| | <<<Back 1 day (to 2012/06/19) | 2012/06/20 |
| ray_work | marcosw: have you determined why "miles" and "kilometers" cluster servers are down ? | 01:18.02 |
| henrys | chrisl:for the logs, I there some nonsense I need to build 8.70, the jasper configure file fails for me? - something about a /lib/cpp failing the santiy check. | 04:00.57 |
| mvrhel | henrys: so I am close to having this copy alpha stuff done. on my last forced run with AA on, most of the remaining segvs occur with the round trip ps2write pkmraw and ps2write ppmraw. Not sure what is going on there. I will take a look | 05:28.36 |
| | this should not be from my changes | 05:28.49 |
| | hmm ps2write might not like the forced setting of alphabits | 05:43.06 |
| | and that is what it is | 05:46.08 |
| | ok 13 segvs left for me to figure out | 05:49.46 |
| | aha. pattern accumulator.... | 05:55.25 |
| kens | chrisl ping | 07:56.56 |
| chrisl | kens: pong | 07:57.25 |
| kens | I can't seem to find a /W2, /DW2 or ./internalW2 in the font dictionary | 07:57.44 |
| | Even though it does have a /CDevProc | 07:57.57 |
| | What am I doing wrong ? | 07:58.09 |
| | I made 2 routines based on zchar_get_CDevProc | 07:58.26 |
| chrisl | Hmm, well, there are other source of a CDevProc than our PDF interpreter...... | 07:58.31 |
| kens | D'oh! | 07:58.40 |
| | I'm using the wrong test file.... | 07:58.48 |
| kens | smacks head | 07:58.55 |
| chrisl | That would not help :-) | 07:59.01 |
| kens | Do not use the PostScript test file when chsecking for PDF entries.... | 07:59.22 |
| chrisl | Been there, done that...... | 07:59.52 |
| kens | Well, thanks for the 'carboard prgrammer' moment... | 08:00.13 |
| chrisl | np | 08:00.30 |
| Robin_Watts | tor8: ping | 09:06.15 |
| tor8 | morning robin | 09:06.21 |
| Robin_Watts | morning | 09:06.48 |
| | could I trouble you to cast your eyes over: http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=137c9615245ae110b2077191048906a8055660cf please? | 09:07.01 |
| tor8 | looks better to me; what's your opinion on it? | 09:08.48 |
| Robin_Watts | I'm happy with it. The only qualm I have is the amount of boiler plate for every accessor function. | 09:09.18 |
| tor8 | yeah, not thrilled about that part either, with the fz_foo, xxx_foo_shim, xxx_foo and initializer stuff it adds up to quite a bit | 09:10.24 |
| | but I like not passing void* everywhere | 09:11.19 |
| Robin_Watts | We can avoid the shim layers. | 09:11.36 |
| tor8 | dangerously, by type casting the function pointers you mean? | 09:11.59 |
| | though to be fair, the shim implementation is also prone to type casting typos | 09:12.36 |
| Robin_Watts | If we define a function types for the fz_thing, then we can cast the pdf_thing to the fz_thing when we assign the function pointer. | 09:12.39 |
| tor8 | yeah. I thought about that. maybe we should. | 09:13.17 |
| Robin_Watts | It leads to faster performing code. | 09:13.31 |
| | and it keeps the nastiness to a single point (where we set up the function table), which is broadly equivalent to the nastiness of casting all the args in the shim functions. | 09:14.06 |
| tor8 | or, cast the function pointer to (void*) instead of adding a typedef and let c do its voodoo | 09:14.15 |
| | doc->super.first_annot = (void*)pdf_first_annot; | 09:14.33 |
| Robin_Watts | tor8: That might work (or to void (*)(void)) | 09:14.43 |
| | no, yours is better. | 09:15.09 |
| tor8 | void (*)(void) is not a void pointer, it'll make the compiler barf :) | 09:15.12 |
| Robin_Watts | Some compilers dislike changes in the level of indirection. | 09:15.39 |
| | hence casting a function to (void *) can be complained about, I think. | 09:16.00 |
| tor8 | a function pointer is a pointer, like void*, and void* can be implicitly converted to any pointer | 09:16.08 |
