| <<<Back 1 day (to 2015/05/26) | 20150527 |
avih_ | !seen tor8 | 07:24.42 |
| has tor been around recently? | 07:24.57 |
kens | Yes | 07:25.01 |
| But he is about to start an extended vacation | 07:25.14 |
avih | ah, thx. extended as in months? | 07:25.44 |
kens | Well weeks at least | 07:25.56 |
avih | i see. thx | 07:26.06 |
Robin_Watts | kens: Did you unzip the entire directory structure? | 08:56.05 |
kens | Yes | 08:56.10 |
| I'm nearly finished installing VS 2013 community edition after which I'll have to get hold of the WDK | 08:56.33 |
| Which is currently refusing to download in my regular browsers. I may have to use Internet Exploder | 08:56.54 |
Robin_Watts | Yes. I had to use IE. | 08:57.43 |
kens | https://www.microsoft.com/en-us/download/details.aspx?id=42273It not working for me with IE either | 08:57.53 |
| Tried Firefox, Chrome and IE now and not downloading in any of them | 08:58.10 |
| I wonder what hte point of the checkboxes i the VS 2013 installer is. I said 'not silverlight, Blend, ....' and its gone and installed them all anyway | 08:59.37 |
| <sigh> IE had blocked a popup, which is why the downlad didn't work..... | 09:00.55 |
| Oh, and you can't install the WDK to anything except the C drive, that's just cr*p | 09:02.25 |
| Robin_Watts : did the VS 2013 installer actually exit for you ? It seems to be stuck for me..... | 09:10.04 |
Robin_Watts | kens: yeah, worked OK for me. | 09:14.55 |
| but I did leave it running while I did other stuff. | 09:15.15 |
kens | I had to kill it and now VS 2013 won't run. Guess I'll have to start again if I actually want to use it. I'm hoping I don't have to | 09:15.21 |
| Almost finished installing the WDK. SHould be done just in time for me to go out. | 09:15.55 |
| Well, it looks like inf2cat will actually run now | 09:19.46 |
| Doesn't seem to work though. II have to go out, I'll carry on later. | 09:23.57 |
tor8 | avih: morning. | 10:04.07 |
avih | tor8: hey :) hope your time off is for good reasons... | 10:04.26 |
tor8 | summertime :) | 10:04.39 |
avih | good enough in my book :) | 10:04.48 |
| so, while not of any high importance, i was thinking that the array implementation could be optimized for shift/unshift in a not too complex way, to make these ops in O(1) per item, and without meaningful hit to perf otherwise. | 10:06.19 |
| the reason for this is that arrays are the "natural" implementation for queues in javascript, and V8 already optimizes shift pretty impressively | 10:07.09 |
tor8 | avih: what sort of optimization did you have in mind? | 10:07.49 |
avih | so the only extra property an array would need to hold is "offset", which only gets modified for shift/unshift | 10:08.03 |
tor8 | remember, we don't actually treat arrays any different from other objects, they are string indexed dictionaries | 10:08.12 |
avih | exactly for this reason | 10:08.23 |
| otherwise there could be other approaches for optimizations | 10:08.38 |
tor8 | right, that might work | 10:08.44 |
avih | so the getter/setter will take this offset into account, which is pretty trivial both code wise and perf wise, since most of the perf goes on accessing the item at the tree anyway | 10:09.35 |
| i implemented such array in js just to check its feasibility, and it looks good | 10:10.12 |
tor8 | yeah, all the code goes through the "high" level accessors anyway, so tweaking it at a low enough level should be transparent | 10:10.26 |
avih | exactly | 10:10.33 |
tor8 | I'm only working today and tomorrow, before my vacation, so I don't expect I'll be able to merge any patches until I come back | 10:11.04 |
avih | and then shift is merely changing the offset value by 1, and unshift changes the offset by number of args it has, and voila. shift and unshift in O(num args) rather than O(arr-length) | 10:11.45 |
| yeah, no worries, just wanted to see what you think | 10:12.25 |
