| <<<Back 1 day (to 2015/10/06) | 20151007 |
emendelson | A question about adding fonts to the 9.18 version of pcl6.exe (for my own private use, not distribution). For various reasons, I have been adding Prestige Elite fonts to recent versions of pcl6 this way: I add the TrueType fonts to the urwfonts folder, then ... | 01:25.30 |
| .. I add the font information to plftable.h, making sure that I number the added fonts after the existing fonts. This lets me build a pcl6 that will create a PDF with Prestige fonts if send PCL data with the correct font info .. | 01:27.43 |
| ... in the PCL font string. I can't make this work in 9.18. I see that the LinePrinter fonts are now part of the numbering scheme, and I've added my Prestige Fonts with numbers after the LinePrinter fonts but ... | 01:28.36 |
| ... I can't get Prestige in the PCL output (LinePrinter gets substituted). Is there some other place in the code where I have to change something to get back this ability? Again, this is entirely for my own use ... | 01:29.23 |
| not for any commercial purposes. I've been printing with Prestige Elite since the days when you could buy a Prestige Elite cartridge for your LaserJet III; my printer drivers in DOS applications are set up to print to ... | 01:30.18 |
| ... Prestige Elite, and it's been very good to have this same ability in pcl6 so that I can create PDFs with Prestige Elite. Any advice will be gratefully received, though I know that this is NOT something that anyone supports or is ever going to support. | 01:31.21 |
fredross-perry | robin: I added a commit to fix the whitespace issues. Sorry itâs not added to the previous commit. Time to push pins into my git doll now⦠| 03:58.14 |
| robin: I *believe* Iâve uccessfully squashed the whitespace fixes into the previous commit. | 04:48.20 |
kens | chrisl which version of C are we currently mandating ? C89 ? | 07:47.43 |
chrisl | kens: yes, C89 | 07:50.38 |
kens | OK so no isnan or isfinite functions then. | 07:50.56 |
| SO I'm going to ignore that part of the bug report | 07:51.09 |
| Or more accurately decline to address | 07:51.21 |
chrisl | Erm, well, we do use widely supported post-C89 extensions in various places - would those qualify? | 07:51.52 |
kens | I don't thnk there is a portable version. For example VC uses _isnan() and I thnk gcc uses isnan() | 07:52.16 |
| There is a quick check (x !=x) which allegedly works but isn't 100% reliable across all compilers | 07:52.51 |
chrisl | We have isnan() in our sprintf implementation, and that seems okay | 07:53.17 |
kens | Hmm, didn't know that. Doesn't work for infinity though, we get 0 (to 80 places) that way | 07:54.06 |
chrisl | Oh, but it is protected by a preprocessor conditional, so..... | 07:54.07 |
kens | ah | 07:54.12 |
| Oh and allegedly one of the compilers which messes up x!=x is visual C++ | 07:55.37 |
chrisl | Probably best to steer clear, unless you use the "#ifdef HAVE_ISNAN" etc | 07:56.17 |
kens | I oculd ad tht to the mul and div math, but I question the utility of doing so, especially since we never have before | 07:56.54 |
chrisl | Which bug is this? | 07:57.15 |
kens | 696248 brand new ths morning | 07:57.28 |
| THe crash I can reproduce and fix | 07:57.35 |
| I see we don't use isnan with %g :-( | 07:59.15 |
chrisl | Well, I'm not sure we check for isnan in the configure script, so I don't think we'll ever use it! | 08:00.13 |
kens | I was about to check on WIndows | 08:00.27 |
chrisl | And I didn't get a mail for that bug - that can't be good :-( | 08:00.46 |
kens | check spam | 08:00.53 |
chrisl | I did | 08:00.58 |
kens | Tha'ts really odd | 08:01.03 |
chrisl | The last gs-bugs mails I got were the FirstPage/LastPage report | 08:02.22 |
kens | should also have been marcos verifying the fix for bug 696239 | 08:02.59 |
chrisl | Nope - I wonder if I've been kicked off it | 08:03.18 |
kens | Doesn't sound good | 08:03.26 |
| FWIW we don't use isnan or isinf with VC++ | 08:03.42 |
| So I'm going to fix the crash and decline the other points. They are a nonsense anyway | 08:03.59 |
chrisl | Presumably VC++ has both those in some form? | 08:04.21 |
kens | 2008 does, I have no idea about 2005 | 08:04.35 |
| in vs 2008 its _isnan and _isfinite I think | 08:04.53 |
