 08:16.16 |
| pedro_mac | Robin_Watts: mvrhel seems to be doing well with the chart stuff | 08:27.52 |
| Robin_Watts | yeah. | 08:35.35 |
| | pedro_mac: mvrhel found this last night. Looks like a lifesaver. | 09:19.40 |
| | http://visualstudiogallery.msdn.microsoft.com/450a00e3-5a7d-4776-be2c-8aa8cec2a75b | 09:19.42 |
| jogux | really needs to get my windows build environment up and running | 09:20.35 |
| pedro_mac | Robin_Watts: cool - wonder if its better than the openxml SDK package editor | 09:24.00 |
| pedro_mac | downloads | 09:24.07 |
| Robin_Watts | Once you load a piece of XML into a window, it will all appear on 2 lines. | 09:24.52 |
| | click on the second line, Ctrl-A to select it, then Ctrl-K, Ctrl-F to format it. | 09:25.21 |
| | (That's the same as "Edt -> Advanced -> Format Selecton") | 09:25.54 |
| pedro_mac | cool, quite nice | 09:26.44 |
| Robin_Watts | henrys: insomnia? | 09:27.49 |
| mattchz | hi | 09:32.32 |
| ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 09:32.32 |
| pedro_mac | Robin_Watts: thatâll save me using explorer to view the XML ;) - Iâve been using the OpenXML SDK Productivity Tool. It does a hierarchy of documents/tags and does code reflection (not what weâre after, but sometimes makes the construction of the doc more comprehensible). It also has documentation for the ooxml | 09:32.39 |
| Robin_Watts | pedro_mac: Ah, well, that might be worth me getting. | 09:33.02 |
| | mattchz: You picked a good day to be away yesterday. I hit a customer problem (we think) to do with Image component subsampling :) | 09:33.34 |
| mattchz | heh, sorry | 09:33.46 |
| | did you sort it out? | 09:33.55 |
| Robin_Watts | no worries. I think it's just a question of the static subsampling threshold being set too low. | 09:34.23 |
| mattchz | ah. | 09:34.34 |
| jogux | robin_watts: I thought you decided both builds probably had the same setting? | 09:34.42 |
| Robin_Watts | dynamic subsampling sounds posh, but it's basically "don't statically subsample", right? | 09:34.48 |
| | jogux: No, I was being dim. | 09:34.58 |
| mattchz | There's a dynamic subsampling deluxe thing iirc | 09:35.04 |
| | with partial decode. | 09:35.14 |
| jogux | robin_watts : ah, d'oh :) | 09:35.14 |
| pedro_mac | Robin_Watts: http://twiki.ghostscript.com/do/view/GhostDocs/OfficeDocumentation | 09:35.21 |
| Robin_Watts | Copying a static lib into a test project that expects a .so and rebuilding, strangely doesn't magically use the new lib. | 09:35.36 |
| mattchz | iirc, basic dynamic subsampling is 'try each level until we find one we are able to decode'. | 09:35.45 |
| jogux | robin_watts : ah :-) | 09:35.56 |
| Robin_Watts | so one of theirs *is* the fvdk, and the other is pageview. | 09:36.06 |
| mattchz | it's basically broken, as we don't know if we're actually going to have enough RAM to render the decoded image anyway. | 09:36.09 |
| | IIRC | 09:36.10 |
| | this is all very hazy. | 09:36.15 |
| | In fact, I think the only reason we didn't remove it was because some obscure customer used it | 09:36.43 |
| Robin_Watts | Right. | 09:36.50 |
| mattchz | the new, partial decode thing is much better though. | 09:37.01 |
| | although it can perform badly, due to lack of caching. | 09:37.12 |
| Robin_Watts | we have partial decode? | 09:37.17 |
| mattchz | Yeah | 09:37.33 |
| | for JPEG | 09:37.35 |
| pedro_mac | grafted from NTP iirc | 09:37.58 |
| mattchz | I can't remember if we ever managed to implement it for any other formats. | 09:37.59 |
| | ped> not entirely. Possibly some of it was (although I can't fully remember) | 09:38.19 |
| pedro_mac | donât think we did | 09:38.21 |
| mattchz | I spent months working on it for JPEG. | 09:38.33 |
| | (this was not long before I left, although before the many miserable months working on partial layout) | 09:39.04 |
| Robin_Watts | While I'm waiting for the customer to get back to me, I've been looking at something else I spotted in their test doc. | 09:39.55 |
