| <<<Back 1 day (to 2014/01/20) | 2014/01/21 |
ray_laptop | I'm glad I'm not the only one that writes emails that then don't get sent somehow. | 00:10.10 |
marcosw | ray_laptop: you referring the email to customer I sent today that I wrote 4 week ago :-) | 00:43.20 |
mobjax | We integrated MUPDF in one of our Android Apps. Inside the MupdfActivity.java, we added a button and on the click of that button, we are closing the activity by issuing FINISH(). We are able to successfully open the PDFs and also are able to return to the main activity on that button click. But we are facing a strange problem when the PDFs are large (>10 pages), when we scroll a few pages, and in between if we click on the Done but | 07:38.35 |
| Please help us in fixing this, Thanks in Advance! | 07:39.12 |
| Hi paulgardiner, Good Morning, can you please answer couple of questions that we have issues in Mupdf library? | 11:21.52 |
paulgardiner | Sure. I'll try | 11:22.06 |
mobjax | Thanks | 11:22.13 |
| We integrated MUPDF in one of our Android Apps. Inside the Main Activity of our App, we are calling the class MupdfActivity.java by passing the PDF location to it. We encountered with the below issues: We added a button and on the click of that button, we are closing the MupdfActivity by issuing a FINISH() inside the program, MupdfActivity.java. We are able to successfully open the PDFs and also are able to return to the main acti | 11:22.27 |
| But we are facing a strange problem when the PDFs are relatively large (>10 pages), when we scroll a few pages, and in between if we click on the Done button, the App is force closing. | 11:23.17 |
| Second is If we reopen a closed PDF, the last visited page is shown instead of showing the first page. Ex: PDF has 10 pages, we traversed to the 5th page, closed the PDF. Now, if we reopen the PDF, 5th page is shown instead of showing the first page. | 11:23.38 |
paulgardiner | mobjax: before getting into the detail, just something I should check in case you get far along with this and then find you've wasted your time: you do know that MuPDF is released under GPL, so you'll have to release the source of your app unless you get a seperate licence from Artifex. | 11:26.23 |
mobjax | that means, shouldn't i use MuPDF at all | 11:27.29 |
| if i do not want to release the source of my App | 11:27.42 |
Robin_Watts | mobjax: If you are not prepared to release the source of your App, then you have 2 choices. | 11:28.37 |
| 1) don't use MuPDF at all. | 11:28.45 |
| 2) get a commercial license for MuPDF from Artifex. | 11:28.52 |
| Either of those is fine. | 11:29.22 |
mobjax | but I already built my app around MuPDF, and right now i don't find any equivalent libraries that can open all the PDFs | 11:29.58 |
| so I would need to think about second option, but I am not sure how much that costs me | 11:30.22 |
| as I want to distribute my App for free for more number of users | 11:30.36 |
Robin_Watts | Well, we tune our licensing to need. | 11:30.48 |
| If you contact scott.sackett@artifex.com and say that you want to use MuPDF in an android app, that'll start a conversation. | 11:31.24 |
mobjax | ok, | 11:31.40 |
Robin_Watts | The first thing he'll do is send back a huge list of questions, many of which won't apply to you, but just try and answer them as best you can. | 11:31.44 |
mobjax | ok, | 11:31.54 |
| i will | 11:31.56 |
| send an email to him right away | 11:32.04 |
Robin_Watts | The idea is to give him an idea of exactly what you want to do with the lib. He'll then make you a licensing proposal. | 11:32.22 |
mobjax | ok | 11:32.30 |
| or one quick question, I do not know if i can ask it or not | 11:32.53 |
| do you know any other libraries that perform as good as MuPDF | 11:33.09 |
kens | Depends what you mean by 'sd good as', but even if there is such a thing, it is likely to be commercial, or open source with a licence similar to AGPL, which means you have to open source your code too. | 11:34.59 |
mobjax | oh ok | 11:35.35 |
chrisl | mobjax: so, if you want to give your app away free, why not open source it? | 11:36.56 |
mobjax | i can't open the source | 11:39.20 |
Robin_Watts | Possibly he wants to give the app away for free, but use it as a portal to a service of his that is potentially charged for. | 11:39.27 |
mobjax | true | 11:41.00 |
| not potentially charged though! | 11:41.30 |
| :-) | 11:41.34 |
| it is going to be a secured Application, so I can't open the source | 11:41.58 |
paulgardiner | Explain the details to Scott and I'm sure he'll be helpful. | 11:44.55 |
mobjax | ok | 11:45.06 |
| i sent an email already | 11:45.26 |
paulgardiner | mobjax: when you say "calling the class MuPDFActivity.java" above, do you mean via an "intent"? | 11:47.16 |
mobjax | yes | 11:51.29 |
| Uri uri = Uri.parse("path to pdf file"); Intent intent = new Intent(context, MuPDFActivity.class); intent.setAction(Intent.ACTION_VIEW); intent.setData(uri); context.startActivity(intent); | 11:52.01 |
| and inside the class, i added a button to the relative layout | 11:52.30 |
| and inside the OnClicklistener, I have finish() statement | 11:52.55 |
Robin_Watts | mobjax: If you don't hear back from Scott within 2 days, let me know. | 11:52.57 |
mobjax | ok | 11:53.03 |
| shall i forward the mail to you as well | 11:53.12 |
Robin_Watts | no need. It's just that occasionally mail to scott has gone missing. | 11:53.31 |
mobjax | oh ok | 11:53.40 |
paulgardiner | And the problematic pdf files show no problems displayed using our app? | 11:53.42 |
mobjax | true | 11:53.48 |
| they open just fine | 11:53.54 |
| i tried VuPDF | 11:53.59 |
| that failed in opening some of the PDFs | 11:54.14 |
paulgardiner | Have you tried repeatedly opening them and then returning to the file picker in MuPDF? That's effectively doing the same thing. If that doesn't lead to a crash, it much be a problem in your code to where the activity returns. | 11:56.54 |
mobjax | I haven't tried this | 11:57.26 |
| one more question which is related to this | 11:57.43 |
| as I am issuing a Finish inside the MuPDFActivity.class, | 11:58.00 |
| it should eventually close the subsequent classes like PageView, ReaderView etc | 11:58.20 |
| but the strange thing is, | 11:58.29 |
| if I reopen the same pdf, it opens the last visited page, | 11:58.52 |
| so the question is, is it not properly finishing the activity? | 11:59.18 |
paulgardiner | mobjax: there is specific code in MuPDFActivity to store details of the last viewed page and reinstate it. You just need to remove that. | 11:59.30 |
mobjax | or rather, how is it able to open the last visited page | 11:59.31 |
| oh is it? | 11:59.54 |
| can you please tell me the function name | 12:00.13 |
paulgardiner | I can tell you where in a minute because I happen to be looking at that class | 12:00.22 |
mobjax | i found out | 12:00.57 |
| it is storing in sharedpreferences | 12:01.14 |
paulgardiner | yeah | 12:01.23 |
| Be warned though: if you remove that without care, you may find that the display move to the first page on device rotation. | 12:02.10 |
mobjax | oh true | 12:02.27 |
paulgardiner | "moves to" that is | 12:02.28 |
kens | Robin_Watts : ping | 12:04.14 |
| git ping :-( | 12:04.21 |
mobjax_ | my Tab is a rectangle | 12:05.37 |
