| <<<Back 1 day (to 2012/03/27) | 2012/03/28 |
mvrhel | ray_laptop: for the logs: starting to work on the pdf14 device stuff for tiffsep. yuck about the way that it is done now! | 00:14.51 |
webmonk | I am moving to a new server that does not have X installed. Is there any way to get ghostscript to install without X? Having trouble finding a solution online. | 02:03.30 |
ray_laptop | mvrhel: Saw your comment. It'll sure be good to have that cleaned up ! particularly the 'put_image' will be a BIG performance boost. | 04:15.45 |
mvrhel | yes. working on it right this second. this thing had to have stunk performance wise the way it is now | 04:16.26 |
kens | chrisl were you really working on that bioPDF rubbish at 6am ? | 07:12.36 |
chrisl | No, 7am, maybe.... not much better, really :-( | 07:13.11 |
kens | Madness..... | 07:13.18 |
| He should beat up on BioPDF, they took his money.... | 07:13.33 |
chrisl | He's already done that (apparently), and they've come up with nothing! | 07:13.55 |
kens | Which is exactly what I would expect, since they are parasites ;-) | 07:14.13 |
| If more people had a bad experience maybe we could extract some dosh from them for support, or put them out of business.... | 07:14.44 |
chrisl | Can you remember who the PDF creation people are that we "like"? | 07:15.04 |
kens | Um, NitroPDF maybe ? | 07:15.20 |
| I seem to have bug reports from them. | 07:16.16 |
| http://www.nitropdf.com/free/ | 07:16.44 |
| And they appear to be a paying customer | 07:16.52 |
chrisl | That's ActivePDF isn't it? | 07:16.53 |
kens | No, diffrent people | 07:17.03 |
| Actually, Ray says they are the same people | 07:17.57 |
chrisl | Hmm, have a look at Joann's latest e-mail.... | 07:18.00 |
kens | What threw me was they use different domains in their emails :-) | 07:19.44 |
chrisl | Mind you, NitroPDF is 120 usd, whilst bioPDF is 29 usd :-( | 07:19.58 |
kens | But it works and you get support :-) | 07:20.15 |
| Also I thought it was free :-( | 07:20.21 |
| The web site says it is | 07:20.49 |
chrisl | There is a free version, not sure if it's just for evaluation, or personal use, or whatever | 07:20.54 |
kens | The site syas 'no popups, no registration, no adware' | 07:21.27 |
chrisl | My intention was, if that guy comes back and says I've solved his problem, to urge him (his company) to seek a (partial) refund from bioPDF, and to point him at a more reputable vendor. | 07:22.02 |
kens | Seems reasonable. But of course if you've solved his problem.... | 07:22.24 |
chrisl | Well, the only way they could reasonably demand a refund from bioPDF would to show that there *is* a solution, and they failed to find it - so a little catch 22-ish | 07:23.24 |
kens | Hmm, debatable. | 07:24.02 |
chrisl | Well, there is nothing actually *wrong* with the PDF they're getting, or the Postscript that it's created from - it is fit for the intended purpose | 07:24.51 |
kens | Depends on the intended purpose, and what they advertise the capabilities as. | 07:25.37 |
| But in general, yes. | 07:25.43 |
chrisl | They make (not surprisingly) relatively few claims about their PDF writing product. Besides, I was fairly sure I knew the answer when I saw the Postscript file...... | 07:27.24 |
kens | Actually, frm teh description, I was just assuming that it was the fact that pdfwrite subsets fonts by default :-) | 07:27.53 |
chrisl | The subset thing is a red herring, the problem is that the driver is converting the TTF to a Type 1 (yeuch!), and mangling the font name in the process (I suspect for good reason!) | 07:28.57 |
kens | Yeah, I can see the real problem, it was the title I was originally looking at. | 07:29.30 |
| Good one from Raed overnight. | 07:29.56 |
chrisl | I didn't really follow what he was asking - although I only scanned the mail briefly | 07:30.32 |
kens | He is supplying a 'pdfmark' file to add (or more accurately re-embed) annotations in a PDF file. Clearly he knows that pdfwrite doesn't honour annotations in a PDF file, because he supplied the pdfmark file to regenerate them. | 07:30.43 |
| So he adds a 'Note' annotation. | 07:30.59 |
| Then complains that the 'stamp' annotation, which he hasn't supplied a pdfmark for, goes missing..... | 07:31.20 |
chrisl | LOL! Fairly typical query then :-) | 07:31.42 |
kens | Yes, I do have to take a quick look at the annotation he does supply though. | 07:32.21 |
| Which of course he hasn't supplied a command line for. | 07:32.35 |
| chrisl it seems to me we should close that BioPDF issue as 'invalid'. Since you did the work, do you want to do that ? | 08:19.19 |
chrisl | kens: I was hoping we get confirmation from chappy that my suggestion had worked | 08:21.21 |
kens | I was thinking of adding a comment 'we'd be intereted to hear your results, but this does not appear to be a Ghostscript problem and so we are clsoing the issue'. | 08:22.00 |