| | They have a 'star12' autoshape, with the word "Done" inside it. | 09:40.10 |
| | Both us and LibreOffice display the Done in the top level corner. | 09:40.29 |
| | whereas MS displays the Done in the middle. | 09:40.41 |
| | If I change the autoshape to 'rect', then the text appears in the top left. | 09:40.59 |
| | I suspect that certain autoshapes imply the use of centred text. | 09:41.22 |
| pedro_mac | could be | 09:41.38 |
| | no mso- alignment/anchoring properties involved? | 09:42.24 |
| Robin_Watts | no. | 09:42.32 |
| pedro_mac | just bad M$ then ;) | 09:42.45 |
| Robin_Watts | It kinda makes sense. | 09:43.03 |
| pedro_mac | it does, but would have been nicer to actually use styling rather than implicit rendering assumptions | 09:43.31 |
| | s/rendering/layout | 09:43.47 |
| Robin_Watts | pedro_mac: I don't have a feel for this well enough to know if it's reasonable for the prstGeom tags to put a 'centred' style on the enclosing sp... | 09:44.55 |
| jogux | shouldn't show pedro the code I was looking through last night then | 09:46.43 |
| pedro_mac | jogux: :) | 09:48.03 |
| jogux | ah great, within the code in the area of this ppt styling issue there's a comment "This may seem odd but it seems to work..." | 10:05.33 |
| pedro_mac | Robin_Watts: I just created a simple star autoshape in pptx and I get a <a:bodyPr anchor=âctrâ> on the ptxBody for my text | 10:06.06 |
| | this is office 2010 | 10:06.26 |
| | (anchor=âctrâ centres the text in its bounding box | 10:07.23 |
| Robin_Watts | pedro_mac: Ah. I'd just added that manually, and I got the right results. | 10:13.29 |
| | but it's not in the original file. | 10:13.35 |
| | Hmm. I have that on the bodyPr, not on the txBody | 10:14.13 |
| pedro_mac | Did I mention I hate Office documents? ;) | 10:14.25 |
| | Robin_Watts: thatâs right : | 10:15.22 |
| | <p:txBody> | 10:15.27 |
| | <abodyPr rtlCol=â0â anchor=âctrâ/> | 10:15.55 |
| | ... | 10:15.59 |
| Robin_Watts | sorry, I misread what you said. | 10:16.44 |
| paulgardiner | pedro_mac: GFE provisioning is still not working for me. I've sent to you my reply to Brad for sanity checking. Can you check the attached photos do show my correctly entering the login details? | 10:19.27 |
| pedro_mac | did he give you more GFE keys? I thought it was just GoodShare | 10:20.30 |
| | (having said that Iâd guess they should work for any app on the domain, but I reckon he gave us different sets for GfE and Goodshare last time | 10:22.08 |
| paulgardiner | pedro_mac: oh possibly. I did try Good Share first but that attempt complained that GS was delegated to GFE and GFE wasn't present | 10:23.15 |
| pedro_mac | I think they keys *should* work for any, but you never know what oddities they have lurking around with their legacy apps | 10:24.00 |
| paulgardiner | I'll try Good Share again | 10:24.13 |
| pedro_mac | GoodShare stopped working for me but GfE was fine | 10:24.19 |
| | go figure | 10:24.26 |
| paulgardiner | But images show what I think they do? | 10:24.58 |
| | :-) Now Good Share wont even ask for the authentication details. I just get a pop up "Set-up Failed" "Infrastructure Activation Failed" | 10:27.21 |
| pedro_mac | paulgardiner: yeah, just triple-checked the access key - definitely fine. email address too | 10:28.25 |
| paulgardiner | Thanks Pete. | 10:28.43 |
| | I wonder if tgnisv001 isn't set up for the latest SDK. I'm using 1.8 now and probably the latest Good Share it too | 10:29.15 |
| pedro_mac | its possible their policy is duff, but thatâs a big oversight if so | 10:31.26 |
| Robin_Watts | winfvdk seems to use static subsampling. With a threshold of 320. No wonder it looks crap. | 10:49.28 |
| mattchz | probably chosen based on an ipaq screen ;) | 10:50.05 |
| jogux | robin_watts : that would very definitely match that photo. wtf. | 10:51.24 |
| Robin_Watts | Ah, I think 320 is the default, and they are not setting the Pal_Property. | 10:51.44 |
| jogux | nods | 10:53.25 |