| Samsung | 12:05.40 |
Robin_Watts | kens: pong | 12:05.45 |
mobjax_ | resolution: 1280 X 800 | 12:05.49 |
| so if I try to scale my PDF to the screen | 12:06.03 |
kens | Robin_Watts : I htink I answered my question..... | 12:06.12 |
mobjax_ | how can I do that | 12:06.13 |
Robin_Watts | mobjax: Depends what you mean. | 12:06.25 |
kens | Robin_Watts : I want to push my local branch to casper I htink | 12:06.31 |
| THat's 'me' on casper, I hope | 12:06.47 |
mobjax_ | I din't get you, can you please elaborate | 12:06.49 |
Robin_Watts | kens: git push me branchname | 12:06.56 |
kens | Hmm. | 12:07.21 |
Robin_Watts | mobjax_: Suppose the PDF has portrait pages that are natively twice as high as they are wide. | 12:07.38 |
| That means that in portrait mode on your device, we will display them as 1280 high, but 640 wide. | 12:08.10 |
mobjax_ | whatever the size of the pdf is, i want to set them to the screen | 12:08.10 |
kens | Robin_Watts : what is 'me' in that syntax ? | 12:08.13 |
Robin_Watts | i.e. we preserve the aspect ratio. | 12:08.18 |
mobjax_ | but what if i want to fit them to the screen | 12:08.31 |
| whatever the original size of the PDF might be | 12:08.39 |
Robin_Watts | kens: I interpreted what you said as being that "me" was the remote name for your remote repo on casper. | 12:09.00 |
kens | Oh, I think I can just use git gui and push it to casper as teh 'remote' I think that's what I've done before | 12:09.28 |
Robin_Watts | Personally, I have the main repo called 'golden' and my personal repo called 'origin'. | 12:09.35 |
kens | As long as I don't use origina I shouldn't screw anythign up | 12:09.40 |
Robin_Watts | hence I would use: git push origin branchname. | 12:09.54 |
| but for you it might be: git push casper branchname | 12:10.07 |
kens | I htnk origin for me is the master, its where I normally push to | 12:10.12 |
Robin_Watts | You can do: git remote -v to see a list. | 12:10.20 |
kens | Its going to kens, so that's OK | 12:10.34 |
Robin_Watts | mobjax_: Right, so you want to ignore the aspect ratio. | 12:10.44 |
mobjax_ | yes | 12:10.50 |
Robin_Watts | That's easy to do with the underlying MuPDF lib as it's just a question of tweaking the transform matrix you pass in. | 12:11.18 |
mobjax_ | where i need to do the change | 12:11.35 |
Robin_Watts | but the problem is that the android app assumes that you wouldn't want to do anything like that. | 12:11.36 |
mobjax_ | means | 12:11.51 |
| it should automatically happen | 12:11.56 |
| like as it happens today | 12:12.04 |
| of perserving the aspect ratio | 12:12.12 |
| it should automatically set the pdf to the tablet | 12:12.21 |
| whichever way i rotate | 12:12.29 |
Robin_Watts | The current app derives the matrix used for rendering from the position/scale of the views. | 12:12.59 |
mobjax_ | yes | 12:13.34 |
Robin_Watts | All the code that handles the moving of the views around the screen, flipping of the pages assumes that it's going to preserve the aspect ratio. | 12:13.46 |
mobjax_ | true | 12:14.06 |
Robin_Watts | What you want it to do requires someone to read through all that code and understand it, and rethink how it needs to change with your new requirement in mind. | 12:14.27 |
| I can't predict how long that would take. | 12:14.49 |
mobjax_ | :-) | 12:15.03 |
| i achieved it some how | 12:15.27 |
| by overriding the scale field | 12:15.34 |
Robin_Watts | and it's not something we'd generally look at for an unsupported user. | 12:15.37 |
mobjax_ | just before | 12:15.39 |
Robin_Watts | Well, if you have something that works, great. | 12:15.45 |
mobjax_ | but the problem is, i can do it for width | 12:16.03 |
| but for height it is again a problem | 12:16.13 |
Robin_Watts | OK. As I say this it not something I'm going to get involved in for an unsupported user, because I don't know the code well enough. | 12:17.22 |
mobjax_ | But my main issue is closing of my main activity | 12:17.27 |
| when i do some speedy flips on the pdf | 12:18.07 |
| and returning to the main acitivity | 12:18.21 |
Robin_Watts | paulgardiner *might* be able to help you (if he has a clear view of a simple change you can make), but again, he has work to do for paid customers, so that has to take priority unless he can see a quick fix you can use. | 12:18.39 |
mobjax_ | oh ok | 12:18.56 |
Robin_Watts | mobjax_: If you can show us that there is a problem with our existing code, then we might be tempted to look at it. | 12:19.02 |
mobjax_ | i would wait for his time | 12:19.04 |
Robin_Watts | But if it's a question of "I've modified the code and now it crashes", well, that's just not something we can get into, because it's not our business to debug someone elses code. | 12:19.59 |
paulgardiner | mobjax_: I can't do anything with that limited information. You'd need to look at the logs and trace where it's failing. | 12:20.00 |
mobjax_ | no what my doubt is, even i issue Finish(), it is not properly closing the subsequent activities that our MuPDFActivity.class triggers (ex: PageView.class, ReaderView.class, etc) | 12:20.12 |
paulgardiner | mobjax_: one thing: is the crash only if you close immediately after many flips? What happens if you let it settle then close? | 12:21.15 |
mobjax_ | it works normal | 12:22.00 |
| no issues then | 12:22.18 |
paulgardiner | So it crahes only if you close immediately after a few flips? | 12:22.45 |
mobjax_ | exactly | 12:22.51 |
paulgardiner | and what do you see in the logs? | 12:25.40 |
mobjax_ | it says it is unable to allocate in PageView.class | 12:26.17 |
paulgardiner | So a java exception? There should be a traceback | 12:27.12 |
mobjax_ | I am not sure, | 12:27.26 |
Robin_Watts | That sounds like you've killed off the app while one of the classes is still working. | 12:27.33 |
mobjax_ | yeah | 12:27.49 |
| finish() does the same | 12:27.54 |
Robin_Watts | That class tries to allocate something which fails because the rest of the code has died, and that makes the whole thing fall in a heap. | 12:28.22 |
paulgardiner | The original app uses finish() though, so it is strange | 12:28.27 |
Robin_Watts | Probably you need to be smarter about where you call finish() from. | 12:28.32 |
mobjax_ | oh ok | 12:29.20 |
| got you | 12:29.22 |
| i am calling it just inside the button click, which i shouldn't | 12:30.06 |
| i may need to use the existing finish(), am i right ? | 12:30.34 |
Robin_Watts | mobjax_: Maybe make the button set a flag, and check that flag from elsewhere ? | 12:31.07 |
paulgardiner | mobjax_: certainly a look at the logs would help. Stick them on pastbin or similar | 12:31.53 |
mobjax_ | what is pastbin? | 12:34.01 |
Robin_Watts | pastebin.com | 12:35.09 |
mobjax_ | oh ok | 12:35.14 |
| got it | 12:35.15 |
| so, if i do a few flips | 12:37.13 |
| that means i am asking the class to do something, and in the mean time, i am closing the activity | 12:37.36 |
| so it is not finding the triggering activity, so it is failing | 12:37.47 |
| i will try to find if i can do anything | 12:38.33 |
| Thanks Robin and Paul for your time | 12:38.54 |
| I am very thankful to you | 12:39.08 |
Robin_Watts | best of luck | 12:39.19 |
mobjax_ | Thanks Again ! | 12:43.20 |
kens | ..\691143 | 13:20.46 |