| Robin_Watts | well, we can try it - you may well be right. | 09:16.27 |
| tor8 | so I don't think so, but I could well be wrong. function pointers are obscure things in a lot of compilers | 09:16.30 |
| | I've done the (void*) trick with function pointers before. want me to have a go at eliminating the shims or are you already knee deep in the editor? | 09:16.56 |
| Robin_Watts | tor8: I'm not yet. I'm about to go for a run, so if you want to, go for it. | 09:17.25 |
| | Or I'll do it when I get back. | 09:17.30 |
| tor8 | Robin_Watts: a git trick I learned about yesterday: "git remote update" | 09:17.44 |
| | similar to "git fetch --all" | 09:17.50 |
| Robin_Watts | ooh. | 09:18.02 |
| | I'd like to add an enumerated type accessor function to fz_annot. | 09:18.30 |
| tor8 | Robin_Watts: I'll bang on it; do you want it on master or on forms? | 09:18.35 |
| Robin_Watts | At the moment, on forms, I guess. | 09:18.49 |
| | although if you do it on master, I can merge across. | 09:19.14 |
| | so whatever you feel is best. | 09:19.19 |
| tor8 | forms has your annotation shims, so may as well do it there | 09:19.51 |
| Robin_Watts | let me push my commit to forms then? | 09:20.07 |
| | done. | 09:20.35 |
| tor8 | the branch graph looks a bit confusing with all the merges going on :( | 09:21.08 |
| Robin_Watts | tor8: hopefully it should just be repeated merges of master into forms from now on. | 09:22.25 |
| tor8 | you and paul are both using windows on the forms branch right? | 09:23.53 |
| | we'll need some makefile variable hacking for the js implementation switching :) | 09:24.14 |
| | Robin_Watts: eh, your commit doesn't actually compile... wrong prototype in fitz.h | 09:25.29 |
| Robin_Watts | tor8: We are. | 09:33.01 |
| | really? | 09:33.04 |
| | builds fine here. | 09:33.41 |
| | trying a rebuild. | 09:34.06 |
| tor8 | you forgot the doc argument on fz_next_annot prototype | 09:34.32 |
| | I have an amended commit and a few build fixes on tor/forms | 09:34.56 |
| Robin_Watts | yeah, and on fz_bound_annot | 09:35.09 |
| sebras | tor8: does */forms build on x11 yet? | 09:35.14 |
| tor8 | sebras: making it do just that now, but without the v8 js engine | 09:35.29 |
| | I added a stub/lazy wintextinput that uses fgets :) | 09:35.41 |
| Robin_Watts | tor8: Do you want me to force push a fixed version? | 09:35.44 |
| tor8 | Robin_Watts: please do, I think we should make sure all commits at least compile so we can bisect | 09:36.05 |
| Robin_Watts | tor8: yeah. | 09:36.21 |
| | sorry about that. done. | 09:36.28 |
| tor8 | in this case it doesn't really matter, since it doesn't compile on x11 anyway, but for procedure's sake :) | 09:36.45 |
| sebras | tor8: fgets ok with me. | 09:37.58 |
| | tor8: I just want to try out the */forms branches... | 09:38.17 |
| | tor8: couldn't you have the forms branch re-use the search input box? | 09:53.21 |
| | tor8: a bit trickier but aalso a bit more nice. :) | 09:53.43 |
| tor8 | I'll leave that as an exercise for the reader ;) | 09:54.05 |
| | the search input box is a terrible terrible hack | 09:54.15 |
| sebras | tor8: ok, make sure fgets works and I'll do the next terrible hack. | 09:54.25 |
| | tor8: I'm sure the dialog box in the win32 case is not what has been planned as a solution. | 09:55.34 |
| | tor8: so having a equivalent terrible^3 hack for x11 seems ok to me. | 09:56.15 |
| | I saw something about Robin wanting to implement caret things to be able to write into the PDF page itself. that seems like a good solution to me (and the same as in acroread). | 09:56.56 |
| tor8 | Robin_Watts: http://git.ghostscript.com/?p=user/tor/mupdf.git;a=commitdiff;h=acfcf144a708aa3afc739904dfafccbec48e64bb;hp=69bcec1c5204bcc64a9405c818e65cc260fb0078 | 09:57.32 |