chrisl | Yeh, I'll do that. I just didn't want to "scare" him off reporting results, if the bug was closed | 08:22.38 |
kens | Fair enough, its just that it crops up in my assigned list. | 08:22.54 |
| And the severity is assigned as 'major' | 08:23.08 |
| Oops, sorry, wrong bug | 08:23.20 |
chrisl | well, it's closed now | 08:24.56 |
kens | OK I think it should be anyway :-) | 08:25.05 |
chrisl | Yeh, I was hoping he'd have tested by the time I started this morning, and I could close it then | 08:25.34 |
kens | I wouldn't be surprised ot find he has, and since it works he's just gone away happy | 08:25.56 |
| Right, back in a couple of hours. | 09:23.10 |
robin_watts_mac | Morning all. | 10:39.28 |
chrisl | hi robin_watts_mac, how are the Himalayan wilds? | 10:42.01 |
robin_watts_mac | less wild today. | 10:42.14 |
| We arrived at the hotel last night to find that there was no power (anywhere in the region). | 10:42.34 |
| Power came back, fortunately, but no wifi. | 10:43.03 |
robin_watts_mac | wonders what VDP is | 10:43.29 |
chrisl | In PDF? Variable Data Printing | 10:45.56 |
robin_watts_mac | right, just reading the logs. | 10:47.02 |
chrisl | Ah, well that's what it is..... | 10:47.26 |
robin_watts_mac | tor8, sebras: I thought I'd added basic documentation for the text extraction stuff. | 11:00.22 |
| and the link dest stuff ? | 11:00.36 |
| Did I forget to push it before I left home? | 11:00.49 |
| chrisl: thanks. | 11:02.36 |
sebras | robin_watts_mac: link_dests will not have their implementation changed, but hopefully I'll put together some docs tonight. if you did then you forgot to push... ;) | 11:12.10 |
| robin_watts_mac: I guess it depends on when tor8 intends to prepare 1.0 | 11:12.33 |
robin_watts_mac | But did I do the text device ? | 11:13.08 |
| ok, I have *some* link stuff here. | 11:15.09 |
sebras | robin_watts_mac: push it to your own branch and I'll dig into it tonight. | 11:16.03 |
robin_watts_mac | It's all pushed. | 11:16.55 |
| oh, there are still XXXX. | 11:17.22 |
| Damn. Cos I remember writing it :( | 11:17.45 |
| It must be languishing at home on my PC :( | 11:17.54 |
tor8 | robin_watts_mac: I was going to replace the XXX by a big "NOTE: this struct is scheduled for imminent change, use at your own peril" | 11:18.07 |
robin_watts_mac | Damn. I can't believe I didn't push it. | 11:18.35 |
| I actually documented everything that was there, and had a 'This interface is subject to change' note. | 11:19.20 |
paulgardiner | paulgardiner: is there any way I can get in and do it? | 11:19.30 |
sebras | robin_watts_mac: does paulgardiner have a key to your hous... :) | 11:19.40 |
robin_watts_mac | paulgardiner: I doubt that's a worthwhile use of your time :) | 11:20.26 |
chrisl | tor8: I've got an idea what the problem might be with that font issue reported yesterday - I'll post the info on the mupdf bug when I'm a bit surer | 11:20.56 |
tor8 | chrisl: the freetype gid 3 warning thing? | 11:21.15 |
chrisl | tor8: I guess so: "font rendering examples" | 11:22.09 |
paulgardiner | I'll be going as far as Charlbury tonight in any case, so if you did want me to, and there's a key I can pick up somewhere close, let me know. | 11:23.49 |
chrisl | tor8: it *looks* like at least one problem is that the font program contains scale dependent opcodes - that's what's causing Ghostscript's problem, anyway. | 11:24.14 |
robin_watts_mac | paulgardiner: I *thought* I'd pushed everything. If it's not on casper, then it's entirely possible that I've lost it :( | 11:26.41 |
paulgardiner | Eek :-) | 11:27.15 |
| Oops. That was supposed to nbe :-( | 11:27.30 |
robin_watts_mac | sebras, tor8: I've rewritten my link dest docs. | 12:27.25 |
| I was about to do the text ones, but I've run out of time. | 12:28.14 |
| I've pushed to my repo on casper. | 12:28.24 |
| I may do some more later. | 12:28.28 |
| ttyl. | 12:28.30 |
| I will check logs if you want to leave me a message here. Night all. | 12:28.41 |
kens | goodnigth robin_watts_mac | 12:28.53 |
tor8 | robin_watts_mac: okay, I'll leave those out of the rc1 but we can try to get them in for 1.0 proper | 12:28.55 |
chrisl | tor8: actually, this font's byte code is just broken in *so* many ways, I'm not going to try to patch Freetype to work with it, I'm going to augment Ghostscript's fallback code to try unhinted rendering for more error cases. | 12:44.45 |
tor8 | chrisl: and it starts with "DF" triggering the dynalab workarounds in mupdf :/ | 12:50.40 |
| otherwise we'd render unhinted | 12:50.53 |
chrisl | Actually, it looks like re-running the font program for every scale call works, but I doubt I'd get that past Werner. Although the spec is a little vague on that point....... | 12:53.19 |
| bbiaw | 13:15.28 |
Salil | henrys:http://pastebin.com/tsxEjuei Here is my code snippet. Do you think there are errors on how should I be using the API? | 13:49.49 |