| | woohoo. my testdoc now matches powerpoint. | 11:30.41 |
| paulgardiner | Brilliant. Unpresedented brilliance in fact | 11:31.43 |
| | three little commits on my master branch for review | 11:32.36 |
| | Oh possibly containg tabs. I'll just check | 11:33.12 |
| jogux | paulgardiner: I was vaguely wondering if I should try and make gen-patch have a flag to strip tabs/trailing spaces, albeit not 100% sure how to do such. | 11:34.36 |
| | actually maybe it should be another tool. maybe someone has already written such a tool. | 11:34.54 |
| paulgardiner | Is there a way to get git to do it every commit? | 11:35.28 |
| jogux | hrm. possibly. I dislike revision control tools changing my source under me automatically though ;-) | 11:36.21 |
| | does the good sdk really instead to /Good Technology? ie. in the root directory? | 11:36.37 |
| | s/instead/install/ | 11:36.43 |
| pedro_mac | should be in the xcode frameworks iirc | 11:37.15 |
| jogux | I'm confused by FIPSConfig.xcconfig in Paul's first commit then | 11:37.48 |
| | paulgardiner: did you mean to remove Default-568h@2x.png from the project? | 11:38.31 |
| paulgardiner | jogux: it builds and runs. Ship it! | 11:38.38 |
| | jogux: I did intend to remove that, but in a previous commit I did not intend to add it!! | 11:39.07 |
| jogux | ur. okay ;-) when did that slip past me I wonder. | 11:39.40 |
| | it's deliberate we're dropping support for iOS 5? | 11:39.47 |
| | and iOS 4 for that matter :) | 11:39.54 |
| paulgardiner | The stuff in the config file is just copied from GD instructions and seems to work | 11:40.05 |
| | plus without it, it didn't work | 11:40.18 |
| | Target iOS 6 was also part of the FIPS instructions | 11:40.41 |
| jogux | paulgardiner: fair enough. I guess xcode must have a search path that means /Good Technology/ doesn't mean that directoy in / :) | 11:40.47 |
| paulgardiner | I believe the lastest SDK requires 6 | 11:40.57 |
| jogux | what is FIPS? | 11:40.58 |
| | I think I should encourage you to put a bit more detail in your commit messages :-) | 11:41.29 |
| paulgardiner | Some sort of security thing. It legislates that you obey certain criteria. Use only certain encryptions algorithms. | 11:41.59 |
| | jogux: yeah. good point. I'll add some explanation to that on | 11:42.22 |
| jogux | ah, that'll be the US GOV FIPS then. | 11:42.23 |
| pedro_mac | retracts his previous statement - it *used* to install to frameworks, but the latest SDK appears to install to root. Odd. | 11:43.54 |
| jogux | pedro_mac : oo. yep. I can see that on the ATS machine. seems naughty to me. | 11:44.33 |
| pedro_mac | yup, and very unnecessary, unless perhaps they though their users wouldnât know where to find the SDK sample apps ;) | 11:46.04 |
| paulgardiner | jogux: okay, updated the commit message | 11:59.46 |
| jogux | paulgardiner: the language in that document is confusing, but I presume "by default enforces FIPS compliance" means we have no choice in the matter and there's no way to change the default and not use FIPS? | 12:02.20 |
| paulgardiner | It has to at least start in FIPS mode before it can check with the GC to see if it needs to stay in that mode. | 12:03.12 |
| | Apparently you cannot start in non-FIPS mode and then enter it | 12:03.28 |
| | .. which I can believe | 12:03.41 |
| jogux | paulgardiner: I was more wondering if supports FIPS at all was a choice for us - ie. could we keep FIPS out and continue to support iOS 5? | 12:04.04 |
| paulgardiner | Not using 1.8 I believe | 12:04.53 |
| jogux | okay | 12:05.20 |
| paulgardiner | I'm guessing the whole Good suite of apps are about to require 6.0 | 12:05.22 |
| jogux | notes that md4/md5 aren't allowed. don't we use those for something? | 12:05.36 |
| | somewhere in the pdf da I think | 12:06.17 |
| paulgardiner | Hmmm possibly. I think I'll allow vericode to worry about that | 12:07.04 |
| jogux | I presume the FIPS module probably doesn'tm anage to see / block those. | 12:07.08 |
| | hehe | 12:07.16 |