| kens | chrisl it 'looks like' the parentheses problem is somethgin to do with the CID, when I look at the output file there is no /W, /W2 entry for the parentheses, but there is for the 'C'. The parntheses have really odd CIDs of 0x1921 and 0x1922 while the C is a reasonable value of 0x0026 | 10:31.31 |
| | I don't know why this is happening yet of course :-) | 10:31.40 |
| | Hmm 200+ differences I wasn't expecting :-( | 10:33.50 |
| chrisl | Yeh, I noticed the weird CID for the parentheses - seemed odd | 10:33.53 |
| kens | Odder is the fact that we don't end up with entries for them in teh Widths arrays | 10:34.13 |
| | But I need to look at my bmpcmp first and probably then work out what I've broen this time | 10:34.33 |
| chrisl | It's probably 15-20 to twenty actual tests changing | 10:35.23 |
| | Er, you know what I mean.... | 10:35.43 |
| kens | yes... | 10:36.34 |
| | I'll get a better idea in a minute, bmpcmp is nearly finished | 10:36.50 |
| Robin_Watts | tor8: Lovely. | 11:13.44 |
| kens | lunches | 11:22.29 |
| ghostgum | .quit | 12:39.27 |
| Robin_Watts | paulgardiner: ping | 14:12.13 |
| paulgardiner | pong | 14:12.20 |
| Robin_Watts | http://git.ghostscript.com/?p=user/paulg/mupdf.git;a=commitdiff;h=8af238e4206f6519711a4bc9f8e9f2b8c509902e | 14:12.28 |
| | Second section of the diff... looks like debug code to me. | 14:12.45 |
| paulgardiner | Aagh. Yes. Done it again! | 14:13.36 |
| Robin_Watts | As a rule, don't mark bugs as being closed until we've published the patch because if problems are found in review, you have to change the bug message too. | 14:14.50 |
| | I've been bitten by that in the past :( | 14:15.02 |
| paulgardiner | Oh right ok. I did wonder. | 14:15.23 |
| Robin_Watts | I found the commit message slightly confusing. | 14:15.37 |
| paulgardiner | I've fixed that and pushed = not fastforward | 14:15.38 |
| | Robin_Watts: Ok ok. In what way? | 14:15.55 |
| Robin_Watts | oh, yes, one other thing... | 14:15.55 |
| | commit messages should have the first line fit on 1 line. | 14:16.18 |
| chrisl | <50 characters, ideally | 14:16.34 |
| Robin_Watts | If I'm understanding the commit correctly, you now ALWAYS make an annotation to go with a widget. | 14:17.26 |
| paulgardiner | Ok. Where have I not done that? | 14:17.30 |
| Robin_Watts | even if it has no appearance stream. | 14:17.43 |
| | Sorry, I'm misreading the web views display of your commit message. What you have done is absolutely perfect. Ignore me. | 14:18.17 |
| paulgardiner | Robin_Watts: always make a pdf_annot type object for a PDF annotation even if it's invisible | 14:18.26 |
| Robin_Watts | paulgardiner: Right, that's a better description, IMO. | 14:18.42 |
| paulgardiner | I'll change that too | 14:18.52 |
| Robin_Watts | I read the commit as "only make widgets that have an annotation", which is precisely backwards. | 14:19.13 |
| paulgardiner | Updated | 14:22.26 |
| | Hang on a minute | 14:23.46 |
| | No. It's ok. | 14:24.30 |
| Robin_Watts | paulgardiner: How hard would it be to add a 'type' field to the annotations? | 14:32.51 |
| | My mujstest stuff should treat comboboxes differently to text fields etc. | 14:33.21 |
| | and probably the cursor in the viewer should change differently too. | 14:33.46 |
| | I've committed the thirdparty stuff to the mupdf branch on my v8 repo. Likewise the gyp checkout. And I've added a make.bat. Just waiting for the push of that to casper to go through. | 14:35.00 |
| paulgardiner | Robin_Watts: pdf_form.c has a static function get_field_type which returns an enumerated type. | 14:35.01 |
| | I may have already made a public version | 14:35.28 |
| Robin_Watts | At the widget level, not the annotation level. | 14:36.11 |
| | ? | 14:36.14 |