robin_watts_mac | tor8: Give me 30 mins and I'll do the text device stuff. | 13:58.57 |
| tor8: Give me 30 mins and I'll do the text device stuff. | 14:09.44 |
tor8 | robin_watts_mac: okay. | 14:11.47 |
sebras | tor8: I guess it's hard to fall asleep in the himalayas... | 14:17.16 |
robin_watts_mac | My darling wife has infected me with her cold. If I fall asleep I may drown... | 14:18.04 |
henrys | Salil:it looks plausible but I didn't read it carefully, what is the problem with it? | 14:25.53 |
robin_watts_mac | tor8: Trying to push now... | 14:31.51 |
henrys | robin_watts_mac:well you did go to the pox bar right? | 14:39.51 |
robin_watts_mac | :) | 14:40.32 |
| Gah. I think they have a broadband line here, and about 200 twitchers all trying to update facebook at the same time... | 14:51.27 |
| tor8: At last. I think that's gone. | 14:52.51 |
| tor8: sebras has spotted trailing whitespace in my patch :( | 15:00.31 |
| I haven't got the energy to fight the connection to try and fix it though - sorry! | 15:00.57 |
sebras | robin_watts_mac: np. I'll fix it when I get home from work, unless tor8 has already done it. I never expected you to rewrite this in the middle of the night while on vacation... | 15:02.30 |
robin_watts_mac | It's 8:30pm here - so not that late. | 15:02.49 |
| Right. Bedtime for me though. Night all | 15:07.48 |
Salil | henrys: I have compiled the 8.63 using make framework and then using ar I have created the static lib from object files (.o s). I have this included in the XCode project and it builds and run but crashes with -15. -dDEBUG doesn't show any log. | 15:59.06 |
| henrys:If I just do make, it fails after creating mkromfs as mkromfs is created for arm7 and can't run. And from obj files I create a static file. XCode have missing dependencies in CUPS and _gs_romfs. | 16:02.37 |
chrisl | tor8: as I mentioned, for Ghostscript I've made the "retry glyph unhinted" condition happen on any byte code error now - I don't think revising Freetype to handle such a clearly out of spec font is the right solution, so..... | 16:14.46 |
henrys | Salil:I thought we agreed you were going to use the xcode project and use the gs debug target in the project. | 16:16.03 |
tkamppeter | Anyone from Ghostscript will come to the OpenPrinting Summit? | 16:16.41 |
henrys | tkamppeter:I can call in. | 16:19.22 |
| I'll ask Miles to go - he lives in the area. | 16:19.52 |
ray_laptop | hi, all | 16:23.13 |
henrys | hi ray_laptop | 16:23.59 |
ray_laptop | nitropdf is different from ActivePDF. As I understand it from Miles, ActivePDF sold 'primopdf' to Nitropdf | 16:24.02 |
ray_laptop | is reviewing the logs | 16:24.13 |
chrisl | PrimoPDF was the name I was trying to remember! Well, Nitro was close enough..... | 16:26.00 |
tkamppeter | henrys, what about mvrhel? | 16:26.23 |
henrys | whatever mvrhel thinks, Artifex will pay for his trip if he wants to go. | 16:28.30 |
tkamppeter | henrys, OK, thanks. | 16:31.57 |
ray_laptop | on the clist change to not compress bitmaps (when 'file' is used) speeds up 300 dpi pbmraw of the PLRM-3rd by about 8% -- better than nothing I guess. | 16:38.25 |
| I'm really eager to see how much the switching over to the spot color planar mem device and an optimized 'put_image' will speed up the files with transparency from cust 393 (bug 692618) -- not to mention avoiding the non-encodeable colors problem. | 16:42.55 |
mvrhel | henrys and tkamppeter: I was planning to go to the open printing meeting. I need to get that all arranged | 16:44.31 |
tkamppeter | mvrhel, | 16:45.48 |
| OK, we will let you present the state of the art an CM and GS. | 16:47.15 |
Salil | henrys: even if I make shared library (dylib) from that I won't be able to include it in the XCode project for iOS. Last time I tried with that and it compiled fine but while running it couldn't find symbols. It was looking for the library on a path where it rests on MacOS disk. | 16:47.41 |
mvrhel | tkamppeter: do I need to do any registration at all? | 16:47.53 |
ray_laptop | mvrhel: glad you are going -- just try not to pick any fights ;-) | 16:47.54 |
mvrhel | ray_laptop: I will try not to. It depends upon if anyone from the openicc group is there.... | 16:48.32 |
ray_laptop | are any of the OpenICC guys going to be there ? | 16:48.33 |
tkamppeter | mvrhel, on the PWG page there are instructions. | 16:48.38 |
mvrhel | tkamppeter: ok thanks | 16:48.48 |
tkamppeter | ray_laptop, afaik noone definitely coming yet. | 16:49.37 |
Jenkens | Howdy Chrisl. | 16:55.34 |
chrisl | Jenkens: hi | 16:55.56 |
Jenkens | We've been chatting about Bio | 16:56.06 |
| I figured this would be a faster discussion than the bugzilla | 16:56.15 |
| been banging my head on this and nitro bugs all the same... | 16:56.34 |
chrisl | Sure, but I've imparted all the wisdom I have on the subject..... | 16:57.06 |
Jenkens | Oh I know | 16:57.18 |