| paulgardiner | Okay to push? | 12:08.24 |
| jogux | yeah, think so. | 12:09.29 |
| henrys | paulgardiner: I see the good saga continues | 12:55.55 |
| paulgardiner | Yes. On the iOS side it's only for testing. I believe everything is working okay in the app because I can test with our own test app | 12:56.49 |
| jogux | robin_watts : hm, I think we use a bad subsampling setting in the testshell too. the text on slide 12 of Artifex2014.pptx is unreadable in places as it seems to be a huge bitmap :-S | 12:59.54 |
| Robin_Watts | jogux: probably, yes. | 13:01.26 |
| jogux | possibly we should increase the default? | 13:01.51 |
| | unless you have something fancier in mind :) | 13:02.06 |
| Robin_Watts | I think we should just set the Pal_Property. | 13:02.42 |
| pedro_mac | that would make sense anyway, given the difference in spec of current targets | 13:02.42 |
| henrys | paulgardiner: ron is working on the web site problems. Having a bit of trouble locating where those logos are in the CMS, but it is off our table. | 13:04.27 |
| | paulgardiner: andhttp://www.picsel.com will eventually behttp://www.artifex.com/smartoffice but I think you should depend of a redirect at least until that is live | 13:05.43 |
| | s/of/on | 13:05.53 |
| paulgardiner | henrys: yeah that sounds best | 13:16.01 |
| | No Tor? | 13:16.23 |
| chrisl | henrys: there was some question about whether GoDaddy allowed the cms that picsel.com uses - if it doesn't, then the site would need rebuilt from the ground up :-( | 13:18.35 |
| henrys | not a web expert but a lot of these cms are sounding like âdata jailsâ | 13:19.37 |
| chrisl | Well, they all use their own underlying data formats, databases etc..... so migrating isn't easy | 13:20.32 |
| | It's not much different to a wiki for that matter - the other option is hacking html manually | 13:21.30 |
| jogux | the fun bit is some of the CMSes you can't even hope to keep data across (major version) upgrades. | 13:22.26 |
| paulgardiner | jogux: Eek! Looking at transferring Android SO2 between accounts. Potential problem: "You will be required to provide the Transaction ID for both developer registrations prior to transferring." That might be hard to find for the Picsel account | 13:22.28 |
| jogux | paulgardiner: what's transaction id? | 13:22.51 |
| | or where would it be found | 13:23.09 |
| paulgardiner | In Google Wallet dashboard | 13:23.28 |
| | That's user specific, right? | 13:23.40 |
| jogux | I would presume we have the login for the picsel google account? | 13:23.55 |
| paulgardiner | Ah Wallet associated with their google account. That's not so bad | 13:24.42 |
| | I thought it might be associated with the person who performed the registration | 13:24.59 |
| jogux | is hoping not. :) | 13:25.12 |
| paulgardiner | That's two of us then | 13:25.22 |
| jogux | I could also go searching through the picsel email if you luck out. | 13:25.28 |
| paulgardiner | I will take a look | 13:25.46 |
| | Whoop! I think I have it. | 13:29.18 |
| | jogux: does our app use integrated services such as Google Analytics or APIs managed through Google Developers Console? | 13:43.07 |
| jogux | I'm not sure, pedro might know that. | 13:43.50 |
| mattchz | paulgardiner/robin_watts: mind looking at my latest review please? http://git.ghostscript.com/?p=user/matt/mupdf.git;a=commit;h=e8f0394761e24ad1cdb143aa40e5594e93a97f0b | 13:44.59 |
| paulgardiner | looking | 13:46.09 |
| | Coo! You've found out what's wrong with bitmap update on 3.1. Magic! | 13:47.13 |
| mattchz | Ah, you already knew about that? ;) | 13:47.24 |
| pedro_mac | paulgardiner: I havenât seen any evidence of any google API usage in the app, other than the licence/DRM stuff but I suspect thatâs not part of it | 13:47.43 |
| mattchz | paul: Yeah, I remember having similar problems when working on another project, but I don't think I ever found a solution at that time other than disabling hardware accel | 13:47.58 |
