| | <<<Back 1 day (to 2018/10/09) | 20181010 |
| ray_laptop | erunks: drawing something in a particular spot color after the rest of the page has been painted isn't that hard (with an EndPage procedure), but as was mentioned back then, does require PostScript knowledge) | 13:56.21 |
| | erunks: it's similar to adding a watermark on top of a page (there are examples of that kind of thing that I'm sure google can find) | 13:57.09 |
| erunks | That's good to know! What about the layering portion of the issue? | 13:57.37 |
| ray_laptop | as kens described there, "layer" can mean just about anything. PS paints on top of whatever's there, so whatever is painted last can be considered a layer on top | 13:58.57 |
| | and with a spot color, what you draw in that color can be "invisible" -- some apps paint a "varnish" layer or a "cut" layer in a spot color which has is defined to have 0.00 for its equivalent color in CMYK | 14:00.55 |
| | then an app (like ghostscript) that can do separations can produce that as a separate image | 14:01.45 |
| | or, if written to a psdcmyk file will have that spot color channel -- Photoshop can show it (and I think gimp can) | 14:02.44 |
| erunks | That's essentially what I'm going for. I need to be able to define a cut line in a spot color. The only issue is that the machine looks for a genuine AI/PS style layer with a name, yet somehow we can include that in a PDF if exported directly from the Adobe program | 14:03.27 |
| ray_laptop | making it a Optional Content into a PDF is more difficult | 14:03.40 |
| erunks | So trying to do that without the middle man of AI or PS is where I feel like I'm stuckl | 14:03.50 |
| | Especially if EPS files have no concepts of layers | 14:04.13 |
| | I suppose it's worth a shot though, just trying to include the spot color 'cut' layer without anything else first | 14:05.33 |
| | Thanks, Ray. I appreciate the help | 14:08.04 |
| chrisl | kens: Before you finish for the day, could I get you to look over a few commits? | 15:54.49 |
| kens | Sure we cna trade favours on that, I have one nearly ready | 15:55.02 |
| chrisl | Okay - I didn't want to bug you if you were in the middle of something horrid | 15:55.49 |
| kens | No I was weriting up my commit log for your perusal :-) | 15:56.02 |
| | I finished opening the bug report :-) | 15:56.08 |
| | top 3 commits ? | 15:56.12 |
| chrisl | Yes | 15:56.17 |
| kens | LOL so .countexecstack and .execstack finally get a real use :-) | 15:58.17 |
| chrisl | Yeh, I'm rather glad you hadn't removed them completely! | 15:58.33 |
| kens | I was circumspect :-) | 15:58.45 |
| | Fear of getting it badly wrong really | 15:59.04 |
| chrisl | Although, I'm also rather horrified we have a procedure that does what that one does | 15:59.16 |
| kens | Its no worse than many, many others :-( | 15:59.37 |
| | The operator arrays one looks good, pleased to have that fixed so quickly. Harder than I had expected | 16:00.29 |
| chrisl | Well, fixing the exploit was easy - fixing the stuff that broke because of that.... | 16:00.59 |
| kens | Subclass one looks good too. I'm *sure* I used to do that stype_is_dynamic stuff once upon a time | 16:01.22 |
| | Yeah it was teh 'follow on' that was harder than I expected | 16:01.35 |
| | Ah you altered the order so we close devices before finalizing them, that makes sense to me too | 16:02.20 |
| chrisl | Well, I think we only added the subclass finalize back in July | 16:02.51 |
| kens | Yes you're quite correct, it was part of the work Nancy did | 16:03.05 |
| chrisl | And the order of operations: clearly makes no differences to our current device roster, but it felt oh so wrong as it was | 16:03.44 |
| kens | Yeah it feels badly wrong to finalize the device before closing it. | 16:04.10 |
| | But it does feel like part of the general confusion about create/open/close/finalize | 16:04.29 |
| NancyDurgin | I remember some confusion about that finalize stuff when I was working on subclass. | 16:04.49 |
| chrisl | Well, create/finish_copy/open/close/finalize | 16:05.11 |
| kens | Yes I feel that the line Chris has put back in was present in an earlier incarnation, btu I couldn't remember what it was about | 16:05.21 |
| NancyDurgin | btw I did put a command line switch to disable my epo optimization, so if it is suspected that there is a bug related to it, you can disable it to confirm the theory or whatever | 16:06.18 |
| chrisl | NancyDurgin: I've used it a few times, yes | 16:06.42 |
| NancyDurgin | good... I am glad it's useful then. | 16:07.03 |
| chrisl | It proved it did cause a problem in one case - and proved I had a lot more work to do in a couple of others! | 16:07.33 |
| NancyDurgin | it's nice to avoid wild goose chases if it's not the problem | 16:07.51 |
| chrisl | Wild geese are a perennial problem with Ghostscript :-) | 16:08.40 |
| kens | chrisl if you could cast an eye over this: | 16:13.10 |
| | http://git.ghostscript.com/?p=user/ken/ghostpdl.git;a=commit;h=16c5d4f51d08e12c972bdb9e4fd573879cda7e87 | 16:13.10 |
| chrisl | I guess I should also see if we can hide .setglobal..... | 16:13.17 |
| kens | Hopefully it's not too offensive..... | 16:13.18 |
| chrisl | " should be absolutely fin" | 16:13.59 |
| kens | Yeah I was just fixing that | 16:14.05 |
| NancyDurgin | kens: BTW I went ahead and put in a fix in pdf_shading.c so it won't crash if colorspace is null. It's on my branch. (no rush!) | 16:14.19 |
| kens | I'll try and catch up Nancy, got a few more things to deal with | 16:14.35 |
| chrisl | kens: I assume "If its o" should be "If its 0" ? | 16:15.01 |
| kens | Huh and an o for 0 mistake in teh commit log too | 16:15.11 |
| NancyDurgin | kens: No problem | 16:15.14 |
| kens | Ah great mind chrisl | 16:15.22 |
| chrisl | kens: Oh, the actual fix is much simpler than I feared | 16:16.18 |
| kens | Yeah, once I got there, it was miniscule | 16:16.32 |
| | Getting to it was the hard part | 16:16.39 |
| chrisl | Ah, mine was the opposite! | 16:16.59 |
| kens | Yes, I'd noticed! | 16:17.09 |
| chrisl | Anyway, that looks good | 16:17.17 |
| kens | Thanks I'll go ahead with it then and tell Phil | 16:17.29 |
| chrisl | Push yours, then I'll push mine | 16:18.10 |
| kens | done | 16:18.26 |
| | chrisl I did experiment with hiding .setglobal in the course of that bug, and it seemed to be OK | 16:32.28 |
| | Probably as a result of all the oother changes we've done recently | 16:32.42 |
| chrisl | kens: I'll take a look at it - been distracted by e-mails | 16:33.38 |
| kens | Yeah me also | 16:33.47 |
| chrisl | Holy crap, it's used everywhere | 16:34.22 |
| kens | I htink that was the problem last time I tried to hide it | 16:34.39 |
| | It was too extensive for me to easily change. I assume it because the DPS version does 'special stuff' | 16:35.03 |
| | Anyway, I have to dash Stella is baby-sitting for next door this evening. Goodnight everyone | 16:35.40 |
| chrisl | Good night | 16:35.49 |
| taviso | I'm a little scared to see the SAFERERRORS change... are you completely opposed to making it on by default, and adding NOSAFERERRORS instead? | 21:51.45 |
| | Forward 1 day (to 2018/10/11)>>> | |
Log of #ghostscript at irc.freenode.net.