| paulgardiner | Yeah, but internally I reckon they are the same thing | 14:36.49 |
| | ish | 14:37.23 |
| Robin_Watts | right, I don't want to be doing string comparisons etc every time I query this, so can I query it once when we create the annotation entry? | 14:37.24 |
| | and store it in the annotation entry? | 14:37.39 |
| paulgardiner | Sure | 14:37.46 |
| Robin_Watts | cool. | 14:38.02 |
| paulgardiner | I don't believe JavaScript can change the type of an annotation | 14:39.30 |
| | Hmmm, merging an FDF file could potentially, but it would be a weird thing to do. | 14:40.29 |
| Robin_Watts | well, we'd need to regenerate the annotation at that point too, right? | 14:41.21 |
| kens | OK cluster push, once more with feeling.... | 14:41.50 |
| paulgardiner | Robin_Watts: FDF could update the flags. | 14:42.07 |
| | My tendency would be not to worry about optimisations for now. | 14:42.29 |
| henrys | assumes marcosw will field support now. | 14:43.27 |
| Robin_Watts | paulgardiner: optimisation? | 14:49.02 |
| paulgardiner | Caching the value to avoid the strcmps | 14:49.24 |
| Robin_Watts | I was thinking I'd need to query the type of every widget on every mouse move. | 14:49.53 |
| | but actually, I only need to query the type of the one I'm over. | 14:50.05 |
| | so strcmps are not a problem. | 14:50.16 |
| | BUT.... I know I'm over an annotation, not a widget. | 14:50.37 |
| paulgardiner | Ah good... although now you've explained I can see why it might have been important | 14:50.43 |
| Robin_Watts | so I do need to be able to go from annotation -> type. | 14:50.53 |
| | And the easiest way to do that may be to cache the type in the annotation because I can query the widget when the widget/annotation are created, right ? | 14:51.15 |
| paulgardiner | Yeah | 14:51.27 |
| Robin_Watts | or is there some way to go from annotation -> widget? | 14:51.32 |
| | Hmm. Maybe there should be a way to go from annotation -> widget, and I should be doing that. | 14:52.20 |
| paulgardiner | You could maybe have a convert-to-widget fn that returns NULL if it isn't | 14:52.54 |
| Robin_Watts | yeah. | 14:53.04 |
| paulgardiner | But on the other hand, perhaps all that I have done with fz_widget should really be more generally applicable to fz_annotation | 14:53.27 |
| Robin_Watts | Well, I do like the fact that widgets separate out the "interaction" nicely. | 14:53.59 |
| chrisl | henrys: your jasper problem with 8.70 was a bug with the jasper configure (or rather, I think, the autotools version used to create that version of it): see http://git.ghostscript.com/?p=ghostpdl.git;a=commit;h=f3429e21 | 14:54.28 |
| henrys | ah thanks chrisl, I don't know how frequent that kind of stuff is but it's got to be maddening when doing a bisect. | 15:02.56 |
| chrisl | henrys: doesn't affect me, I have a g++ installed..... | 15:03.47 |
| | henrys: you can of course build without jasper | 15:04.33 |
| henrys | chrisl:oh so I'll just install g++ also. I didn't have a problem on the mac just linux | 15:05.54 |
| Robin_Watts | Poor mvrhel. | 15:21.46 |
| | (Well, poor Stephanie...) | 15:21.51 |
| alexcher | henrys marcosw: The sample file from RSP-4187 uses JPX streams. We render it correctly since v. 9.05. | 15:25.41 |
| henrys | alexcher:I'm out of the support business marcosw is back - yeah! | 15:48.03 |
| kens | Just in time for Gemma :-) | 15:48.15 |
| chrisl | henrys: I don't understand that last mail from Norbert - if they're using UFST, surely they are using the Microtype fonts? | 15:53.32 |
| henrys | chrisl:he wants to add fonts | 15:54.54 |
| | tt fonts | 15:54.59 |
| | like cartride fonts on older style printers. | 15:55.14 |