| just trying to figure out how biopdf mangled postscript now | 16:57.50 |
chrisl | In what way "mangled"? | 16:58.26 |
Jenkens | reading the irc log... lol beat up biopdf | 16:58.35 |
| well, honestly I don't know. PDFCreator prints these same documents, without changing any of the settings you recommended and 'works' as one might expect | 16:59.10 |
chrisl | Well, we don't take kindly to folks who use our work to make money, and don't pass some of it along to us | 16:59.25 |
Jenkens | so, i'm assuming bioPDF has somehow mangled something with postscript | 16:59.26 |
| well I don't blame you | 16:59.33 |
| Just don't be upset with me, I'm just the sucker on the end that paid out cash | 16:59.48 |
| (25 licenses, afaik) | 16:59.55 |
chrisl | If I was upset with you, I wouldn't have helped! | 17:00.09 |
Jenkens | obviously it's not a ghostscript problem as you pointed out. I'm just trying to figure out _why_ PDFCreator does it right, and bioPDF does not. | 17:00.25 |
chrisl | I suspect that bioPDF has either messed up the virtual printer install, or doesn't have a correctly configured PPD for the Ghostscript backend | 17:01.20 |
ray_laptop | Jenkens: you might want to look at the PPD (inf) file that the PSCRIPT5 is using for the two printers. That has the settings that tell PSCRIPT5 what kind of PS to generate (or at least used to -- I haven't looked with Win 7 recently) | 17:02.05 |
Jenkens | ray_laptop: I did a bit of comparing between PPD's (bio and pdfcreator) .. they are quite different | 17:02.30 |
| not sure which of those would impact what's happening though, i dont specifically see any font subsetting or the like | 17:02.51 |
ray_laptop | Jenkens: can you use the PDFCreator PPD with BioPDF and see if it fixes the problem ? | 17:03.02 |
chrisl | Jenkens: Well, they really shouldn't be very different, since they are both "printing" to Ghostscript. | 17:03.12 |
Jenkens | ray_laptop: that was my next step.. just not sure how to go about that. I guess just overwrite the PPD is enough? | 17:03.28 |
ray_laptop | Jenkens: I'd overwrite the PPD and reboot (in case it was already resident in RAM somewhere) | 17:04.12 |
chrisl | Jenkens: Oh, and you almost certainly are seeing font subsetting (unless you're producing PDF/A) | 17:04.15 |
Jenkens | ray_laptop: looks like overwriting was enough. I see a bit change in printer "settings" | 17:05.04 |
| let me see if printing works as expected | 17:05.10 |
ray_laptop | is only slightly hopeful | 17:05.42 |
Jenkens | hrmm | 17:06.15 |
| ok, if i change postscript output option: optimize for portability | 17:06.28 |
| it doesn't save | 17:06.33 |
| (this was a problem I thought was biopdf related too).. | 17:06.44 |
| locking down things that chris suggested like truetype font download option and substitute font dont seem to 'stick'. | 17:07.02 |
ray_laptop | sheesh. working on bug 692688 (finally) and it takes 189 seconds (180 of it rendering) to png16m, and takes 654 seconds (634 rendering) to tiffsep :-( | 17:07.07 |
chrisl | Jenkens: that not retaining your settings is one of the places I think bioPDF have messed up their install of the printer | 17:08.00 |
Jenkens | unfortunately the ppd alone does not solve it | 17:08.22 |
chrisl | Jenkens: one thought: can you make the changes as administrator? Maybe that would make them stick? | 17:08.36 |
ray_laptop | Jenkens: oh, well. | 17:08.43 |
Jenkens | I am administrator on this machine (hell I'm domain admin) | 17:08.50 |
chrisl | Jenkens: what version of Windows is it? | 17:09.22 |
Jenkens | I'm not sure if you read my ticket updates, but turning off "enable advanced printing features" on the advanced tab for the printer properties 'fixes' the weird TTXXXYZY font issue (XP SP3) | 17:09.35 |
henrys | Salil:not sure why that would happen - it would be nice if we had a library build in the xcode project and we'll put it on our list... other than that I'm afraid you're on your own. | 17:09.44 |
Jenkens | windows test page goes from 23KB to 25KB and the fonts identify right instead of TTXXYYZZ | 17:09.56 |
| re-enalbing "enable advanced printing features", size goes back down to 22KB and my fonts are TTXXYYZZ again. | 17:10.19 |
| enabling* | 17:10.24 |
| which of course makes no sense whatsoever. | 17:10.56 |
chrisl | Hmm, unnecessarily converting a Truetype font to a Type 1 seems like a "retarded printing feature", not an "advanced" one :-( | 17:11.08 |
Jenkens | indeed | 17:11.19 |
| not sure why simply toggling that option fixes the font issue | 17:11.43 |
chrisl | The difference in the file size is probably just down to the specifics of embedding Truetype outlines, rather than Type 1 outlines. | 17:12.00 |
Jenkens | I told bioPDF about and whoever replied gave it a head scratch | 17:12.03 |
| right | 17:12.09 |
henrys | Jenkens:probably I missed something why don't you return this biopdf product and buy something that works? | 17:12.16 |