| drat, I hate it when a starting application grabs the keybaord focus while you're typiung | 13:21.17 |
mobjax_ | Hi paulgardiner | 13:24.34 |
| this is the log cat of the MuPDF | 13:24.44 |
| java.lang.NullPointerException at com.artifex.mupdfdemo.PageView.addHq in PageView.java on Line 606 at com.artifex.mupdfdemo.MuPDFReaderView.onSettle in MuPDFReaderView.java on Line 243 at com.artifex.mupdfdemo.ReaderView$2.run in ReaderView.java on Line 781 at android.os.Handler.handleCallback in Handler.java on Line 725 at android.os.Handler.dispatchMessage in Handler.java on Line 92 at android.os.Looper.loop in Looper.java on L | 13:24.46 |
Robin_Watts | mobjax_: Do NOT Paste here. | 13:24.55 |
| Use pastebin and drop the URL here. | 13:25.04 |
mobjax_ | oops | 13:25.07 |
| oh ok | 13:25.09 |
| I am sorry | 13:25.22 |
| this is the url | 13:25.26 |
| http://pastebin.com/NEXppjLz | 13:25.27 |
Robin_Watts | no worries. | 13:25.40 |
| That looks like a useful backtrace. | 13:26.19 |
mobjax_ | what should i do now, do i need to catch any exception? | 13:27.10 |
Robin_Watts | well, except that you've changed PageVew, right? | 13:27.12 |
mobjax_ | no | 13:27.19 |
| i haven't changed PageView, I changed just the MuPDFActivity.java | 13:27.32 |
| i included a button which on click finishes the activity | 13:27.51 |
paulgardiner | mobjax_: that's not the most recent version of the source from git. Which version are you using? | 13:28.19 |
mobjax_ | I took the latest version | 13:29.44 |
Robin_Watts | mobjax_: From where? | 13:29.59 |
| The "latest version" is always the version in git. | 13:30.07 |
| The "latest release" would be 1.3 | 13:30.17 |
mobjax_ | I followed the steps listed in this url | 13:31.12 |
| http://mupdf.blogspot.in/ | 13:31.13 |
Robin_Watts | So you're using 1.1 ?!? | 13:31.41 |
mobjax_ | yes | 13:32.03 |
Robin_Watts | How is that "the latest version" ? | 13:32.12 |
| 1.1 is horribly out of date. | 13:32.16 |
| We've done all sorts of fixes and improvements since then. | 13:32.27 |
mobjax_ | no ait | 13:32.34 |
| wait* | 13:32.35 |
| I am using 1.3 only | 13:32.43 |
| I checked my dump | 13:32.55 |
Robin_Watts | OK, so the first thing for you to try is to update to using the version from git. | 13:33.02 |
| We may have fixed the problem already. | 13:33.14 |
mobjax_ | oh ok | 13:33.28 |
| so do you want me to build it again | 13:33.36 |
| from the scratch | 13:33.40 |
Robin_Watts | Yes. | 13:33.42 |
mobjax_ | did u fix anything in this week, as i built it some 10 days back only | 13:34.25 |
Robin_Watts | When you built it is irrelavent. It's when the source code dates from. | 13:34.50 |
mobjax_ | I took 1.3 only | 13:35.05 |
| when I built | 13:35.10 |
Robin_Watts | Which dates from last august. | 13:35.18 |
| I'm telling you to grab the latest version from git and to build that. | 13:35.36 |
| which will date from yesterday. | 13:35.51 |
mobjax_ | ok | 13:37.08 |
| i will build it again | 13:37.14 |
| I will build it and test it, if it fixes my problem I am lucky | 13:40.08 |
| Thanks again for your time | 13:41.05 |
henrys | paulgardiner: test flight worked for me, nice. | 14:15.10 |
paulgardiner | henrys: great. Nice system they've set up. | 14:16.28 |
| Does the app seem okay? | 14:16.36 |
henrys | It starts I don't remember how to get files to it on iOS, are you looking in a particular place where I should put pdf files? | 14:17.31 |
tor8 | henrys: if you open a pdf in safari, you should get a button near the top where you can pick "Open in MuPDF" | 14:17.53 |
henrys | tor8:oh I tried to open a pdf directly with dropbox and didn't get asked. | 14:18.30 |
paulgardiner | henrys: also in itunes you can add files to the app | 14:18.32 |
tor8 | paulgardiner: a minor tweak to pdf-util.js on tor/master for your review | 14:19.42 |
henrys | tor8: why does it do that in safari but not from iOS or should it? | 14:19.59 |
paulgardiner | henrys: on a different subject, appologies if I've just sent Raed your way! | 14:20.15 |
tor8 | henrys: I have no idea. I think it might just be a missing feature in Dropbox to show a list of alternative file handlers. | 14:20.25 |
| henrys: what if you try an XPS file? | 14:20.31 |
henrys | tor8: let me try that | 14:20.54 |
tor8 | paulgardiner: is it possible to test the app without having a mac handy? | 14:21.22 |
paulgardiner | If you register with testflight, I can add your device to the list and regenerate the test version | 14:22.22 |
| tor8: I put a post in tech about testflight. | 14:22.36 |
tor8 | paulgardiner: oh, right! that one. I'll go do that now. | 14:23.06 |
| and charge my ipad | 14:23.09 |
henrys | tor8: it doesn't know what to do with xps files | 14:23.27 |
tor8 | henrys: Dropbox or Safari? | 14:23.54 |
henrys | Dropbox | 14:24.08 |
| tor8: but I assumed this would just something that gets shelled out to iOS and would work uniformly in any app that opens a file - have we tried file managers in iOS? | 14:25.25 |
| paulgardiner: np about raed | 14:26.46 |
tor8 | henrys: I don't think there are any file managers in iOS... every app lives in its own sandboxed directory with very few ways of passing files to other apps. | 14:28.16 |
henrys | tor8:dropbox does seem to identify all the multimedia stuff right and do the right thing as you'd expect but it may just be core iOS apps that can be started | 14:30.27 |
paulgardiner | tor8: I may already have included your iPAD. You could try installing the app already uploaded to testflight | 14:30.56 |
tor8 | paulgardiner: it might be a while until it decides to start. battery's drained... | 14:32.16 |
paulgardiner | henrys: dropbox provide an iOS SDK. Possibly that is the mechanism by which apps integrate with it. | 14:33.20 |
henrys | paulgardiner: yeah I guess each app has some different thing, check out good reader: http://www.goodiware.com/gr-man-howto.html#transfer | 14:34.20 |
| paulgardiner: I don't think we want to fool with all that. | 14:34.31 |
tor8 | paulgardiner: I just got an email from apple that ios 7.1 beta 4 and xcode 5.1 beta 4 are out. if you want, I could set up to forward all such emails from apple (they're to our developer account) to you. | 14:35.02 |
paulgardiner | something we can always consider down the line. | 14:35.04 |
| tor8: thanks, but I seem to get them anyway for some reasone | 14:35.41 |
henrys | paulgardiner: well item 4 looks good if it is simple. | 14:36.10 |
| paulgardiner: but probably not important right now | 14:36.34 |
tor8 | henrys: item 4? | 14:37.26 |
paulgardiner | I could have a look. I've been mainly concentrating on exposing the new MuPDF features | 14:37.29 |
henrys | tor8:see the link above | 14:38.39 |
tor8 | Apple's document interchange? yeah, that'd probably be a good idea. | 14:39.20 |
| something else is AirPrint, do we do that yet, paul? | 14:39.30 |
paulgardiner | tor8: no. Not yet doing printing or signature support, but the rest is in. | 14:39.58 |
henrys | paulgardiner: I do wonder if that would be more important than the other features. | 14:41.06 |
paulgardiner | Yeah. I should probably concentrate on that next when I get back to iOS. | 14:41.47 |
| ... after the office doc push. | 14:42.06 |
tor8 | paulgardiner: did you see my poke about a commit on tor/master or did that fall off your scrollback? :) | 14:42.21 |