| paulgardiner | pedro_mac: okay thanks peds | 13:47.58 |
| Robin_Watts | mattchz: Bargain. LGTM. | 13:48.38 |
| paulgardiner | mattchz: is there any risk that a new texture gets uploaded to OGL inbeween erase and render? | 13:48.45 |
| mattchz | Hmm, that's a point. | 13:49.06 |
| Robin_Watts | paulgardiner: tor is on hols. | 13:49.37 |
| paulgardiner | Robin_Watts: oh okay ta | 13:49.44 |
| | mattchz: were you able to test on a 3.1 device? | 13:50.19 |
| mattchz | I tested it on a 3.2 device | 13:50.24 |
| | which showed the same problem. | 13:50.27 |
| | but I think you're right. There could be an issue. | 13:50.38 |
| paulgardiner | Well if your patch seems to work go with it | 13:50.43 |
| mattchz | I do occasionally see 'white screens' appear | 13:50.52 |
| | I'm not sure if that's the problem you are predicting. | 13:51.05 |
| | or if that's just a redraw thing that we would see anyway. | 13:51.17 |
| paulgardiner | Do the white screens just flicker or does it get stuck that way somtimes | 13:51.28 |
| | ? | 13:51.43 |
| mattchz | just flicker, and I think it's when bringing a new page on, so it's probably just that the inital draw hasn't finished. | 13:51.44 |
| | rather than being anything to do with this. | 13:51.54 |
| paulgardiner | No worries. I think go with it. | 13:51.55 |
| | Results without your patch on 3.2 are hopeless | 13:52.12 |
| mattchz | Do we currently render the bitmap from a thread while drawing into it? | 13:52.15 |
| | (from another thread) | 13:52.26 |
| paulgardiner | drawing? render? Don't understand | 13:52.44 |
| mattchz | erm, I mean. I guess we blit the Bitmap in the main thread while updating it in the AsyncTask. | 13:53.04 |
| | so there's already a sort of race. | 13:53.13 |
| paulgardiner | We don't blit do we? | 13:53.27 |
| | ... if I understand what you mean | 13:53.46 |
| mattchz | I mean, we set the ImageView's bitmap to the bitmap, and then do invalidate. | 13:53.53 |
| | i.e. there's no double buffering | 13:54.05 |
| paulgardiner | right. Those are in different threads | 13:54.07 |
| | No double bufferring | 13:54.21 |
| mattchz | and they can run at the same time, so potentially we could get an half-rendered buffer on screen. | 13:54.31 |
| paulgardiner | But we invalidate only when the render is complete | 13:54.43 |
| mattchz | oh, I see. | 13:54.56 |
| paulgardiner | so really OGL should have no reason to reupload until render is complete | 13:55.11 |
| mattchz | in that case, I guess we can't reupload then | 13:55.16 |
| | hopefully | 13:55.19 |
| | ok, so I think it's fine then. | 13:55.40 |
| | ok to push? | 13:55.55 |
| paulgardiner | The only worry I'd have is that erase triggers an upload behind the sceens | 13:56.05 |
| | Yeah push | 13:56.09 |
| | It's a definite improvement. | 13:56.25 |
| mattchz | it seems to work pretty well, anyway. | 13:57.06 |
| | Honeycomb was really a load of broken poo | 13:57.53 |
| | No wonder they didn't want to release the source ;) | 13:58.06 |
| | thanks. | 14:00.54 |
| jogux | robin_watts: http://intranet.picsel.com/~ats/cgi-bin/tgvath-scriptview.pl?resultid=1037&scriptid=3885&diffs=1 | 14:08.14 |
| Robin_Watts | Bargain! | 14:08.38 |
| jogux | minor downside: all circles are now squares | 14:08.48 |
| Robin_Watts | bullet points ? | 14:09.08 |
| jogux | nah. http://intranet.picsel.com/~ats/cgi-bin/tgvath-scriptview.pl?resultid=1037&scriptid=2219&diffs=1 | 14:10.01 |
| Robin_Watts | Oh, all shapes are now squares. | 14:10.26 |
| | That's more of a problem :) | 14:10.29 |
| jogux | :) | 14:10.32 |
| pedro_mac | text formatting looks much improved though | 14:13.33 |
| jogux | robin_watts: there's an issue on page 2 of artifex2014.pptx, alignment is off and there's some square boxes, probably worth looking at that? I don't think it's related to this. | 14:19.43 |
| Robin_Watts | jogux: That's the timeline? | 14:19.51 |
| jogux | that's the one | 14:19.52 |
| Robin_Watts | Yes, I had that down as something to look at, but I was leaving it until you'd finished in case it was a similar styling thing. | 14:20.00 |