Jenkens | I figured as much, I don't mind a little file swell | 17:12.20 |
| henrys: haha, well... I haven't explored that avenue yet. | 17:12.38 |
| henrys: really this problem only manifested itself last week. we have a couple reports that print is MS San Serif that we can't change.. and doing things like highlighting or striking through fonts identified as TTXXYYZZ gives inconsistent results (half highlighted) | 17:13.17 |
| for _most_ printing it seems to be doing a good enough job | 17:13.37 |
| though had I known they don't support GS I would have looked elsewhere (they were recommended by our paperless office developers) | 17:14.01 |
henrys | we are going to tire helping you with biopdf given our relationship with the company. | 17:14.02 |
Jenkens | I realize that | 17:14.26 |
| I'm just trying to provide info that might lead someone in the right direction as to what's causing this | 17:15.08 |
chrisl | Jenkens: regardless of other things, the bioPDF printer's refusal to keep your settings is just a bug, and the only people who might be able to fix it are bio themselves. | 17:15.38 |
Jenkens | chrisl: yes, I have a report out on that.. but as to why toggling advanced features 'fixes' this is another mystery. | 17:16.09 |
| now why pdfcreator is trying to get me to download something from mickyfastdl.com (which ESET) blocks... that's no bueno | 17:18.12 |
chrisl | pdfcreator used to be freeware, but now seems to be adware - and I don't think they're very selective about how the ads are displayed :-( | 17:19.49 |
| Jenkens: http://helpdesk.hs-lab.com/Disabling-Advanced-Printing-Features-a169.php | 17:21.14 |
Salil | henrys: As I compile for arm7 executables like echogs can't run so I am copying the echogs compiled for x86 architecture to finish the compilation for arm7. Do you think it would create problem? What does echogs and other executables do actually? | 17:21.18 |
ray_laptop | Jenkens: have you tried primopdf ? they are actually supported by us, and I've had good success with them. It is free (nitropdf uses it to up sell to their other products) | 17:21.57 |
Jenkens | ray_laptop: Not yet. Basically bioPDF was the first we looked at (due to recommendation.) I need something that'll autosave files to a location and let me supply variables for the filename <author>_<datetime>_<documenttitle>.pdf is preferred.... and then allow me to run an "after save" script. bioPDF allows this... and it looks like pdfCreator does too. | 17:23.20 |
| pdfcreator lets me save advanced preference (print for speed/portability).. but I think they are all registry entries so why bio's are jacked up.. who knows. | 17:24.37 |
| hrmm | 17:24.45 |
ray_laptop | Jenkens: OK. Free PrimoPDF does not allow this -- I know the purchased NitroPDF allows for the autosave to location (I have it), but I don't know about author, datetim, etc. | 17:25.08 |
Jenkens | ray_laptop: we have 25 licenses of NitroPDF as well | 17:25.22 |
chrisl | Jenkens: are you running bioPDF "locally", or as a shared printer? | 17:25.30 |
Jenkens | nitropdf has it's own set of printing issues (slow) | 17:25.31 |
| chrisl: locally | 17:25.36 |
| nitro also doesn't like some lines in our reports (omits them when we print them to the nitro printer, but they work fine coming through bioPDF).. | 17:25.57 |
ray_laptop | Jenkens: so why did you buy biopdf ? | 17:25.58 |
Jenkens | got a bug out with nitro about that | 17:26.05 |
| ray_laptop: because nitro wont let me automate printing to pdf and launch an "after print" program like bio can | 17:26.21 |
ray_laptop | Jenkens: I see. | 17:26.36 |
Jenkens | and as I said above there are other problems related to nitropdf's printer | 17:26.39 |
ray_laptop | Jenkens: right. | 17:26.48 |
Jenkens | all and all our experience with Nitro is a mixed bag | 17:26.48 |
chrisl | Jenkens: well, based on the link I posted above, a "PDF printer" should never have "Advanced printing features" enabled. | 17:27.01 |
Jenkens | we use stamps quite heavily and the first revision of 7.x crashed repeatedly using stamps | 17:27.03 |
ray_laptop | too bad none of these folks are open source -- then you could get help fixing things :-) | 17:27.27 |
Jenkens | got a fix that fixed stamp crash, but now quick access bar settings get overwritten with each update (that fixes other bugs we're encountering).. but support has been excellent about listening to our issues | 17:27.52 |
| chrisl: Sounds reasonable, pdfCreator has it enabled by default but handles the documents right (and the advanced settings save) | 17:28.29 |
chrisl | Jenkens: it *probably* depends on what the virtual printer is setup to do when that option is selected. Basically, there's no documentation I can find about it so, again, it's punt it back to bio..... | 17:30.00 |
Jenkens | yep | 17:30.11 |
| i got tickets out with bio about all of this. just not hopeful | 17:30.21 |