| maybe i have hidden gotchas which i didn't notice | 10:12.41 |
| (though my js implementation does seem to work well) | 10:12.51 |
tor8 | jsR_hasproperty and jsR_setproperty would be where the guts live | 10:13.45 |
| they already check if an object is an array to do the automatic length property update | 10:14.01 |
| doing a further check to apply an offset to a numeric index should be fairly simple | 10:14.28 |
avih | fwiw, that's my js implementation. http://pastebin.com/uFFWkkv3 | 10:15.08 |
tor8 | though I'm sure there are a few gotchas hidden in the code | 10:15.10 |
avih | yeah. my new array prototype supports get/set (methods rather than props), push/pop, shift/unshift, delete, reverse and change length. i wanted to see if there would be any inherent gotchas in any of those, and i don't think i could notice once | 10:16.58 |
| one* | 10:17.16 |
| anyway, enjoy your time off :) | 10:25.00 |
| it's probably possible to do reverse in O(1) too, but making the accessor taking into account both offset and the reverse bit could be a bit tricky, so my first attempt has only offset. but once offset works, i think it should also be trivial to add an inverse-direction bit | 10:28.03 |
| though now that i think of it, there could be an inherent gotcha. if you push and shift repeatedly, the offset keeps changing in one direction, which could introduce issues of integer overflow. | 12:35.54 |
| so ultimately one would need to manage those too, possibly "reindexing" the array if it gets to that | 12:36.45 |
Robin_Watts | avih: You might find that overflow in 2s complement maths just does the right thing. | 12:51.29 |
avih | possibly, but it needs to be validated | 12:51.58 |
kens | OK vs 2013 stup completed this time, so I need to restart, brb | 12:54.57 |
| Wel, I'm completely unable to make inf2cat do anythign useful :-( | 14:13.26 |
Robin_Watts | kens: Presumably you've tried following: https://technet.microsoft.com/en-us/library/dd919238%28v=ws.10%29.aspx ? | 14:16.44 |
kens | I have yes | 14:16.50 |
| #Whenever I invoke inf2cat (on any .inf file) it says the file does not exist | 14:17.10 |
Robin_Watts | Is there a particular stage where it all breaks down? | 14:17.14 |
| ah. | 14:17.16 |
| http://www.osronline.com/ShowThread.cfm?link=121726 | 14:17.44 |
| It wants the directory that contains the .inf file, not the .inf file itself. | 14:18.04 |
kens | wibbles | 14:18.16 |
Robin_Watts | Don't shoot the googler :) | 14:18.33 |
kens | I don't understand how it knows aht the device is called then, but let me try that | 14:18.50 |
| Well its doing 'soemthing' now | 14:20.06 |
| How long can this take?..... | 14:21.15 |
| Maybe I should have made a nice simple directoty for it,I bet its doing all the files in that folder, and its huge | 14:21.44 |
| Well, that's progress, I seem to have a .cat file | 14:23.44 |
| Now to do the certificate dance | 14:24.06 |
| Robin_Watts : did you add certificates to your store ? | 14:27.50 |
Robin_Watts | kens: I think so. | 14:28.54 |
kens | :-) | 14:29.09 |
Robin_Watts | I had to do something like that to extract the certs from the supplied .p12 file into the .spc .pem etc files. | 14:29.17 |
kens | I'm not totally sure how to use the certificate, but I htnk I need to start by importing the Comodo cert | 14:29.49 |
| WHich I'm just doing now | 14:29.55 |
Robin_Watts | There is a URL in the README in the Certificates dir that contains the procedures I followed. | 14:34.12 |
kens | Yeah I'm having trouble figuring out what to do with all the things you sent | 14:34.31 |
| Hmm, I installed it to my computer, rather than to Internet Explorer, but I suspec the two are equivalent | 14:35.34 |
Robin_Watts | http://www.itninja.com/question/guide-to-signing-unsigned-drivers | 14:36.05 |
| You don't need steps 2 and 3 from that. | 14:37.21 |