| kens | chrisl my latest code shows one definite progression and one apparent regression, both with PostScript files, but all other files are OK, and the test file is OK apar from he parentheses. SO progress of a sort | 15:56.24 |
| chrisl | henrys: Oh, I read it like he thought we were using UFST for the TTF fonts we supplied | 15:56.43 |
| | kens: That's sounds good | 15:57.00 |
| kens | Its 'close' | 15:57.08 |
| | I'll look at the apparent regression tomorrow, and then that leaves the parentheses to go | 15:57.24 |
| chrisl | Is that your current bmpcmp results? | 15:57.55 |
| kens | Yes | 15:58.00 |
| | Have a look if you like | 15:58.05 |
| | But don't do any work on it :-) | 15:58.13 |
| | Got to go | 15:58.16 |
| | Goodnight all | 15:58.20 |
| chrisl | Not going to touch it - gnite! | 15:58.26 |
| | henrys: well, additional TTFs will be able to be passed to UFST through the FAPI interface | 15:59.32 |
| paulgardiner | Robin_Watts, tor8: Any reason why I shouldn't have a single string object appearing twice as values of different fields in a dictionary object? I guess ownership will be fine because of ref counting, but maybe there is someother sought of problem? | 16:00.26 |
| Robin_Watts | paulgardiner: a pdf dictionary? | 16:01.05 |
| paulgardiner | Yep | 16:01.18 |
| Robin_Watts | I don't see a problem. As you say the refcounting should just have 2 refs to the pdf string object. | 16:01.46 |
| paulgardiner | And there aren't methods to update a string object that might lead me to unintentionally updating both when wishing to change just one. | 16:02.52 |
| Robin_Watts | I'd hope not. | 16:03.57 |
| henrys | what is the prefix for france? | 16:05.26 |
| | I have some number that start +33 368... | 16:05.45 |
| Robin_Watts | 33 ? | 16:05.45 |
| henrys | it doesn't go through for me. | 16:06.06 |
| Robin_Watts | So 00 33 368 ... | 16:06.08 |
| | oh, 011 33 368... | 16:07.01 |
| | (in the states it's 011, right? ) | 16:07.11 |
| | henrys: If you want to PM me the number, I can call in and say you're on your way? | 16:08.04 |
| | mvrhel_laptop: Hey, sorry to hear about Stephanie. Send her our love. | 16:22.23 |
| mvrhel_laptop | ok thanks. I am at the hospital now. Surgery is in a couple hours | 16:22.50 |
| Robin_Watts | Buy her some chocolates :) | 16:23.06 |
| mvrhel_laptop | good idea :) | 16:23.16 |
| henrys | Robin_Watts:I ended up having them call me, but I should look into it I thought I did everything righ. | 16:34.48 |
| | mvrhel_laptop:wow glad you got her in quickly. | 16:35.33 |
| mvrhel_laptop | yes. she actually drove herself at about 4am after doing a self diagnostic on the internet.... | 16:36.07 |
| henrys | mvrhel_laptop:oh you weren't there? | 16:37.17 |
| mvrhel_laptop | it is a tricky thing to diagnose without imaging and getting a white blood cell count | 16:37.19 |
| Robin_Watts | and you woke up to find a note saying "have gone to hospital" ? Clearly either you're a very heavy sleeper, or she's much more stoic than my missus :) | 16:37.32 |
| mvrhel_laptop | no. I had to stay with the kids | 16:37.34 |
| henrys | ah good job waking up and all ;-) | 16:38.37 |
| mvrhel_laptop | hehe | 16:38.44 |
| | I am trying to make up for not driving her now.... | 16:38.56 |
| Robin_Watts | I am reminded of that bit in "When Harry Met Sally" where he says that he avoids the inevitable "How come you never drive me to the airport any more?" questions when a relationship matures by never driving anyone to the airport. | 16:40.13 |
| | "How come you never drive me to the hospital any more?" | 16:40.24 |
| henrys | ;-) | 16:40.53 |
| | something you might here when child number 4 or 5 is coming. | 16:41.22 |
| mvrhel_laptop | :) | 16:48.32 |
| henrys | mvrhel_laptop:I just got off the phone with our customer we want to upgrade... I'm not hopeful this will happen in the forseeable future. But they do want a list of upcoming color features. Maybe we can talk about that next meeting. | 16:49.25 |