chrisl | We could add "#define isnan(a) _isnan(a)" for Windows, and add the checks in configure for other platforms | 08:05.42 |
kens | Well it might be 'nce', though I think I'd rate it as an enhancement (unlike the crash) | 08:06.19 |
| And I certaionly don't want to take on configure :-) | 08:06.43 |
chrisl | It's up to you, I'm happy to do the builds part of the work | 08:07.16 |
kens | OK what I'll do is fix the crash, change it to an enhancement and add a note about isnan | 08:07.37 |
tor8 | kens: we use _isnan with VS2005 in mupdf | 08:15.02 |
kens | tor8 that's great, thanks tor | 08:15.21 |
tor8 | and _finite | 08:15.35 |
kens | I'll feel comfortable with using it then, I was having trouble finding informaton on it from teh MS help | 08:15.41 |
tor8 | we guard the defines with an #ifdef MSC_VER < 1800 | 08:16.12 |
kens | Why < 1800 ? | 08:16.27 |
tor8 | newer versions support isnan directly :) | 08:16.48 |
kens | Oh, well I guess I need to use that too in the #define, use either _isnan or isnan depending | 08:17.10 |
tor8 | http://git.ghostscript.com/?p=mujs.git;a=blob;f=jsi.h;h=40821e9bc35aeed01f360986059e218ba8571d49;hb=HEAD#l15 | 08:17.37 |
| for mujs, we use a lot of floating point functions that we needed to cover for in MSVC | 08:17.53 |
| no round, isnan, isinf, isfinite, signbit, INFINITY or NAN | 08:18.28 |
kens | OK well I thnk I can write somethgin which will work, but not before I head out. I'll look at it when I get back | 08:18.31 |
Robin_Watts | Bugger. Should have taken the day off. | 12:58.01 |
| http://www.telegraph.co.uk/news/worldnews/northamerica/usa/11915644/Doomsday-coming-End-of-the-world-is-imminent-say-Christian-believers.html | 12:58.07 |
chrisl | What? Again??? | 12:58.28 |
kens | Can't happen my birthday is after that | 12:58.35 |
Robin_Watts | 10 hours to go, apparently. | 13:00.05 |
| Will the cluster have cleared the backlog by then ? :) | 13:00.22 |
kens | Ah you thnk its the nine billion names of god ? | 13:00.39 |
| I wonder if he'll announce a retraction tomorrow | 13:02.03 |
Robin_Watts | I feel bad taking the piss, cos it's clearly a deeply held belief. I mean, there is no upside to it. | 13:03.13 |
| If he's right, it's not like he gets to say "I told you so", or that people will go "Fuck me, he knows his stuff, I'm going to follow him..." | 13:03.46 |
kens | The Branch Davidians felt that way too, for them there was truly no upside, for thje rest of use though | 13:04.00 |
| Its not the fact that its a belief, its the steady stream of 'I'm better than you and I'm going to Heaven and you all aren't cause you're heathens and sinners and today is the day you all get your comeuppance' I find annoying | 13:05.30 |
Robin_Watts | kens: Given the weather, I suspect it'll be rained off. | 13:59.59 |
kens | maybe it'll be flood again not fire | 14:00.19 |
Robin_Watts | That's so 4000BC. | 14:08.05 |
kens | I htought 4004 BC was Creation | 14:09.27 |
chrisl | Is the support mailing list hosted on casper? Or is it hosted on gmail somehow - does anyone know? It *looks* like gmail, from the message header..... | 14:47.38 |
kens | suppot@artifex.com, or the gs-bugs list ? | 14:48.08 |
| Not that I know either way..... | 14:48.15 |
Robin_Watts | Nothing @artifex.com is handled on casper, AIUI. | 14:48.53 |
chrisl | Well, the mails I've been getting from the bugzilla-daemon have been addressed to support | 14:49.09 |
kens | Really ? The ones I get are mailed to me | 14:49.43 |
chrisl | For bugs assigned to me, I get the the mails addressed to me, but all the others, get to me through a list | 14:50.26 |
kens | Oh possibly, I didn't check those. So many of the gs-bugs mails I get are assigned to me...... | 14:50.54 |
chrisl | Which up to 1st Oct was gs-bugs, but now seems to be support..... | 14:51.09 |
kens | Yes I see that change also | 14:51.40 |
chrisl | And now, I'm not getting them | 14:52.05 |
| One for Marcos, I guess | 14:52.46 |
kens | All the mails I've had since then are assigned to me. Of course, I could be missing mails, its difficutl to tell | 14:52.49 |
chrisl | kens: I made a comment on the gs installer bug from a few days ago. I haven't had any mail from bugzilla either directly or mailing list - can you check whether you got an update? | 14:59.43 |
kens | Err let me see | 14:59.54 |
chrisl | You'd have got a mailing list copy | 15:00.13 |