| when I told them that advanced settings dont "stick", they remarked about some .ini setting they were going to implement to fix that | 17:30.53 |
| i just groaned | 17:31.01 |
chrisl | Well, this is all their stuff, so if they can't fix it....... | 17:31.03 |
Jenkens | hah yeah | 17:31.08 |
| i just dont know of any alts other than pdfCreator right now and this has got that adware or whatever | 17:31.31 |
| and my uncheck "enable advanced settings" workaround is 'good enough' for the machines exhibiting the issue | 17:32.12 |
| for now | 17:32.15 |
chrisl | FWIW, I have seen issues vaguely similar to this with other PDF creator type products, which is how I knew where to point you in the first place - I've *never* seen one that refused to retain the user's settings, though | 17:32.52 |
Jenkens | yeah with bioPDF anything under preferences/advanced don't stick.. don't have a reason why and they don't seem to either | 17:34.24 |
| these seem to get stored in HKCU\Printers\DevModePerUser | 17:35.23 |
| but it's just a binary entry so it's not easy to decipher | 17:35.37 |
chrisl | Jenkens: well, I'm out of ideas, and I'm heading out now - good luck with bio....... | 17:40.13 |
ray_laptop | bug 692688 is really interesting. It is definitely NOT a tiffsep issue -- it is a clist issue. It is restarting the glyphs from the top down on every band. | 17:53.07 |
| If I remove the TextAlphaBits=4 then the problem disappears -- which explains why we probably haven't seen it (at least often) | 17:56.02 |
| should be easy to track down. For the logs: Thanks to Ken for the reduced file -- it runs a lot quicker and I can use -ZL without as much "noise" | 17:57.30 |
| kens: I didn't realize you were still online -- see thanks on previous line | 17:58.04 |
| the AA mode is using copy_color_alpha and the non-AA uses copy_mono_planes. Digging further. | 18:02.52 |
Salil | Is there any way around for GS to not use cups? | 18:07.46 |
kens | ray_ I'd completely forgotten I'd done that :-) | 18:20.14 |
kens | guesses Ray's laptop crashed again | 18:21.35 |
henrys | kens:unfortunately it looks like gl/2 -> pdfwrite is live and well again and the customer wants a pdfwrite solution for rops sigh... | 18:33.29 |
ray_laptop | henrys: unlike transparency and gradients, ROPs have no bbox info -- we pretty much would have to render the whole page to an image, but I bet they want searchable text, too, right? | 18:39.17 |
henrys | I need to study something - I think that if pdfwrite could intercept the rop device install in the image code we'd catch a lot of the issues. Images are 90% of the problem. kens and I were looking at that a few weeks ago while debugging one of the problems. | 18:41.43 |
mvrhel | ray_laptop: so I have the put_image working now. | 18:45.14 |
henrys | going swimming bbiaw | 18:45.20 |
mvrhel | it was not too hard since it just needs to call the copy_planes operation for the planar device | 18:45.36 |
| off to run some errands | 18:46.26 |
| bbiaw | 18:46.28 |
ray_laptop | ah ha. The bug 692688 is doing transparency -- I bet the bug is in pdf14_copy_alpha (good thing mvrhel just left) | 18:53.27 |
| mvrhel: I think I found it -- pdf14_copy_alpha wasn't adjusting the 'data' pointer when it bumped y up to the bottom of the rect (and decreased h). It was missing that part of fit_copy. Testing now | 19:06.45 |
vtorri | hey | 19:29.27 |
ghostbot | niihau | 19:29.27 |
vtorri | so mupdf is near 1.0 ? | 19:29.36 |
tor8 | vtorri: yes, we're preparing release candidate now | 19:46.29 |
vtorri | tor8: about thread stuff, on windows, are you using native thread functions ? of pthread-win32 ? | 19:55.19 |
| or* | 19:55.36 |
kens | henrys I see your comments, and I'll read the logs for details in the morning, maybe we can talk about this tomorrow or next week ? | 20:03.15 |
ray_laptop | fixed bug 692688 - running cluster test now, but it works on the bug's file. | 20:54.49 |
| mvrhel: do you want to see the patch (it's really tiny) | 20:56.53 |
| since it requires TextAlphaBits>1, transparency AND a band boundary in the glyph, I now see why this hasn't been a bigger problem. | 20:58.58 |
mvrhel | ray_laptop: no I don't need to see it. please go ahead and commit if it fixes the issue | 20:59.28 |
| thanks for finding it | 20:59.43 |
ray_laptop | mvrhel: OK -- as soon as I review the cluster results | 20:59.47 |
Jenkens | ray_laptop: it looks like replacing the ppd and bpd for bio with the one from pdfcreator 'works'. | 20:59.52 |
ray_laptop | Jenkens: Awesome ! | 21:00.07 |
Jenkens | just FYI.. not sure what the BPD is specifically, but when I tried with just the PPD this morning it did not work | 21:00.08 |
| also I can now save "advanced" printer settings and they 'stick' | 21:00.23 |
| not sure why the ppd and bpd would do that | 21:00.29 |
ray_laptop | Jenkens: I think the BPD is a binary form of the PPD that gets created when the printer installs | 21:00.35 |