paulgardiner | I was just thinking it might be worth a release on iOS of what we have so far | 14:42.29 |
| tor8: sorry. Saw it and then got destracted | 14:42.50 |
| tor8: just fectched and I don't see it. | 14:44.21 |
| tor8: would help if I had you set up as a remote. Weird. I thought I did. | 14:46.09 |
henrys | tor8:I got no prompt from safari | 14:47.30 |
| tor8: I wonder if that is a test flight issue. | 14:47.56 |
paulgardiner | tor8: ah I do. Still can't see your commit | 14:48.27 |
| ah no found it, just not rebased yet | 14:48.52 |
tor8 | paulgardiner: yeah, it's a bit old by now | 14:49.37 |
| I've been forgetting to poke you | 14:49.51 |
paulgardiner | Looks fine, but how did you test it? | 14:51.10 |
tor8 | paulgardiner: I haven't, I should probably warn you about that | 14:51.26 |
| I just ran it through a javascript interpreter to catch any syntax errors | 14:51.58 |
henrys | paulgardiner: let's talk about a release at the meeting in a half hour, others may want to weigh in. | 14:52.21 |
paulgardiner | henrys: okay. Makes sense. | 14:52.47 |
| I don't think the releases cost us anthing or require much effort. | 14:53.31 |
tor8 | I think we are very overdue for a release, I think the app in the store is still at mupdf version 1.1 or 1.2 | 14:54.07 |
paulgardiner | tor8: probably a cluster test would pick up problems. | 14:54.09 |
tor8 | paulgardiner: the mujstest stuff is run on the cluster, that's right. | 14:54.34 |
paulgardiner | tor8: all the javascript that's there was added because of being used somewhere in our test files, so hopefully... | 14:55.50 |
| tor8: the commit certainly looks good. | 14:55.59 |
tor8 | paulgardiner: apart from changing the prototype of the objects from Array to Object, nothing other than syntax has changed | 14:56.38 |
paulgardiner | Certainly looks to be the case | 14:58.18 |
tor8 | henrys: I went and did an alex... I spent a week (or two) implementing a javascript interpreter. | 14:58.21 |
henrys | tor8:wow | 14:59.16 |
tor8 | so soon we shan't need the v8 or javascriptcore bloated monsters anymore | 14:59.32 |
| it's missing regular expression support, but it can already do most of the spec completely with weird corner cases covered | 15:00.00 |
henrys | tor8:are you going to release it separately as well? | 15:00.03 |
tor8 | henrys: I was going to ask about licensing for it. I've currently hosted it on github. it started as a weekend project, but now that it's consumed so much of my time, and it's actually looking realistic to be able to use it | 15:00.51 |
henrys | tor8:if I had any idea it could be done in that amount of time I certainly would be all for it. So I don't think it was Alex like. | 15:01.14 |
tor8 | it may make sense to do something proper with it | 15:01.16 |
| the current license I stuck on it is the same as the OpenBSD license, but if you want I'd be okay with AGPL instead (but that really depends if you think there's a market for it) | 15:02.05 |
| I doubt there is, anybody who cares about performance will go for spidermonkey or v8 anyway, the focus here is on correctness and small code bloat | 15:02.31 |
henrys | tor8:AGPL is best for selling it dual license, of course. | 15:03.23 |
paulgardiner | tor8: I'd have thought there would be a market for that. | 15:03.24 |
| amazing | 15:03.40 |
tor8 | henrys: yeah, but with BSD we could still get support contracts (and the question is, do we want to?) | 15:03.54 |
| anyway, the question is out there. it's still a few weeks from being complete. | 15:04.21 |
henrys | tor8:good question | 15:04.23 |
tor8 | complete enough to do a first release, that is. | 15:04.35 |
| https://github.com/ccxvii/libjs/ that's where the code lives | 15:05.01 |
paulgardiner | tor8: When you get to the API for integration of native code there's some issues I'd like to bring up. e.g., v8 provided for a much more efficient connection with MuPDF than Javascriptcore | 15:05.18 |
henrys | tor8:okay | 15:05.20 |
tor8 | paulgardiner: yeah, that's something I wanted to ask you | 15:05.31 |
ray_laptop | morning, all | 15:05.31 |
tor8 | paulgardiner: I've designed the API to be very similar to Lua | 15:05.48 |
| paulgardiner: my current thinking is that you'll create a new Object and attach a userdata pointer (and destructor) to it | 15:06.11 |
| and then use property accessors to implement native functions on the object | 15:06.25 |
| paulgardiner: yeah, I saw the javascripcore wrapper was a lot heavier than the v8 one | 15:07.24 |
paulgardiner | v8, strangely, allows native data to be attached to the property accessors, and that turns out to be important when trying to create a layer to putting one native API over another. | 15:07.29 |
| In MuPDF we have two levels of API, one above the other, with the DOM definitions lying between the two. If you are able to implement the lower interface, you are able to reuse the DOM definitions. | 15:08.54 |
| That's attractive if you wish to allow use of several engines. | 15:09.09 |
tor8 | paulgardiner: how is the native data attached to the property accessor in v8? (any particular function I should be reading) | 15:09.48 |
paulgardiner | If you are forced to implement the upper interface then you have to create a load of engine specific DOM stuff | 15:09.51 |
| tor8: hang on I'll loook | 15:10.09 |
| ... well not that hard, I'll look maybe | 15:10.31 |
tor8 | is it attached to the javascript object (that gets passed as the 'this' object) | 15:10.34 |
| or is it attached to the property itself? | 15:10.52 |
paulgardiner | Actually attached to the property (which seems strange, but is surprisingly useful) | 15:11.41 |
tor8 | paulgardiner: yeah. my main worry is people detaching any native functions and calling them on the 'wrong' object | 15:12.33 |
| like you can do with Number.property.toString.call(new Boolean(true)) | 15:13.03 |
| that'll give you a TypeError (since the 'this' object that gets passed to Number's toString isn't a Number) | 15:13.25 |
| s/property/prototype/ | 15:13.34 |
paulgardiner | pdf_jsimp_addmethod is where the issue arrises | 15:13.59 |
henrys | tor8: what sort of testing has libjs been subject to? | 15:14.20 |
tor8 | henrys: not much yet, just loaded up weird test cases of the specification manually while working on the features | 15:15.05 |
| the plan is to run it through the ecma-262 test suite | 15:15.14 |
| http://dmitrysoshnikov.com/ecmascript/the-quiz/ stuff like that | 15:15.31 |
henrys | ray_laptop:are you going to be okay for the next staff meeting, driving and all? | 15:17.05 |
ray_laptop | henrys: I sure hope so! | 15:18.17 |
| I'm getting better with keeping the eye moist, but I hope the whole thing goes away before then. | 15:18.51 |
| the staff meeting would be at the outside of the worst case "3 months" window for recovery | 15:19.35 |
| I just passed 5 weeks (time flies when your having fun, so it feels like 5 months) | 15:20.23 |
henrys | ray_laptop: I was just thinking about the drive from dallas to denton I haven't even checked how long it is. | 15:20.50 |
Robin_Watts | 40 mins, apparently. | 15:20.59 |
ray_laptop | henrys: I think it's about 45 min | 15:21.06 |
Robin_Watts | longer if you get stuck in the queue behind my accident :) | 15:21.31 |