| | has there been marcosw sighting today? | 16:51.28 |
| Robin_Watts | I don't think so. | 17:01.32 |
| | OK. I have v8 built on linux, and I've got local changes to the makefile so it builds mujstest and mupdf-v8 if the v8 libs are present. | 17:10.37 |
| mvrhel_laptop | henrys: good grief. I would think the features we have now compared to 8.71 would be enought | 17:18.33 |
| henrys | they aren't complaining there is a dearth of features they just want to be kept up to date. I specifically asked are there any missing features holding up integration to which they said no. | 17:19.39 |
| mvrhel_laptop | ok | 17:19.52 |
| Robin_Watts | henrys: I suspect they won't upgrade until we start telling them "no, we can't provide bug fixes for that version". | 17:21.23 |
| mvrhel_laptop | yes. | 17:21.32 |
| henrys | mvrhel_laptop:I think we have, haven't we? | 17:21.53 |
| mvrhel_laptop | they recently asked for some fix and I thought marcos found something for them | 17:22.47 |
| henrys | ray_work:are you around anything to say about craig's problem. | 17:25.53 |
| | ? | 17:26.22 |
| Robin_Watts | tor8, paulgardiner: ping | 17:28.30 |
| | 4 new patches on my casper repo. tor8 probably needs to look at the Makefile one at least. The others are more in paulgardiners area. | 17:30.02 |
| | all are tiny though. | 17:34.11 |
| henrys | alexcher:are you around? | 17:34.25 |
| paulgardiner | Robin_Watts: Looks fine to me | 17:35.14 |
| Robin_Watts | Thanks. | 17:35.31 |
| | I'm about to update the solutions etc, so they will use a different directory structure under v8-3.9 | 17:35.58 |
| | (to match the positions of the built versions) | 17:36.18 |
| | and that will mean we pick up release or debug as appropriate. | 17:36.32 |
| chrisl_away | henrys: If you think it expedient, and if Alex is already busy, I'll look into some kind of woraround for Craig tomorrow - I feel it important that it's clear that is *not* our bug, though | 17:40.35 |
| | s/woraround/workaround | 17:40.57 |
| henrys | chrisl_away:isn't this solved by the customer changing their tray config. | 17:41.49 |
| | ? | 17:41.50 |
| | for the device? | 17:41.55 |
| chrisl_away | Well, in "real" printers, trays are usually physical entities, and the software changes to suit them - in this case, though, the software isn't actually integrated with the hardware, so..... | 17:42.57 |
| alexcher | henrys: I'm back from lunch. | 17:43.20 |
| henrys | alexcher:have a look at the support email and let's decide what we want to do. | 17:44.49 |
| chrisl_away | henrys, alexcher: if you want me to look into it, leave a note on here, and I'll see it in the logs....... | 17:45.43 |
| henrys | chrisl_away:I thought we had a tray default setup in the startup code. | 17:45.45 |
| alexcher | henrys: IMHO, the user need an IdiomSet resource that switches off tray selection. | 17:45.55 |
| chrisl_away | henrys: we do, that's the problem - they're basically complaining because the order of our default tray selection changed | 17:46.29 |
| henrys | right so the customer can put back the old defalt for their product yes? | 17:47.12 |
| chrisl_away | Hmm, possibly, but it's not a simple reversion of the change, the way it gets setup has changed | 17:47.56 |
| | henrys: FWIW, I think alexcher's idiom suggestion is a reasonable suggestion | 17:48.46 |
| henrys | well I leave it to you guys and I'm more than happy to say "invalid" to the customer. | 17:48.52 |
| | let me know. | 17:49.55 |
| alexcher | henrys: OK, I'm making an idiom and attaching it to the bug report. | 17:50.36 |
| | henrys: Should we keep such hack somewhere in the repository? | 17:51.08 |