antkong | hi, i use imagemagick to convert pdf file to image files. The latest version somehow stuck when trying to convert a pdf file. `gs` run for 60 min already. It created some 12 tmp files and stop creating more files. What can I do? Here is the related command line âconvert -verbose -strip -quality 100 -alpha off -density 165% -scene 1 infile outfileâ | 15:01.42 |
kens | I don't have a copy of your comment on the 7th October. So it looks like we are *all* missig emails for bugs not assigned to us | 15:02.04 |
henrys | kens: I copied you in on a bug too, and I don't think you saw it, the one shelly is working on. | 15:02.50 |
kens | I'm sure I didn't see that henrys, no | 15:03.02 |
Robin_Watts | antkong: This is an irc channel for gs, not for imagemagick :) | 15:03.04 |
| If you can find us a gs command that goes wrong, we'll try to help. | 15:03.16 |
kens | henrys if you give me the bug number I'll look at it | 15:03.21 |
henrys | http://bugs.ghostscript.com/show_bug.cgi?id=696222 | 15:03.50 |
Robin_Watts | There is probably an imagemagick command line option to get it to tell you the gs command it's calling. | 15:04.00 |
| something to do with debug or verbose I would imagine. | 15:04.12 |
antkong | Robin_watts: certainly. Here you go: "gs" -q -dQUIET -dSAFER -dBATCH -dNOPAUSE -dNOPROMPT -dMaxBitmap=500000000 -dAlignToPixels=0 -dGridFitTT=2 "-sDEVICE=pamcmyk32" -dTextAlphaBits=4 -dGraphicsAlphaBits=4 "-r165x165" -dUseCIEColor "-sOutputFile=/tmp/magick-21771CV2oAhY9g6Fi%d" "-f/tmp/magick-217719N59QFMIg5Vq" "-f/tmp/magick-21771z0o2ZwmIq9bz" | 15:04.20 |
kens | henrys I don;'t seem to be on the cc list for that bug | 15:04.29 |
Robin_Watts | antkong: And if you run that command locally, do you get the same problem? | 15:04.54 |
henrys | oh I mad the comment and never added you ;-) | 15:05.08 |
| s/mad/made | 15:05.14 |
kens | :-) | 15:05.15 |
| I did just find the comment | 15:05.21 |
Robin_Watts | antkong: You'll need to capture the content of those /tmp files, obviously. (The -f listed ones) | 15:05.51 |
antkong | Robin_watts: what do you mean by âcaptureâ? | 15:07.55 |
kens | We will need to run those files ourselves, so you need to have the files. | 15:08.11 |
| I presume you do or you could not have run the same command locally | 15:08.29 |
| Assumign you do have the files, then open a bug report, attach the input files, and give teh command line | 15:08.53 |
Robin_Watts | antkong: As kens says, in order to see if you can get the same problem running just gs (and leaving imagemagick out), you'll need to have copies of those files to pass to gs. | 15:08.59 |
chrisl | marcosw: ping | 15:17.13 |
marcosw | chrisl: morning | 15:17.19 |
chrisl | morning. I seem to have dropped off the bugzilla mails in the last day or so | 15:17.42 |
kens | chrisl its clearly not just you | 15:18.02 |
| We've shown I'm missing at least one email also. Its terribly hard to know how many others might ahve gone AWOL | 15:18.31 |
chrisl | Well, I'm not getting *any* - not even for my bugs. | 15:18.39 |
kens | Oh. well that's even worse than me | 15:18.48 |
marcosw | great. email is the bane of my existence. | 15:18.50 |
kens | I wonder if its related ot the change you did for smart office bugs | 15:20.10 |
marcosw | everyone just needs to write a program that scrapes bugzilla every hour or so and alerts them to changes. You could even customize it so more important changes wake you up or something. | 15:20.16 |
Robin_Watts | We should all move to using a totally reliable and industry standard host (that doesn't snoop on our private data) for all our email requirements! | 15:20.20 |
| Oh, wait... | 15:20.22 |
marcosw | we used to run bugzilla on some shared system. Thatâs why all our bug numbers are in the 690 000 range. | 15:21.16 |
chrisl | Was the change to having bugzilla use support rather than gs-bugs for SO? | 15:21.40 |
marcosw | obviously that wouldnât work for customer bugs. | 15:21.41 |
| chrisl: yes, but that âshouldnâtâ have broken anything (other than possibly SO bugs). | 15:22.07 |
kens | Well I no longer get emails for any bugs other than the ones assigned to me | 15:22.31 |
chrisl | marcosw: no, it was working fine up to yesterday - but that was the only other change I could see recently | 15:22.48 |