ray_laptop | Scott probably does it in 30 min ;-) | 15:21.47 |
| Robin_Watts: at least we don't have roundabouts to contend with as you drive on the right side for a change :-) | 15:22.26 |
| "right" in both senses | 15:22.50 |
Robin_Watts | ray_laptop: Driving on the right is fairly easy, as long as it's an automatic. | 15:23.01 |
| Working a manual with my left hand is not something I've ever had to do. | 15:23.20 |
ray_laptop | Robin_Watts: tell me about it. I picked up a car at Gatwick and had to drive to Heathrow to pick up someone else. All they had was a stick and no AC on a really hot day. There were lots of breakdowns on the motorway as well (overheating) so lots of start-stop | 15:24.48 |
Robin_Watts | ray_laptop: Urk. | 15:25.29 |
| Over here, automatics are rare. | 15:25.42 |
chrisl | So are really hot days! | 15:25.56 |
ray_laptop | by the time I got to Heathrow, I had plenty of practice shifting. I would have been hosed if the pedals were reversed as well | 15:25.57 |
Robin_Watts | chrisl: true! | 15:26.04 |
chrisl | ray_laptop: we may have insisted on the anachronism of driving on the left side of the road, but we don't require cars to retain the pedal layout of the Model-T...... | 15:27.35 |
henrys | in texas they just drive in the middle so they won't notice any missteps | 15:28.45 |
| we should start the meeting. | 15:29.23 |
| iOS release? I'd rather put something out that can get files better but I defer to the wisdom of the mupdf'ers | 15:30.25 |
chrisl | How do other PDF viewers handle the "file chooser" issue? | 15:31.01 |
ray_laptop | painfully ? | 15:31.18 |
henrys | chrisl: see the link above | 15:31.18 |
| http://www.goodiware.com/gr-man-howto.html#transfer | 15:31.40 |
Robin_Watts | good reader implements a WebDAV thing, AIUI. | 15:31.59 |
chrisl | So there's no standardised method? | 15:32.09 |
Robin_Watts | chrisl: It's ios. There are no sharp corners cos you might hurt yourself. | 15:32.36 |
henrys | I think #4 at the link is a reasonable way to do things | 15:32.39 |
Robin_Watts | Just play nicely in the sandpit. | 15:32.47 |
paulgardiner | Not sure what this is saying, but maybe: http://knowledgetransferinobjectivec.blogspot.in/2012/04/inter-app-communication-document.html | 15:33.00 |
| oh maybe not. Seems scheme driven | 15:33.39 |
chrisl | My dislike of iOS is in way no being assuaged - I'll crawl back under my Ghostscript rock, now........ | 15:33.52 |
kens | Who said command liens were too hard ? THis is all hideous.... | 15:34.27 |
ray_laptop | chrisl: pull your head into your linux shell ? | 15:34.28 |
chrisl | ray_laptop: if only! I'm looking Windows stuff for kens just now :-( | 15:35.04 |
kens | :-) | 15:35.11 |
| At least I got it to compile.... | 15:35.20 |
chrisl | Yes, well, when Windows finally decides to stop updating/restarting, I'll be able to look at this properly | 15:36.21 |
mvrhel_laptop | morning. sorry I am a bit late | 15:36.37 |
kens | Hi Michael | 15:36.55 |
henrys | chrisl: unfortunately it is the likely place for us to make money. iOS consumers actually pay for apps unlike what seems to be happening in the Android market | 15:38.05 |
Robin_Watts | henrys: urm... really? | 15:38.31 |
kens | is unconvinced | 15:38.48 |
henrys | yes I posted a forbes article sometime back with numbers | 15:39.05 |
kens | Must have slipped my mind, I do recall reading it | 15:39.23 |
henrys | everyone I know doing development on mobile thinks android is no place to sell something but that is more anectdotal | 15:39.58 |
Robin_Watts | henrys: right. and from memory that had ios and android both worth naff all, but MS surprising worthwhile. | 15:40.03 |
| probably because there is no competition on ms yet. | 15:40.29 |
kens | Presumably because there are so few apps for MS Phone that everyone reluctantly has to buy them ? | 15:40.32 |
chrisl | paulgardiner: we are scheduled a "proper" release in just over a month, maybe we could wait until then? | 15:40.45 |
Robin_Watts | http://www.forbes.com/sites/tristanlouis/2013/08/10/how-much-do-average-apps-make/ | 15:40.49 |
| no, sorry, you are right. ios is worth 10x as much as android per download for a paid app. | 15:41.31 |
kens | Seems to be the case yes | 15:42.05 |
mvrhel_laptop | being the more expensive device, I suppose it makes sense that they would be more willing to spend money for apps | 15:42.07 |
henrys | Robin_Watts: that arcticle does show money is being made. These are averages. A powerful app say like "Office" should do very well. | 15:42.20 |
kens | Apple and MS about the same in terms of revenue | 15:42.27 |
Robin_Watts | henrys: except I suspect you'll get far fewer downloads for expensive apps. | 15:42.44 |
| but, there is progress to report on the GhostDocs stuff... | 15:43.22 |
henrys | it is what you would expect culturally - I mean the difference between the platforms if we through android in with linux and open source it looks like the desktop market. Except the numbers are smaller | 15:43.35 |
paulgardiner | chrisl: I'm in no hurry. It was just that I'd fixed a number of problems and it seemed viable for a release. I hadn't realised that there were things other than add the new MuPDF features that we needed to do. Certainly makes sense to get file opening sorted. | 15:43.45 |
henrys | s/through/throw | 15:43.50 |
Robin_Watts | I got the FBS to the stage where we can build exes. | 15:43.58 |
| I haven't tested the exe, but we can build it. | 15:44.07 |
paulgardiner | And I have the testshell producing PDF files | 15:44.27 |
henrys | paulgardiner: wow greate | 15:44.36 |
| Robin_Watts: cool | 15:44.42 |
paulgardiner | Getting the app to call it is looking hellish though. | 15:44.52 |
Robin_Watts | The current test-shell build requires things like fonts and resources on disc. | 15:45.22 |
paulgardiner | Actuall not true. Easy to call the command line, but getting the exe included in the build and then finding where it is. | 15:45.29 |
| Robin_Watts: just getting the exe on and calling it is looking nasty so far | 15:45.50 |
Robin_Watts | There is a mechanism within the picsel build system for building a "rom filing system", akin to the gs one. | 15:45.50 |
| so if we can tweak the build to use that, it will mean we can have a single self contained exe rather than an exe + a tree of files. | 15:46.30 |
mvrhel_laptop | sounds like some nice progress | 15:46.55 |
chrisl | paulgardiner: given everything else you've got going on, is the file opening on iOS likely to get a look-in in the next 6 weeks or so? | 15:47.36 |
Robin_Watts | I was thinking that I might spend some time looking at trying to make the romfs build work? That's independent of paulgardiners work, so I won't be getting in his way. | 15:47.48 |
marcosw | we should have a projector at the staff meeting so that paulgardiner can give a demo/tutorial at the staff meeting | 15:48.10 |
paulgardiner | chrisl: difficult to know. The initial office stuff could just drop out, but equally well it could be hell | 15:48.17 |
| marcosw: yeah, everyone will need a nap. | 15:49.49 |
Robin_Watts | Everyone with an android device can download SmartOffice Lite for free from google play and try it out. | 15:49.50 |