kens | Yeah but they all want to use signcode, which I don;t have | 14:37.29 |
Robin_Watts | signtool sign | 14:37.37 |
kens | The MS instructions say ot use signtool | 14:37.41 |
Robin_Watts | I think "signtool sign" does what signcode used to do. | 14:37.59 |
kens | Possibly | 14:38.06 |
| OK looks like I need a /f switch | 14:38.45 |
Robin_Watts | And I don't think you need any steps after 4. | 14:38.54 |
kens | No all I need to do is sign the thing | 14:39.07 |
| Hmm, well it *says* its signed it.... | 14:41.56 |
Robin_Watts | kens: Want to throw it at me, and I'll try to install it? | 14:42.12 |
kens | OK one moment | 14:42.19 |
| OK its on its way | 14:44.14 |
| Doesn't install anything when I double click it though | 14:44.33 |
Robin_Watts | kens: That's a security catalogue. It's not a signed driver | 14:53.28 |
kens | I've sent the .ppd and .inf as well | 14:53.41 |
Robin_Watts | Oh, right. They haven't arrived yet :) | 14:53.54 |
kens | SOrry :-) | 14:53.59 |
| I thnk you may need a different .cat file, which OS are you trying ? | 14:54.10 |
Robin_Watts | 7 | 14:54.17 |
kens | OK that should work | 14:54.23 |
| As long as its 32-bit | 14:54.34 |
| I forgot to add 64-bit and Windows 8 at the time | 14:54.43 |
| But when I try it here I'm still getting a security warning | 14:55.00 |
| Admittedly it isn't one of the naty warnings, so maybe that's just UAC being officious | 14:55.43 |
Robin_Watts | I think we still expect UAC to bleat at us. | 14:56.11 |
kens | Ok then tha't sprobably fine. I guess I'll boot up my WIndows 8 VM and try it there. | 14:56.30 |
Robin_Watts | We just don't expect the "Ew! This isn't signed! It'll break everything!" | 14:56.41 |
kens | Yeah I don't get that any more, it says its signed by Artifex | 14:56.54 |
| Let me mail you a new .cat, for all 4 variants | 14:57.09 |
| You hsould get the .inf and .ppd at the same time, for some reason Eudora had decided to sit on them | 14:58.15 |
Robin_Watts | Ok, when I click the .inf I get warned "Publisher: Unknown Publisher" | 15:01.23 |
| So that's not worked. | 15:01.40 |
kens | You have the .cat in the same directory ? THe .inf is supposed to refer to the .cat | 15:02.01 |
Robin_Watts | I have a GhostPDF directoru with .cat, .inf, .ppd in. | 15:02.21 |
kens | Yeah that's what I have too | 15:02.33 |
| Oh, I couldn't doa double clik on the .inf | 15:03.11 |
| I had to add the printer form the control panel | 15:03.18 |
| WHen I tried to double clik it it told me that wasn't supported | 15:03.31 |
| If I right-click on Windows 8 then I get the UAC warning, but it says its signed, I thnk | 15:04.05 |
| And that's the only warning I get | 15:04.26 |
Robin_Watts | Add what sort of printer? | 15:05.13 |
kens | 'Have disk' | 15:05.27 |
Robin_Watts | From Devices and Printers => Add a printer | 15:05.36 |
| it asks me "Add a local printer" or "Add a network, wireless or Bluetooth printer" | 15:05.52 |
kens | Add a local printer | 15:06.01 |
sebras | tor8: is there anything left to do for the gif patches? | 15:06.10 |
Robin_Watts | Use an existing port? or Create a new port ? | 15:06.16 |
kens | Use existing port | 15:06.22 |
Robin_Watts | Which port? | 15:06.28 |
sebras | tor8: if possible I'd like them to be on origin/master before your leave. | 15:06.32 |
kens | Anyo one | 15:06.33 |
| LPT1 is fine | 15:06.38 |
Robin_Watts | Ok, I added it as LPT3, and I have a Ghostscript PDF. | 15:08.07 |
| Can we really not get it so that we can double click the .inf file? | 15:08.18 |
kens | Yeah that's what I get, tho mine is on LPT1: | 15:08.19 |
| Robin_Watts : I have no idea | 15:08.27 |
| WHen I try that here i tells me that's not a supported installaton method | 15:08.44 |
| Although it gives no errors on Windows 8, I don't get a printer either :-)( | 15:09.05 |
| Well double-click for me opens the file in Notepad. RIght-click->Install says 'The INF file you selected does not support ths method of installation'. | 15:11.22 |