kens | As chrisl point sout above, the mails used to come from gs-bugs, now they come from support | 15:22.51 |
marcosw | chrisl: oh good. I donât mind when things break. I hate when I break them. | 15:23.40 |
chrisl | marcosw: careful what you say! | 15:24.19 |
marcosw | hopefully the emails are stacked up somewhere and will all magically show up at once :-) sort of like Torâs mupdf commits. | 15:24.32 |
chrisl | Yes, now he's dropped off, let's blame Tor! | 15:24.55 |
marcosw | speaking of which, wonder if the cluster will catch up today. | 15:24.55 |
kens | getting close now | 15:25.05 |
chrisl | I'm doing my best to slow it down | 15:25.13 |
kens | depends how much chris and I do | 15:25.16 |
chrisl | Although, OTOH, if I never see any bug notifications, I can just assume everything is okay! | 15:27.32 |
marcosw | is annoyed that when starbucks switched from ATT to Google for their free internet that they blocked port 22. I run an alternate port ssh daemon at home but it means when connecting to casper or any other system I have to route via my home. | 15:28.41 |
chrisl | Well, that's just dumb - I wonder why they'd do that | 15:29.54 |
marcosw | I donât know. itâs doesnât stop anyone, itâs just annoying. | 15:30.23 |
| iâve been at places where they block everytihng but port 80 :-) | 15:30.37 |
| (and maybe 443). | 15:30.58 |
Robin_Watts | Often 8080 gets through. | 15:31.11 |
marcosw | i run sshd on 8888 (Iâm already using 8080 :-) ) | 15:31.29 |
| odd, but this starbucks isnât blocking 22. i tried connecting to bugs.ghostscript.com via port 22 and it worked. So never mind. | 15:32.15 |
chrisl | Maybe they've had so many complaints | 15:32.34 |
marcosw | they probably block 59xx | 15:32.35 |
Robin_Watts | Just use telnet. | 15:33.02 |
marcosw | theyâve had a lot of complaints, just google âstarbucks blocking port 22â. | 15:33.30 |
chrisl | Well, loads of VPN traffic is routed over ssh so a lot of business users would be stuck | 15:34.34 |
marcosw | speaking of starbucks, some of them have removed their power outlets. apparently people were using them as offices for the entire day. | 15:34.55 |
chrisl | Just sipping on that one espresso...... | 15:35.21 |
Robin_Watts | Yeah, the profit margins on their coffees only really made sense when you think of them as including office space hire... | 15:37.03 |
marcosw | i will admit that iâve gone into a starbucks and didnât want to stand in line to order a drink so just sat down and used the internets. They presumably would say something if the tables were all taken but the one that Iâve done this at is near my sonâs school, so lots of students who donât stay. | 15:37.24 |
chrisl | TBH, it's not an environment in which I'd want to attempt "real" work | 15:37.57 |
marcosw | you mean like debugging an email problem? | 15:38.26 |
Robin_Watts | chrisl: I've yet to find one of those anywhere :) | 15:38.29 |
chrisl | Points well taken! | 15:38.52 |
marcosw | speaking of which, bugs.ghostscript.com is showing lots of email being deliverd without problemâ¦wonder where itâs going. | 15:39.10 |
| BTW, working at home with you children makes working at starbucks seem like a library (back when librarys were quiet places). When our kids were little I would often go to starbucks to work. Admittedly I went to a less popluar starbucks, which is now closed due to lack customers. | 15:41.04 |
marcosw | wonders if anyone can tell that chatting on irc is better than debugging email⦠| 15:41.22 |
Robin_Watts | https://coffitivity.com/ | 15:42.10 |
marcosw | uhmm, it looks like bugzila emails broken October 1, which is when I last made a change :-( | 15:44.24 |
fredross-perry | robin: nice | 15:52.42 |
| âOur coffee shop sounds are supported by scienceâ | 15:53.07 |
chrisl | They clearly missed the quotes around the word science | 15:55.00 |
Robin_Watts | gets Portal flashbacks. | 15:56.10 |
kens | chrisl I added code to check for NAN and INF into zarith.c and it compiles and works on the cluster. There are several files which now throw errors (unsurprisingly perhaps) and I'll need to check those before proceeding. If I am happy with it I'll commit it and let you do the configure changes seperately if that's OK Wiothout those the code will do nothing. I used the same #define as in gssprintf.c | 15:58.22 |