henrys | marcosw: I think paulgardiner might be too busy for entertainment. | 15:49.58 |
chrisl | paulgardiner: I'm just thinking that it would be nice to start to get the iOS and Android builds to fit with the regular release schedule | 15:49.59 |
Robin_Watts | chrisl: I think miles would really like to have something to show for the japan trip (Feb 20th?) | 15:50.42 |
paulgardiner | just wondering if nap means something completely different in the US | 15:50.44 |
| Robin_Watts: Feb 20th, that's office doc work, not iOS, right? | 15:51.28 |
mvrhel_laptop | Robin_Watts: yes. something for Japan would be nice. I know Miles just bought his ticket | 15:51.28 |
henrys | the high priority thing is having a demo for Japan, if possible | 15:51.32 |
Robin_Watts | paulgardiner: yes. | 15:51.39 |
henrys | paulgardiner: presumably mvrhel_laptop will demo it on the nexus yes? | 15:52.08 |
mvrhel_laptop | yes that would be ideal | 15:52.17 |
Robin_Watts | henrys: The valgrind work for mupdf is getting there. | 15:52.33 |
chrisl | But presumably that doesn't have to be a full release to Google Play? Just an apk to stick on a device or two would be enough | 15:52.34 |
paulgardiner | We might be nearly there, but getting a command-line program included in an app and callable is looking far harder than I'd imagined. | 15:52.42 |
mvrhel_laptop | yes that would be fine (the apk) | 15:52.49 |
henrys | paulgardiner: so as soon as you have something let's get mvrhel_laptop to test it out. | 15:52.52 |
Robin_Watts | We have 3 openjpeg ones left, and I'm working on a cmap valgrind problem now. | 15:52.53 |
henrys | Robin_Watts: okay let me know if I should look at some of that. I've been inattentive to the codecs. | 15:53.47 |
Robin_Watts | henrys: Well, feel free to look at 694880 694904 694906. Or farm them out to Simon/Shelly. | 15:54.28 |
henrys | Robin_Watts: okay I'll dispatch those after the meeting. | 15:54.45 |
| Robin_Watts: hin-tak has been swinging lately too. | 15:54.56 |
Robin_Watts | Does last weeks MuPDF decision affect whether we push ahead with the jni stuff or not ? | 15:54.56 |
henrys | Robin_Watts: I think we should continue. | 15:55.22 |
Robin_Watts | tor8 and I discussed it briefly and figured it probably still made sense to pursue, but it's less of a priority now i would imagine. | 15:55.33 |
| I am tempted to put the JNI stuff as lower priority than the GhostDocs romfs stuff though? | 15:56.03 |
tor8 | I obviously haven't been paying enough attention. where does the GhostDocs name come from and what is it? | 15:56.43 |
henrys | Robin_Watts: if you think you can do something to help paulgardiner with the demo I'd say that would be more important, but I don't know if you helping is going to change the schedule. | 15:56.44 |
Robin_Watts | henrys: My stuff *might* help paul, but I wouldn't swear to it. | 15:57.16 |
| Certainly it won't hurt. | 15:57.35 |
paulgardiner | Possibly I could do with some help finding out how to include an executable in an app and call it. I'm just not finding it. | 15:57.55 |
Robin_Watts | It might help a lot if we decide to try to build a self-contained native lib that we can call. | 15:58.00 |
henrys | Oh by the way I did name this thing GhostDocs and assumed there wouldn't be objection, seemed an obvious name | 15:58.08 |
Robin_Watts | paulgardiner: Sure. | 15:58.13 |
| I thought the GhostDocs name came from Miles? (Possibly from Henry -> Miles -> Me) | 15:58.29 |
henrys | tor8: we bought a "docs" suite. | 15:58.29 |
mvrhel_laptop | tor8: impressive amount of work on the java script interpreter for such a short amount of time. def think this should be APGL. | 15:59.25 |
marcosw | did someone check if Quality Logic has a docs test package? | 16:00.10 |
henrys | marcosw: should we review any customer bugs? | 16:00.12 |
mvrhel_laptop | marcosw: I looked on their website and saw nothing wrt docs | 16:00.37 |
henrys | marcosw: I've been looking, no luck. | 16:00.46 |
| we are at the 1/2 hours anything for the meeting - kens, chrisl gsprint stuff? | 16:01.08 |
kens | henrys I sent chrisl the initial gsprint stuff, it works 'sort of' for me on WIndows 7 | 16:01.33 |
| Hopingchrisl can sort out hte build spaghetti | 16:01.45 |
chrisl | henrys: looking at the build side of that now - some issues to work out on that score | 16:01.57 |
kens | I don't have a real physical printer here, so I'm limited to what comes out of printers on FILE: and similar | 16:02.12 |
| The Adobe PDF printer refuses point blank to work | 16:02.25 |
| I have a sneaking suspicion that class 3 drivers might not work at all in this setup | 16:02.41 |
henrys | kens: what about any HP printer driver then view the pcl in ghostpcl | 16:02.53 |
Robin_Watts | paulgardiner: http://gimite.net/en/index.php?Run%20native%20executable%20in%20Android%20App | 16:03.07 |
kens | henrys All the ones I have a re version 3 and don't work on FILE: | 16:03.12 |
| henrys I could try under Windows 8 of course, assuming the drivers there are version 4 | 16:03.45 |
| My main concern was to get what I have now built in | 16:04.00 |
tor8 | henrys: right, that's what I thought, but then robin said something about romfs and I wasn't as confident anymore | 16:04.05 |
paulgardiner | Robin_Watts: yeah, I got that far, but that just runs existing commands | 16:04.08 |
henrys | kens: ugh, last I captured stuff I used tcpdump to grab packets | 16:04.17 |
Robin_Watts | paulgardiner: "OK, but how can I put my own native execulable in Android App?" | 16:04.31 |
kens | I can't do that with no actual pritner henrys | 16:04.36 |
paulgardiner | Robin_Watts: precisely | 16:04.49 |
mvrhel_laptop | brb | 16:05.03 |
paulgardiner | Robin_Watts: You are telling me I need to read on | 16:05.07 |
chrisl | kens: Is this going to mean two sets of Windows executables for future releases? | 16:05.39 |
marcosw | henrys: it looks like the Quality Logic "Application Test Suite-Intermediate Files" would be a reasonable place to start. Do you know if we already own those? | 16:05.39 |
henrys | kens:once you have something building I'll set it up here I have 2 HP printers I can use, surely we can get some decent output with that. | 16:05.43 |
kens | chrisl I certaqinly hope not | 16:05.49 |
| hgenrys what's on my repo in the XPSPrint branch builds *if* you have the Windows DDK installed | 16:06.15 |
| henrys^^ | 16:06.20 |
chrisl | kens: so I can build this with VS2005? | 16:06.29 |
kens | chrisl I did | 16:06.36 |
| But you need the Windows DDK | 16:06.43 |
| For the .h files and the .lib files | 16:06.51 |
henrys | kens:I have windows 8 but I should use 7 for this? | 16:06.56 |
| kens: I have 7 as well. | 16:07.06 |
kens | henrys either should work. I have used both | 16:07.17 |
henrys | marcosw: I don't have any idea where that would be if we had it. | 16:07.32 |
kens | On Windows 7 the only device I can get to work is the Microsoft XPS printer driver | 16:07.38 |
chrisl | I have the DDK. And the .lib is really static, we aren't dependent on an xpsprint.dll? | 16:07.42 |
kens | chrisl, I have no idea..... | 16:08.02 |
| I'm assuming if we need the DLL its part of WIndows | 16:08.15 |
marcosw | https://storefront.qualitylogic.com/p-72-microsoft-ats-ifs.aspx | 16:08.22 |
kens | Like ole32.lib and so on | 16:08.23 |