Jenkens | right | 21:00.39 |
| thats kind of what i figured | 21:00.45 |
ray_laptop | Jenkens: thanks for letting us know -- chrisl_away will not have bad dreams :-) | 21:01.28 |
Jenkens | i've relayed this info to bioPDF.. so maybe they can figure it out *shrug* I know it's not your guy's issue but I wanted to relay it | 21:01.32 |
| FWIW I told the authors that they should support GhostScript (monetarily).. and they claimed it was too cost prohibitive to do so | 21:01.59 |
ray_laptop | Jenkens: Thanks. I had started typing that, but you got it out first. Maybe if we start bashing them on facebook, they'll smell the roses :-) | 21:05.18 |
Jenkens | lol | 21:06.22 |
ray_laptop | Jenkens: We'll bring it up with our Sales VP and Pres. -- they're the "enforcers" | 21:06.29 |
Jenkens | He claims he's conversed with someone at some point and you are all very helpful.... so it's not like he said anything negative other than a cost issue. | 21:07.09 |
| hopefully this can be a win for me (with a fix) and a win for you (with a sale) if it ever goes that way. :) | 21:07.34 |
Salil | henrys: Are there any other files which needs to be in searchable path other than files in lib and Resources. Also does it look for any directories or can all those files be in single directory and GS will find them? | 21:07.50 |
ray_laptop | Jenkens: I won't be holding my breath ;-) | 21:07.51 |
Jenkens | yeah | 21:08.01 |
ray_laptop | Salil: gs also needs the Font file to be in the searchable path set with -I particularly if you use -dSAFER since only those paths get put on the "PermitFileReading" list | 21:09.14 |
| Salil: the paths on the LIBPATH list (search paths when you do -h) are the ones we search for the Fontmap files -- The Fonts themselves are usually OK, and if you have COMPILE_INITS=1 (the default) then the %romesource and descendants are added automagically | 21:11.59 |
Salil | ray_laptop: If I set -Idirectory_name and put all the files (no folders) in directory_name directory will it find it? The reason why single directory is that by design XCode puts them all in single directory on the device. | 21:12.21 |
henrys | Salil:without debugging you're in trouble. There really is no point in doing anything else until that works IMHO, "sharpen the saw" kind of thing. | 21:12.25 |
Salil | henrys: I created dylib for latest version but can't run with it on the device. On static library -dDEBUG doesn't show any love no matter what :( | 21:14.10 |
ray_laptop | Salil: that _should_ work. But note that PostScript (and thus Ghostscript) expects that the files under Resource/ are in subdirectories. But if you have COMPILE_INITS=1 those will be built into the executable (rom file systems a.k.a. %romesource/ ) | 21:15.09 |
henrys | there are 2 -dDEBUG one for ghostscript and one when you compile the code (CFLAGS) to which are you referring. | 21:15.12 |
| ? | 21:15.19 |
ray_laptop | Salil: -dDEBUG as a command line (or gs_init_with_args) option works on debug or release builds | 21:16.12 |
Salil | ray_laptop:I am not creating rom as no executable is used. I am creating static library to use in iOS project | 21:17.08 |
| ray_laptop: and -dDEBUG doesn't show up anything in debug console | 21:17.28 |
ray_laptop | a common oversight is to set gs_args to an option string, but forget to bump the argc to gs_init_with_args | 21:17.28 |
| Salil: if you specify COMPILE_INITS=1 then gsromfs1.o gets linked into the libraray | 21:18.08 |
| Salil: note that -dDEBUG should come _before_ the filename you are trying to process (if it is the last arg you pass to gsap_init_with_args) | 21:19.19 |
Salil | ray_laptop: oops let me change the order | 21:19.38 |
henrys | he posted his code | 21:19.40 |
ray_laptop | henrys: sorry -- I didn't see it. Is it in today's log ? | 21:20.01 |
Salil | http://pastebin.com/tsxEjuei | 21:20.06 |
henrys | I was hoping he would do a debug build so we could look at -Z flags. | 21:20.18 |
ray_laptop | I found it | 21:20.22 |
Jenkens | hrm, looks like it's not the ppd/bpd afterall | 21:20.51 |
| i'm now using the biopdf ppd/bpd, i switched the driver (advanced tab) to the pdfCreate driver, printed (it printed fine but used biopdf's dialogs).. then reverted it back to the bioPDF driver and now things 'stick' and print fine...... | 21:21.31 |
Salil | henrys: I did debug build with CFLAGS as well but it didn't change anything. But that could be my XCode as I found that it has been caching libraries and not using the latest I had been building. I just found that today so I have to recheck on that one. | 21:21.59 |
ray_laptop | Salil: just as a matter of convenience, usually setting gsargc to 0 then using gsargv[gsargc++] = "..." prevents some mistakes | 21:22.21 |
| Salil: so, are ALL of the Resource/Init files on the "Path_Where_gs_init.ps_can_be_found" ? | 21:23.38 |