chrisl | kens: that's fine, just ping when me when you're ready | 15:59.21 |
kens | It won;t be today, I need to go look at some Genoa test files nad figure out if they are *supposed* to throw an error, and if not, why they are..... | 15:59.46 |
marcosw | sorted, i think. | 16:01.27 |
| tots my fault. i changed the sender for bugzilla email from bugs-daemon@ghostcript.com to bugs-daemon@artifex.com. I did this since bugs-daemon@ghostcript.com doesnât go anywhere (it least I donât think it does, in any case I canât see those emails). So bounce message and mail not delivered messages were lost. | 16:04.00 |
| but I forgot to add bugs-daemon@artifex.com to the gs-bugs mailing list whitelist. | 16:04.36 |
kens | ah that would be it | 16:04.51 |
marcosw | unfortunately the gs-bugs mailing list is set to discard âpostings from non-membersâ, so it did. | 16:05.24 |
chrisl | Strange that it seemed okay up until yesterday | 16:05.24 |
| Maybe I just wasn't paying attention | 16:05.39 |
kens | Absences are hard to spot..... | 16:05.58 |
marcosw | chrisl: you sure? the last email I have from bugzilla that isnât for a bug that I submitted or am assigned is for Bug 696238 and that was Oct 1 at 8:47 AM (PDT), which is when I was fiddling with the settings. | 16:07.42 |
kens | THat's what I see too | 16:08.02 |
chrisl | I see the ones for 696246 - possibly because it was assigned to support? | 16:08.47 |
kens | Well I just got mail on bug 696222 so it loks like its all good again | 16:09.25 |
marcosw | chris: yes. I see those too but they are sent to support@artifex.com (and the cc that is sent to gs-bugs isnât there). | 16:10.10 |
chrisl | marcosw: right, I guessed that was the cause. Anyway, you seem to have worked your usual magic, and we're all go again. Thanks | 16:11.03 |
marcosw | as everyone has probably surmised from my use of the word âdiscardâ the emails are not recoverable unless i do something complicated with the MySQL bug database but Iâm not sure itâs worth the effort. There are 44 bugs that changed in the last 6 days, so everyone could just browse the changes: <http://tinyurl.com/pgoa7ym> | 16:12.55 |
| (ignore the report title, itâs searching for the last 6 days). | 16:13.37 |
| it does mean that the gs-bugs mailing list archive will be missing data. but I belive the gs-bugs archive is offline and has been for a long time so presumably no one uses it (why itâs offline is on my list of things to look into, but in the P5-enhancment-later category, so will never get done). | 16:16.30 |
kens | K I'm off, I'll look at these Genoa files tomorrow. Night all | 16:17.30 |
Robin_Watts | Night kens. | 16:17.42 |
chrisl | IIRC, the gs-bugs archive hasn't work since I joined - nobody seems to care. Maybe we should restrict gs-bugs to just Artifex people, now? | 16:18.28 |
marcosw | chrisl: I didnât even realize that not artifex people were able to subscribe. | 16:19.28 |
| we have 30 members :-) | 16:19.46 |
chrisl | Yeh, and occasionally we get external mails on the list not through bugzilla. | 16:20.39 |
marcosw | unless they are from a member of the list or whitelisted they should be discarded, as was demonstrated over the last 6 days :-) | 16:21.37 |
chrisl | Yes, they are from list members. But if it's fit to print, it should go in the bug, and if not, then sending it to public list is a bit dumb! | 16:22.38 |
marcosw | gs-bugs archive does exist and is up to date. The last entry was 14 minutes ago. | 16:28.27 |
chrisl | Is it accessible? I know the files are there, but I thought all the web links were dead ends | 16:29.03 |
marcosw | chrisl: I canât figure out how to get to it from the web, so no. | 16:29.24 |
chrisl | In theory, you get to them from here: http://ghostscript.com/cgi-bin/mailman/listinfo/gs-bugs | 16:29.54 |
| bug the link is stale | 16:29.59 |
marcosw | geez, we have a lot of mailing lists that are not longer used: fitz-cvs, gs-code-review, mac-gs, and a dozen or so more. | 16:30.37 |
| if you go <http://ghostscript.com/pipermail/> you get a friendly message | 16:30.56 |
chrisl | Yeh, I think I put that there | 16:31.17 |
marcosw | are any of our mailing lists accessible?!? | 16:31.36 |
chrisl | None of the archives are, no | 16:31.51 |