chrisl | kens: yes, but won't that fall over if the DLL isn't there, for example, on WinXP? | 16:08.54 |
marcosw | henrys: I thought you might have them, but presumably not. I suppose ray_laptop or miles. | 16:09.15 |
kens | chrisl, Hmm yes that's probably true, but we aren't supporting WIn XP any more :-) | 16:09.19 |
henrys | paulgardiner: Oh I don't know if I want to ask this⦠but what did picsel do to test this? | 16:09.27 |
kens | chrisl I can give it a try, I have 2000 and XP installed | 16:09.30 |
| (on VMs) | 16:09.34 |
Robin_Watts | henrys: Picsel had various things for testing. | 16:09.54 |
chrisl | kens: I don't want to be in the firing line for dropping XP in near future :-( | 16:10.06 |
kens | :-) | 16:10.13 |
Robin_Watts | Firstly, they had a webspider running that pulled down pretty much any file they could find of a given type on the web. | 16:10.17 |
kens | We cna make 2 executables if really required | 16:10.24 |
| One cna be called gsprint.... | 16:10.33 |
Robin_Watts | Then they had an automated testing system (known as ATS). | 16:10.35 |
chrisl | kens: 4 executables..... | 16:10.43 |
Robin_Watts | ATS does all sorts of clever things. | 16:10.48 |
kens | chrisl only if we want to worry about 32-bit legacy support | 16:11.09 |
Robin_Watts | Firstly it would do automated builds of lots of different executables for lots of different systems. | 16:11.12 |
kens | is inclined to say 'its 64-bit, upgrade!' | 16:11.29 |
Robin_Watts | so we could be sure that anything we wrote would not break some obscure build with some obscure combination of options. | 16:11.40 |
kens | Especially since the existing gsprint will work on anythign except Windows 8 | 16:11.42 |
chrisl | kens: again, can I forward the complaints to you on that, then?? | 16:12.05 |
kens | No henrys :-) | 16:12.13 |
Robin_Watts | Secondly, it would test the executables (run the exe, load files into it, take snapshots, compare snapshots to references etc). It managed that on a wide range of exes. | 16:12.35 |
henrys | Robin_Watts, paulgardiner : a good start for us is a corpus of test files. Qualitly logic would have the advantage of someone trying to use all the various functionalityâ¦. | 16:12.39 |
Robin_Watts | '/devices. | 16:12.44 |
chrisl | kens: I do have an XpsPrint.dll on my Windows7 install :-( | 16:12.55 |
kens | THen I guess its needed | 16:13.06 |
| We should therefore make a different executable I think, one for henrys to ponder | 16:13.29 |
henrys | kens: have we gotten response to your pdf from the customer? marcosw? | 16:13.34 |
Robin_Watts | In particular, as well as the massive array of overnight tests picsel did, there were a smaller set of files that were tested on every commit. | 16:13.35 |
| We haven't got a copy of the full testdocs (I asked, and the liquidators said no), but supposedly we do have a copy of the smaller set of files that were tested on every commit. | 16:14.18 |
kens | henrys, the PCL orientation one / No response other than a grdging agreement not to trim support from CCs and a vague hand wave of 'trying' to get someone to sescribe their workflow. | 16:14.25 |
chrisl | kens: we already setup paths for the Microsoft SDKs, but it looks like rather a mess :-( | 16:14.36 |
kens | chrisl I'm not terribly surprised | 16:14.49 |
| I don't think MS expect people to produce open-source devices | 16:15.01 |
marcosw | henrys: do you want to email miles asking if we have the Quality Logic ATS-IF or shall I? | 16:15.19 |
henrys | marcosw: http://msdn.microsoft.com/en-us/openspecifications/cc816059.aspx | 16:15.33 |
| does anybody have a microsoft account so we can get these tests? | 16:16.34 |
| I'll get one but I'd rather not expose myself if somebody else is already victim ;-) | 16:17.11 |
ray_laptop | henrys: I think I had one, but it's been quite a while. I'll dig around. You aren't talking about the MSDN subscription, but just a MS account, right ? | 16:18.15 |
chrisl | henrys: I'm not sure which tests we'd want.... | 16:18.32 |
marcosw | henrys: maybe I'm missing it, but I don't see any Office test files on that page, just a validation tool and standards documents. | 16:18.38 |
ray_laptop | henrys: I do have an account with partners.microsoft.com | 16:19.38 |
henrys | marcosw: I think you're right I misread it. | 16:20.00 |
Robin_Watts | I've got to shoot off to the Doctors in a mo to see about my back. I'll be back later. | 16:20.41 |
marcosw | I can certainly write a spider to find .doc/.ppt/.xls files on the interwebs, but there are going to be a lot of them :-) | 16:20.56 |
henrys | I certainly don't mean to keep folks you can always leave after a 1/2 hour in. | 16:21.05 |
ray_laptop | henrys: when I try and sign in, it says "account does not exist". I don't mind getting one if you want | 16:21.18 |
Robin_Watts | marcosw: We have the picsel spider. | 16:21.22 |
mvrhel_laptop | I can grab any of these, but it is not clear to me which one we need. They don't seem to be a set of Office test files | 16:21.50 |
ray_laptop | Just a note, I'm getting closer on the approach to accumulate pdf14 transparency using a clist for non-clist devices | 16:22.06 |
henrys | marcosw: maybe a good strategy if we can find stuff with high page rank. | 16:22.08 |
mvrhel_laptop | s/They don't/ There does not / | 16:22.11 |
ray_laptop | Works on windows 'display' and ps2write so far | 16:22.32 |
Robin_Watts | I *thought* that the standard QL files we got came as office format files that had to be converted to pdf by printing? | 16:22.32 |
chrisl | ray_laptop: you got your gc conundrum cleared up, then? | 16:22.52 |
kens | chrisl one solution (at the expense of a more complex installer) is to provide a dummy xpsprint.dll to be installed if it doesn't exist. It need only return an error number from StartXPSPrintJob | 16:23.03 |
henrys | we do have the japanese tests where are those ray_laptop ? | 16:23.06 |
ray_laptop | Robin_Watts: some are office format, right. generally a few Excel, PPT and word ones | 16:23.10 |
marcosw | Robin_Watts: what happens after the files are downloaded? Is it practical to automatically drive office and generate a bitmap? | 16:23.17 |
ray_laptop | henrys: the JEITA ones ? I thought you got those | 16:23.45 |
Robin_Watts | marcosw: Drive office to get a 'reference' bitmap? Not as far as I know. | 16:23.53 |
ray_laptop | chrisl: yes. Thanks for trying to help. I wrote "for the logs" yesterday to you with what the issue was | 16:24.34 |
chrisl | ray_laptop: I read your post, yes, I just wasn't clear that solved the entire problem | 16:25.09 |
| kens: the required DLL might be part of the SDK redistributable stuff, we could make that a requirement for earlier Windows versions | 16:25.27 |
marcosw | Robin_Watts: so what does the spider do with the new office files? Presumably there are hundreds per day. Seems like a mechanical turk sort of problem :-) | 16:25.30 |
kens | chrisl but it won't work with earlier WIndows versions :-) | 16:25.44 |
| No XPS print path | 16:25.48 |
Robin_Watts | marcosw: I think they ran the spider for a brief period :) | 16:25.49 |
ray_laptop | chrisl: enough to give me confidence to try a cluster push (and get back a mess of segfaults). | 16:25.54 |