Robin_Watts | kens: OK, so I think we need a [DefaultInstall] section in the .inf to be able to double click it. | 15:11.54 |
kens | COUld be | 15:12.02 |
Robin_Watts | http://stackoverflow.com/questions/10830242/how-to-verify-inf-file-installation | 15:12.02 |
kens | I was looking for somethig like that | 15:12.09 |
Robin_Watts | https://msdn.microsoft.com/en-us/library/windows/hardware/ff547356%28v=vs.85%29.aspx | 15:12.37 |
kens | Greek to me :-( | 15:13.08 |
| OK so if I add that then it wants me to give it pscript5.dll | 15:14.44 |
Robin_Watts | what exactly did you add? | 15:17.39 |
kens | [DefaultInstall] | 15:17.49 |
| CopyFiles=@GHOSTPDF.PPD | 15:17.49 |
| DataSection=PSCRIPT_DATA | 15:17.49 |
| Include=NTPRINT.INF | 15:17.49 |
| Needs=PSCRIPT.OEM | 15:17.49 |
| Just a copy of the 'Installer' section | 15:18.04 |
Robin_Watts | oh, wait... | 15:18.08 |
| Note The INF file of a driver package must not contain an INF DefaultInstall section if the driver package is to be digitally signed. For more information about signing driver packages, see Driver Signing. | 15:18.20 |
kens | I was about to paste that | 15:18.25 |
| Just read it | 15:18.28 |
| So that's a no-no then | 15:18.34 |
| Seems like if its digitally signed, you can't install ths way | 15:18.46 |
Robin_Watts | Fair enough. | 15:18.55 |
| Excuse my ignorance here... this driver does what? | 15:19.42 |
kens | Nothing much | 15:19.48 |
| Technically it allows you to firstly install RedMon, then select the RedMon port during installation. | 15:20.14 |
| THen you cna have RedMon run Ghostscript in order to convert the PostScript to PDF | 15:20.34 |
| Thereby getting you a 'Print to PDF' solution | 15:20.45 |
Robin_Watts | So the port is where it gets it's input from? | 15:20.50 |
kens | No the port is where the PostScript is sent to | 15:21.04 |
| Could be a serial port, parallel port, network or FILE: | 15:21.21 |
Robin_Watts | Postscript is sent to? | 15:21.37 |
| or PDF is sent to ? | 15:21.41 |
kens | PostScript | 15:21.45 |
| THe .inf installs a PostScript driver | 15:21.54 |
| Windows really only has one PostScript driver for all the PS prtiners | 15:22.14 |
| The .ppd file distinguishes them | 15:22.21 |
Robin_Watts | oh, I see. So this .inf doesn't actually involve gs at all. | 15:22.35 |
kens | Nope | 15:22.39 |
Robin_Watts | Gotcha. | 15:22.51 |
kens | Its a mer convenience, which is why Chris and I have never been terribly bothered about it | 15:22.57 |
| A signed one is 'nice to have' but far from vital. | 15:23.14 |
| And of course it doens't work worth a damn on Windows 8+ | 15:23.26 |
| Because the printing system is utterly different | 15:23.37 |
| OK it seems that you can add our printer in WIndows 8, and it even works afterwards. Whether RedMon will work is debatable, but that's a different story. So I'll leave Windows 8 capability in the signing stuff. | 15:35.46 |
avih | tor8: re array again, i don't think i see how 1. js_setindex changes the array length. and 2. how changing the array length disposes of the truncated elements... | 15:40.55 |
| (in mujs code) | 15:41.03 |
| (js code does behave correctly, i.e. var x=[]; x[3]=3; <- x is of length 4 with first three elements empty, but i don't see where this happens at the code) | 15:42.04 |
| since js_setlength and js_setindex are super simple | 15:42.50 |
| or maybe there's some custom array magic inside js_setproperty in case the object is an array? | 15:45.30 |
| ah, there's some array specific magic. https://github.com/ccxvii/mujs/blob/d9ed73fd717ebbefe5595d139a133b762cea4e92/jsrun.c#L522 for truncating length, and few lines later for adding a value at index>=length | 16:10.43 |
tor8 | avih: looks like you found the answer yourself :) | 17:30.51 |
avih | i got lucky :) | 18:14.26 |
| Forward 1 day (to 2015/05/28)>>> | |