marcosw | the artifex-support ones are (authentication required): <http://ghostscript.com/cgi-bin/mailman/private/artifex-support/> | 16:34.03 |
chrisl | Interestingly: http://gs-bugs.ghostscript.narkive.com/ | 16:35.16 |
marcosw | I actually go through the pending emails for that list once a week or so and moderate them. | 16:35.27 |
chrisl | http://gs-devel.ghostscript.narkive.com/ | 16:35.43 |
marcosw | wow. but neither appears up to date. | 16:36.10 |
chrisl | gs-devel is close - there may be some delay | 16:36.45 |
marcosw | gs-bugs is 12 days old, which is well before I screwed up the email | 16:37.43 |
chrisl | I wasn't suggesting we rely on narkive, I was just surprised | 16:38.20 |
marcosw | much nicer presentation than our archives (even if they were working). | 16:38.24 |
| chrisl: I know. | 16:38.40 |
| <http://ghostscript.com/cgi-bin/mailman/private/gs-bugs> | 16:42.07 |
chrisl | I have no idea what my password would be..... | 16:43.49 |
marcosw | i donât know my password either, but I know the admin passord⦠| 16:46.53 |
henrys | chrisl (or ray for the logs) or anybody. Are there any device parameters that require the device to be closed. I know that if geometry changes can cause the framebuffer and other stuff to be reallocated, but they basically work. Does any strictly fail? | 17:54.34 |
| s/Does any/Does any device parameter/ | 17:55.12 |
rayjj | henrys: looking at gsdparam.c, gs_closedevice is called when HWResolution changes, or LeadingEdge, or width/height or MediaSize. Other devices may also add this in some cases. For example the 'bit' device does when num_components changes (the ForceMono param). grep for 'closedevice' in base and devices to see more examples | 18:58.18 |
| henrys: what do you mean by "do any strictly fail" ? | 18:58.51 |
henrys | rayjj: I'm trying to see a need for the languages to close the device before setting a particular parameter | 19:35.02 |
| rayjj: I know the library will close and reopen a device for some parameters | 19:35.35 |
rayjj | henrys: I don't know why that might be needed. Generally devices are not allowed to do anything before processing all of the parameters (for potential errors), so that once the parameters are "applied" (usually just putting the collected set into the device struct) the device will be guaranteed to be openable (open won't fail) if the device is closed due to some need | 19:37.59 |
| henrys: so AIUI, a device *can* close itself anytime it wants, and return after the close, and the device will be re-opened, but the open better work | 19:38.49 |
| henrys: what are you thinking about that might need a language to close a device *before* setting a param. In general, languages are supposed to let the lib + device take care of things, and making sure that setting params never end up returning to the language with an invalid/closed device | 19:40.42 |
henrys | rayjj: the way pcl/pxl are set up now they share a device but close the device after each job and set some parameters. I don't see why we really need to close. | 19:40.55 |
rayjj | henrys: they shouldn't have to as far as I can tell. Certainly PS doesn't work that way | 19:41.35 |
| henrys: it sort of sounds like "belt and suspenders" -- PCL didn't trust that the graphics_lib+device layer to close if it needed to, so it did it, "just in case" | 19:43.18 |
henrys | rayjj: that's what I suspected thanks | 19:43.56 |
rayjj | henrys: unless the PCL/PXL was not using 'put_params', but was directly setting values in the device structure (which is generally *NOT* the way they are supposed to), in which case closing the device might be needed | 19:45.05 |
henrys | rayjj: I think we only read. I'd call that just a bug that I'd want fixed anyway | 19:46.25 |
rayjj | henrys: agreed | 19:46.40 |
| henrys: and reading the device structure is OK, as long as they are elements of the gx_device class. It takes *very* careful code to access beyond that (generally only if there is an _is_***_device method that someone came up with | 19:48.47 |
| s/device method/method or macro/ | 19:49.18 |
henrys | rayjj: --saved-pages is not used in pcl_process_options is wrong. The device doesn't exist when you open it. | 20:36.57 |
| rayjj: --saved-pages in pcl_process_options is wrong. The device doesn't exist when you open it. | 20:37.15 |
| Forward 1 day (to 2015/10/08)>>> | |