| jogux | from what I can see, it doesn't appear these changes will affect it. | 14:20.33 |
| Robin_Watts | but I'll move that up the list if you think it's unrelated. | 14:20.34 |
| jogux | certainly there's no change in that slide with my current patch | 14:20.44 |
| Robin_Watts | I'm just boggling at how the subsample threshold doesn't seem to be configurable at the AlienConfig level. | 14:21.01 |
| mattchz | really? I thought it was? | 14:21.20 |
| jogux | you mean it needs adding to the alien-config-fv.h? | 14:21.20 |
| | robin_watts: PicselConfigFV_StaticSubsampleThresholdOverride ? | 14:21.34 |
| | (that's in picsel-config-fileviewer.h) | 14:21.35 |
| Robin_Watts | D'Oh. | 14:21.37 |
| | Thanks. | 14:21.41 |
| | I had just painstakingly hunted through picsel-config.h | 14:21.58 |
| | I'm going to duplicate it in picsel-config.h so it's available in page-view too. | 14:43.45 |
| | No, I'm not. Ignore me. | 14:44.03 |
| jogux | :) | 14:45.28 |
| pedro_mac | it probably makes more sense in picsel-config.h rather than any app-specific config imho | 14:48.12 |
| | suspect its not the only one in FV config like that though | 14:48.30 |
| Robin_Watts | It's already abused and used in non FV things (such as page-view). | 14:50.19 |
| jogux | I would expect it can probably be viewed as "configuration for apps that view files" rather than "configuration for FVDK". I'd probably be surprised if anything falls into the latter category. | 14:55.55 |
| Robin_Watts | 1 commit on robin/master then | 15:10.32 |
| | oops. stupid typo. hold on a mo. | 15:12.42 |
| | OK, fixed version on robin/master, and on robin/smartoffice-2-1-28-release-branch. ATS test running now. | 15:18.38 |
| pedro_mac | Robin_Watts: the change LGTM - only comment would be regarding the function description in thumbnailer.h - that could be interpreted as âyou should always set the subsample threshold to screen size as thereâs no benefit in larger, but for devices with more RAM available they could retain better clarity when zoomed in. Probably a thorny topic for documentation/implementation guides though | 15:49.01 |
| Robin_Watts | pedro_mac: Yeah. | 15:51.22 |
| henrys | marcosw: damn going to have to change all the pcl md5sums⦠sigh.. did you see my duplex question in the bug report? | 16:30.20 |
| Robin_Watts | henrys: how come? | 16:34.23 |
| henrys | Robin_Watts: changing the default hardware clipping region to 0 and making it settable by a parameter | 16:35.41 |
| Robin_Watts | So every rendered page is going to change size a bit ? | 16:36.23 |
| henrys | yeah before it was 1/6â now the default will be full bleed and you can set it with -H which I just checked in. | 16:37.57 |
| | normally pcl doesnât write much in that 1/6â but many not all files change because they have output that sits right on 1/6" | 16:39.00 |
| pedro_mac | paulgardiner: Iâve just had a play with the new iOS SO2 on testflight - still has the issue with disappearing topmenu - is that expected? | 16:39.26 |
| henrys | Robin_Watts will even think about PCL problems to avoid the dreaded SOT ;-) | 16:40.19 |
| Robin_Watts | :) | 16:40.37 |
| Robin_Watts | is dangerously close to having fixed a bug within a single day. Something is bound to come up... | 16:41.14 |
| pedro_mac | :) | 16:41.40 |
| Robin_Watts | Of course, this only fixes pptx, not ppt. | 16:41.53 |
| pedro_mac | weâre sticking with that approach for now though arenât we? | 16:42.29 |
| paulgardiner | pedro_mac: oh, how do you provoke that? | 16:42.44 |
| Robin_Watts | pedro_mac: Yes. unless it's trivial to fix the other one too. | 16:42.51 |
| pedro_mac | paulgardiner: I just opened an excel spreadsheet from my googledocs account | 16:43.33 |
| paulgardiner | and what happens? I don't think I've seen the problem. | 16:45.18 |
| | Do you see it if you open something from My Docs? | 16:46.25 |