| henrys | If I email it to the customer it will be in the support archive and we'll have it in the bug report so I think now. | 17:51.59 |
| | s/now/not | 17:52.03 |
| ray_laptop | henrys: I saw bug 693140 and there is an easy way for customer 780 to run a 'prefix' PS file to set specific tray mapping for a specific class of file (maybe based on knowledge about the target printer the driver used). I iput the comment in the bug, but I will leave contacting the customer up to you. | 18:27.50 |
| | henrys: the example works fine with their test file, BTW | 18:28.14 |
| henrys | sure I'll write back to him, alexcher so are you good with that also? | 18:29.51 |
| ray_laptop | henrys: alexcher: it's definitely more general, and less work than IdiomRecognition | 18:30.51 |
| | we can't expect (most) customers to come up with Idioms, but modifying the PS snippet they should be able to handle | 18:31.58 |
| henrys | I'm fine pointing them to 2 solutions also. Anything that wraps up my support responsibilities is fine by me. | 18:36.29 |
| alexcher | henrys: I like the idiom more because user's files, generated for different printers can contradict each other. | 18:37.13 |
| henrys | alexcher:I don't follow why your solution works for that and Ray's doesn't? | 18:41.42 |
| ray_laptop | alexcher: I assumed that they would have a prefix file that sets up for a particular printer emulation, based on their workflow | 18:44.08 |
| alexcher | henrys: What if they get 2 files. One expects tray 3 to be A4, another expects US Letter. | 18:44.11 |
| ray_laptop | different printers emulations would have different setups. | 18:44.29 |
| henrys | I see yes | 18:44.54 |
| ray_laptop | alexcher: but how would an Idiom know what the expectation was ? | 18:45.01 |
| | alexcher: there is nothing that says the same ProcSet (where the Idiom sees the setup) isn't used on printers with different tray setups, is there ? | 18:45.44 |
| alexcher | ray_laptop: The page size is set before tray selection. | 18:45.51 |
| | ray_laptop: So the tray selection is just not needed on gs. | 18:47.08 |
| ray_laptop | I did find it somewhat unusual that the BoundingBox was [ 0 0 612 396 ] which would seriously clip the image | 18:47.36 |
| alexcher | BoundingBox definition for PS files (unlike EPS) makes no sense. | 18:48.45 |
| ray_laptop | alexcher: I don't quite know where you were going to plug in the Idiom (what proc you were going to modify) so I guess I can't comment on your method, sorry | 18:48.47 |
| | this one clearly doesn't | 18:48.59 |
| henrys | ray_laptop:the instructions in the bug look clear is there a problem before I write? | 18:49.41 |
| alexcher | The /Tray procedure. | 18:49.55 |
| | According to PLRM, PS BBox is an union of all BBoxes of all pages. | 18:51.55 |
| ray_laptop | alexcher: yep, and these are A4 pages (595x842), so 612x396 is bogus (but not relevant to the solution) | 18:53.40 |
| | alexcher: if you modify the /Tray procedure to set particular mapping of tray to PageSize or just force PageSize, it still may not be correct for a different printer that uses the same driver (and ProcSet) | 18:55.16 |
| henrys | bbiab I'll look to the logs to guide my customer followup | 18:57.41 |
| alexcher | We don't know the details. The user is always welcome to contact us. | 18:58.01 |
| mvrhel_laptop | why do I always end up in with the crash case that is transparency with pattern clists | 20:34.09 |
| | an of course this one is psdcmyk + AA and the above stuff | 20:35.54 |
| tor8 | Robin_Watts: makefile hackery looks good | 21:02.05 |
| mvrhel_laptop | hi Robin_Watts: are you around? | 21:12.50 |
| henrys | better let marcosw take over support before I scare off the customers. | 21:50.41 |
| | Forward 1 day (to 2012/06/21)>>> | |
IRC Logs