henrys | ray_laptop: yeah I do, that's right. | 16:25.55 |
| I have to think open office or libre has a test suite. | 16:26.10 |
ray_laptop | I've since fixed at least one (glaring) issue and am going to try again. | 16:26.22 |
Robin_Watts | basically they were in the same sort of situation that we are. They can spot changes, but initial correctness has to be established by a human. | 16:26.30 |
ray_laptop | Robin_Watts: if we can make PDF's, then we can fuzzy compare with the PDF's we have (or ones we create from *real* office) | 16:27.54 |
| (compare the rendered PDF, not the PDF's themselves, of course) | 16:28.23 |
| and probably needs very fuzzy comparison | 16:28.44 |
Robin_Watts | ray_laptop: Sounds like a nightmare. | 16:29.04 |
ray_laptop | but nothing beats having human eyes look at them, which lets me out, since I am visually impaired :-) | 16:29.27 |
marcosw | Robin_Watts: I can hack together a xdotool based script (along with Virtual Office) that can open files in office and generate pdfs. | 16:30.09 |
ray_laptop | henrys: so, did you want me to get the partners.microsoft.com account ? | 16:30.11 |
marcosw | paulgardiner: did you say .doc -> .pdf is working in GhostDocs? | 16:30.30 |
tor8 | paulgardiner: Robin_Watts: the javascript changes caused no differences in the regression. there's another commit on tor/master as well (unused for now, but affects the svg reader) | 16:31.06 |
marcosw | is looking to see if he owns a copy of office | 16:31.08 |
henrys | ray_laptop: no unless you find test files | 16:31.17 |
paulgardiner | marcosw: yes, as a stand-alone executable built via epage's build system. Tested only on Windows so far | 16:31.31 |
ray_laptop | henrys: OK. I won't. If Microsoft forgot me, that's probably for the best :-) | 16:32.05 |
henrys | marcosw: http://svn.abisource.com/abiword-testsuite/trunk/impexp/docx/ | 16:32.14 |
marcosw | paulgardiner: so if I send you a couple of .doc files you can convert them for me and email the pdfs back? | 16:32.42 |
paulgardiner | marcosw: sure | 16:33.05 |
henrys | marcosw: they even have a perl script for you / - Revision 33719: /abiword-testsuite/trunk | 16:33.25 |
| sorry I meant: http://svn.abisource.com/abiword-testsuite/trunk/ | 16:34.09 |
marcosw | henrys: np, I figured that out :-) | 16:35.06 |
| abiword seems abandoned, the source hasn't been updated since 13-Jun-2010. I'm guessing the test files are at least that old as well... | 16:39.29 |
henrys | ugh | 16:40.29 |
marcosw | found a copy of office, unfortunately it's office:mac 2011 | 16:44.07 |
ray_laptop | henrys: or the test files are newer and they showed so many problems that they gave up ? | 16:44.15 |
marcosw | actually the svn logs show updates in 2014, all by the same person (linas). | 16:45.33 |
| there svn log messages are "terse" :"A better fix than the previous one" | 16:46.07 |
| s/there/their/ | 16:46.13 |
| I have to run to uni. be back later today. | 16:58.54 |
kens | is off too, goodnight all | 17:02.37 |
Robin_Watts | is back. Did I miss anything ? | 18:08.33 |
| tor8, paulgardiner: ping | 18:37.10 |
| http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=d695605c5d91b951a9c37f64a843e92caaa178e9 | 18:38.17 |
henrys | Robin_Watts: the 3 bugs are just valgrind's right ? no crashes? | 18:40.45 |
Robin_Watts | henrys: Yes, I believe so. | 18:40.56 |
| One of them used to be a SEGV bug, but that was fixed, and now it's just a valgrind issue. I think. | 18:41.20 |
henrys | Robin_Watts: so zeniko is not well placed to work on those, I don't know about shelly. | 18:41.38 |
Robin_Watts | I have a fix for the last non-codec bug now. | 18:41.48 |
| possibly not, though I thought he'd fixed some of the valgrind problems earlier in this set... maybe as a side effect of 'real' issues. | 18:42.27 |
chrisl | henrys: Shelly uses both Windows and Linux as required, so he can do valgrind issues if we need him to. | 18:42.44 |
henrys | I'd like to have a different priority bug that we can do 100.00 for - it makes sense to do these 3 at 100.00 each for example. | 18:44.03 |
| we could use P5 for that. | 18:44.32 |
Robin_Watts | yeah, Shelly has certainly helped with valgrind stuff in the past. | 18:44.59 |
chrisl | Or minor/trivial classification | 18:45.00 |
henrys | chrisl: yes that would be better, use the importance field | 18:47.37 |
Robin_Watts | ok, so I'm going to dive back into the GhostDocs mkromfs stuff. | 18:48.23 |
henrys | chrisl: are you still updating this awki business? Should I change it there or just do the html directly | 18:52.50 |
| ? | 18:52.53 |
chrisl | henrys: I'm still using awki, until I get the important pages switched over to the mupdf style | 18:53.49 |
| henrys: if you'd prefer, you can mail me the changes, and I'll do them. | 18:54.07 |
henrys | chrisl: okay | 18:55.32 |
| chrisl: something like changing the "Accepted fixes â¦" paragraph with: Accepted fixes for bugs at P1 and P2 pay a bounty of US$1000 | 19:14.13 |
| each. Bugs at lower priority and normal importance pay US$500 per bug. | 19:14.13 |
| Bugs designated with trivial or minor importance will pay a negotiated | 19:14.15 |
| amount less than or equal to US$100 per bug. | 19:14.15 |
Robin_Watts | how about "a negotiated amount (typically US$100 per bug)" ? | 19:14.56 |
| as that gives us more scope ? | 19:15.10 |
Robin_Watts | shuts up. being nosey. sorry. | 19:15.56 |
henrys | I'm alright with that either way | 19:16.35 |
chrisl | henrys: I think I prefer Robin's - you know how initially simple problems can spiral especially with Ghostscript | 19:17.13 |
henrys | chrisl: okay | 19:17.29 |
| for our next trick we have to alert folks the policy has changed. | 19:20.51 |
| for now I'll add it to relevant bugs when I make them bountiable | 19:21.49 |
Robin_Watts | sounds ideal. | 19:21.59 |
chrisl | Edits done: http://ghostscript.com/Bug_bounty_program.html | 19:22.32 |
Robin_Watts | US$100, rather than $100 for consistency? | 19:23.05 |
henrys | I'm not big on parens there but I'm ok other then the US$ I'm good | 19:23.47 |
| s/then/than | 19:23.54 |
chrisl | Oops, I thought I did put that - obviously getting late for me.... | 19:24.03 |
Robin_Watts | ooh, The Following S2 starts tonight. | 19:24.22 |
| I bet you got that yesterday or something, Henry? | 19:24.36 |
henrys | Robin_Watts: yes 2 nights ago I think? | 19:24.51 |
chrisl | henrys: if you prefer, it would be equally correct to have: "pay a negotiated amount, typically US$100 per bug." | 19:25.11 |
Robin_Watts | chrisl: that sounds fine to me. | 19:25.29 |
henrys | chrisl: yes the comma is what I'd have used. | 19:26.19 |
chrisl | OKay, I've made that change, too. Looking at it now, I prefer it, too. | 19:26.45 |
henrys | great I'll make the bugs bountiable in a few minutes | 19:28.48 |
Robin_Watts | henrys: https://www.humblebundle.com/ | 20:36.56 |
Gigs- | need attachment to bug marked private please http://bugs.ghostscript.com/show_bug.cgi?id=694874 | 20:46.37 |
Robin_Watts | done? | 20:49.15 |
Gigs- | thanks | 20:50.04 |
| my coworker uploaded that which is why it's so huge | 20:50.11 |
mvrhel_laptop | bbiaw | 21:17.21 |
henrys | Robin_Watts: wow didn't even know there was an audio for blood meridian | 21:27.25 |
| Forward 1 day (to 2014/01/22)>>> | |