| Salil: and what does replacing "-dNOPAUSE" with "-h" produce ? | 21:24.23 |
| Salil: lastly, For testing, try removing the -dSAFER to see if that is causing you problems | 21:25.41 |
Salil | ray_laptop: good point about gsargc. Thanks! yes all the files are at "Path_Where_gs_init.ps_can_be_found" | 21:26.06 |
| ray_laptop:let me see | 21:26.17 |
ray_laptop | Salil: because gs also needs a valid Resource/ directory for things like Encoding Decoding Fonts and a sibling 'iccprofiles' (unless you specify those with the correct option) | 21:27.19 |
Salil | ray_laptop: so I will have to put them in a proper directory structure. Right? | 21:28.26 |
ray_laptop | Salil: once again -- this is why COMPILE_INITS=1 is _strongly_ recommended -- it puts all of the required files into the lib/executable so they will be there | 21:28.39 |
Salil | henrys: Is -Z flag part of gs commandline? | 21:28.45 |
Jenkens | shakes his head at the whole thing | 21:29.03 |
Salil | COMPILE_INITS=1 is not default? | 21:29.08 |
ray_laptop | the romfs is a rom/ram based pseudo filesystem that has (effectively) the needed subdirectories. | 21:29.36 |
| Salil: It is, but you need to check (grep) your Makefile | 21:30.00 |
| Salil: the output from -h will tell you if %romesource/Init/ is on the search path | 21:30.45 |
Salil | ray_laptop: It seems to be in place. -h outputs the help details with this: Initialization files are compiled into the executable. | 21:30.49 |
ray_laptop | %rom%Resource/Init | 21:30.58 |
| Salil: great -- so I recommend removing your -I option | 21:31.26 |
Salil | ray_laptop: after I removed -dSAFER it crashed with -110 error | 21:32.01 |
ray_laptop | Salil: because we assume that the -I path is intended to OVERRIDE the %rom contents (for testing, or wahtever) | 21:32.10 |
| Salil: e_Info is return code 110 (codes > 100 are not crashes, but used as "request" codes) | 21:33.33 |
| Salil: e_Info is returned if you specified -h as an option | 21:34.42 |
| Salil: see psi/iapi.h | 21:35.11 |
| Salil: I have to run an errand -- I'll check back later, but also try changing your "-IPath_Where_gs_init.ps_can_be_found" to "-I." to make sure the cwd is permitted (f you still have -dSAFER) | 21:37.35 |
| bbiaw... | 21:37.47 |
Salil | ray_laptop: sorry for using crash. I hardly think cwd will be permitted and what will be achieved with -I.? I will catch you later. Thanks for all the help. | 21:44.46 |
| henrys: So I should compile with debug and use -Z flag to see debug information? | 21:45.10 |
| henrys: I compiled with CFLAGS-DDEBUG and put -Z option but it doesn't give much info either. | 22:02.31 |
| henrys: I mean it doesn't give any additional info | 22:02.51 |
| ray_laptop: If everything is compiled in gsromfs1.o (rom) then I don't need to add any resources (gs_*.ps and fonts) to the project...right? | 22:06.12 |
ray_laptop | Salil: correct. BTW -- what was the output from using "-h" -- if you didn't see anything, then your stdout and stderr aren't hooked up correctly. You will need both to debug. | 22:13.03 |
Salil | ray_laptop: I saw the usual -h output like we see on commandline | 22:13.39 |
| ray_laptop: I compiled with CFLAGS -DDEBUG and put -Z option but it doesn't give much info either. | 22:14.04 |
ray_laptop | Salil: OK. can you paste it (to pastebin) so we can review it ? | 22:14.21 |
| Salil: -Z doesn't do anything by itself. PLEASE READ THE DOCS -- particularly doc/Use.htm (see Debugging switches) | 22:15.11 |
| Read The Fantastic Manual ;-) | 22:15.58 |
Salil | ray_laptop: :) I will now. Pasting in a minute. | 22:16.21 |
| http://pastebin.com/mY8rbggm | 22:18.23 |
| ray_laptop: Also please note that I am compiling without cups as there is no cups shared library for iOS | 22:19.13 |
| henrys: http://pastebin.com/r9WfTHv8 output with -Zi option. It seems to be doing something. The input pdf file has very little text. Any suggestions for other Z flags? | 22:32.06 |
henrys | Salil:hey you got it to work - now do -Z# and post that. | 22:38.55 |
Salil | henrys:ok in a minute. | 22:40.14 |
| henrys: http://pastebin.com/Hhvy1Frx | 22:42.00 |
| henrys: I just noticed I am using headers from 8.63. I can find iapi.h and ierrors.h in psi but where to find gdevdsp.h? | 22:43.05 |
| henrys: since framework gave the header in 8.63 I had been using that | 22:43.22 |
| henrys: How can I use this debug output to track the problem? | 22:54.31 |
henrys | Salil:sorry got to head out, what release are you trying to use? | 23:31.13 |
Salil | henrys: latest 905 | 23:31.31 |
henrys | Okay the current code must have changed your line numbers for the -Z output are different than mine. I'll have a look at it later if you don't figure it out. | 23:32.46 |
Salil | henrys: I have to head out too. I hope to catch you later. It seems to give error in zusparam in setparam function e_rangecheck. Thanks for all the help. | 23:34.07 |
| Forward 1 day (to 2012/03/29)>>> | |