| pedro_mac | the column header is in the correct place but topmenu has vanished - think its a known issue though (or at least Iâm pretty sure jogux had seen it on his iPhone | 16:48.36 |
| paulgardiner | So the blue bar at the top has the title but no icons? | 16:50.06 |
| pedro_mac | basically the topmenu auto-hides after a few seconds but the document doesnt shift | 16:50.19 |
| | perhaps its just excel docs that make it obvious - the document can be panned around (its a tall spreadsheet) but the column header stays at the position it was when the topmenu was visible (ie just below the menu bar) | 16:51.32 |
| paulgardiner | topmenu is the group of icons on the left of the blue bar at the top of the screen? | 16:52.26 |
| pedro_mac | its basically the whole bar which auto-hides | 16:53.50 |
| paulgardiner | Right. So no bar present at all at the top of the screen? | 16:54.11 |
| | Does it leave a gap betwen the top of the doc and the top of the screen or does the doc expand to fit? | 16:55.00 |
| Robin_Watts | pedro_mac: That sounds like a ruler problem. There were various ruler commits on master/branches just before picsel went under. | 16:55.13 |
| pedro_mac | no - I think thatâs expected behaviour for the app though - tapping on the doc will make it re-appear (then it auto-hides after a few seconds of inactivity) | 16:55.13 |
| paulgardiner | strange that I don't see it here | 16:56.07 |
| | I've created a .docx and a .pptx. I can't get rid of the topmenu whatever I do | 16:56.36 |
| pedro_mac | it is leaving a gap - thatâs the problem - especially for excel where you end up with the column header bar stuck part-way down the screen (and you can scroll the doc behind itâ¦) | 16:56.39 |
| | this is excel | 16:56.51 |
| paulgardiner | Just created a .xlsx and still can't get rid of the top menu | 16:58.36 |
| pedro_mac | is that the build from testflight? | 16:59.26 |
| paulgardiner | yep | 16:59.33 |
| pedro_mac | on iPad or phone/pod? | 16:59.40 |
| paulgardiner | i8117 | 17:00.03 |
| | iPad 2 | 17:00.08 |
| pedro_mac | suspects that iPad may well have it always-on - we do that on most android tablet builds I think | 17:00.22 |
| paulgardiner | Makes sense. | 17:00.52 |
| | Behaviour sounds like a feature. | 17:01.16 |
| pedro_mac | Iâm sure its intentional, just looks broken with excel | 17:01.47 |
| pedro_mac | adds a bug | 17:02.12 |
| Robin_Watts | pedro_mac: Don't the rulers usually move with the topbar? | 17:02.18 |
| paulgardiner | Would be good to test the current app store version see if this is new | 17:03.06 |
| | Thanks for testing anyway. Have to decide whether it's worth fixing that before a release. Clearly the rulers should move with the menu | 17:10.34 |
| pedro_mac | np - its http://bugs.ghostscript.com/show_bug.cgi?id=695330 | 17:15.28 |
| paulgardiner | yeah nicely stated | 17:16.06 |
| pedro_mac | for other doc types itâs less obvious - the doc is just lower down the screen than it could be | 17:16.33 |
| paulgardiner | yeah. I can imagine how odd it looks with the rulers | 17:16.53 |
| pedro_mac | apart from that what Iâve tested has been fine :) | 17:17.59 |
| Robin_Watts | Ok, so I've created a whole bunch of autoshapes in powerpoint, put "Text" into each one, and saved them out. | 17:41.23 |
| | By default they all get created as having a vertical anchor of "ctr" and an inline style to centre the text horizontally. | 17:42.22 |
| | If you manually remove those anchors and those styles, the text all moves. But NOT to the top right. | 17:42.52 |
| | It seems that every autoshape has some kind of inner border that the text is enclosed within. | 17:43.22 |
| pedro_mac | or possibly some default offsets being applied? not something Iâve looked at before Iâm afraid | 17:55.25 |
| pedro_mac | gives up waiting on the good support bod phoning me back and goes home. Time for another email prod. | 17:56.00 |
| Robin_Watts | pedro_mac: (For the logs) it's definitely shape specific (different shape, different margins) | 17:57.16 |
| henrys | Robin_Watts: I guess Joseph and Pete have left, whatâs the protocol for getting an account and password for TWiki GhostDocs? | 18:24.01 |
| Robin_Watts | henrys: You ask me. | 18:34.58 |
| | and I try to remember it :) | 18:35.11 |
| | You already have an account (HenryStiles) | 18:35.48 |
| | If you go here: http://twiki.ghostscript.com/do/view then you can see yourself in the 'Twiki Users' list on the top right. | 18:36.37 |
| | Click on yourself, then there is a "ChangePassword" link. | 18:36.51 |
| henrys | Robin_Watts: Iâm in I must have set a password and donât remember | 18:36.57 |
| Robin_Watts | henrys: cool. | 18:37.03 |
| | Now, I think you should be able to see GhostDocs ? | 18:37.17 |
| myspacee | hello | 19:22.30 |
| ghostbot | Welcome to #ghostscript, the channel for Ghostscript and MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 19:22.30 |
| myspacee | is there any way to create "PDF/X-1a" ? | 19:23.20 |
| pedro_mac | Robin_Watts: thatâs odd with the autoshapes; youâre probably right about some implicit behaviour in the M$ shape processing | 19:29.36 |
| | I couldnât see any shape-specific margins/offsets/indents in the test docs I made. Quite feasible that there are version-specific differences too | 19:31.04 |
| Robin_Watts | pedro_mac: yeah, I need to understand the Edr that we are currently generating too. | 19:31.13 |
| | I think that when we create a autoshape group there are 'some' margins in there, but I can't see where they are specified in the Edr. | 19:31.37 |
| | If they are implicit ones, then possibly I could just change that logic. | 19:31.58 |
| pedro_mac | nods | 19:32.09 |
| Robin_Watts | If they are explicit then I should change them there. | 19:32.13 |
| | so tomorrow, I'll bug jogu about that (trying to avoid bugging you about these things as good is more important) | 19:32.36 |
| Gigs- | make so is gone. How do you just build ghostscript? I don't want all this other stuff | 19:58.15 |
| | the official documentation still says to use make so. The readme in git doesn't actually say how to build regular ghostscript. | 20:05.23 |
| | I don't even see a likely target in Makefile itself | 20:05.44 |
| | oh here it is, gs has been moved into a subdirectory | 20:07.27 |
| mvrhel_laptop | Gigs- that change was made a long time ago... | 20:08.13 |
| dewesh | hi ,if i may introduce myself i am a final year student from iit Kharagpur. | 20:08.46 |
| | i was going throught your code and i think many thinkgs has to be done and would like to contribute | 20:09.24 |
| | some guidance would be best. | 20:09.54 |
| Gigs- | mvrhel_laptop: I must have forgotten. The build system in the upper level directory confused me. | 20:10.11 |
| mvrhel_laptop | Gigs- I forget stuff like that all the time... | 20:11.21 |
| Gigs- | we are seeing a resolution dependent segfault again, but a little different this time | 20:12.03 |
| | looks like it crashes if -r > 113 or < 148 where last time it was only one specific dpi | 20:12.33 |
| | can someone mark the attachment private http://bugs.ghostscript.com/show_bug.cgi?id=695333 | 20:29.56 |
| | Robin_Watts: ^ | 20:30.01 |
| Robin_Watts | someone beat me to it. | 20:30.56 |
| Gigs- | thanks | 20:31.01 |
| mvrhel_laptop | marcosw: running a bmpcmp for the fix with the whitepoint in the conversion of calRGB to ICC | 22:44.26 |
| | quit a few files affected. | 22:44.32 |
| | basically any file that used calRGB | 22:44.48 |
| | and had a white point that was not D50 | 22:44.56 |
| | which was probably all of them | 22:45.01 |
| | that is all that had calRGB did not have a D50 illuminant | 22:45.21 |
| | oops sorry about the long title in the white point commit | 23:43.29 |
| | marcosw: are you there? | 23:43.40 |
| | I sent you an email as I committed the fix for Customer 532 | 23:53.36 |
| | no bug to close though | 23:53.48 |
| | with that done time for a brain context switch back to SOT | 23:54.06 |
| | Forward 1 day (to 2014/06/26)>>> | |
IRC Logs