| | <<<Back 1 day (to 2022/02/24) | Fwd 1 day (to 2022/02/26) >>> | 20220225 |
| artifexirc-bot | <RayJohnston> @Pritam I looked at the file. It has extra <cr><lf> characters that make it invalid. If I manually remove those with an editor, then the file runs. | 01:58.58 |
| | <Pritam> thank you for explaining ... it means you are able to open this file using GhostScript ... please upload the file which is working so that we can also try | 02:13.57 |
| | <enyc> I wonder how the world would be different if unix had not set up 0A(cr)-only line-endings and Apple had not 0D(lf)-only lend-ends, and CRLF (0d 0a) were the only type of text file =) | 02:52.43 |
| | <KenSharp> PostScript is a programming language. When you send a PostScript 'file' to Ghostscript you are in point of fact sending a PostScript program to a PostScript language interpreter. In essence this is the same as sending a Python program to a Python interpreter. That is what I meant when I said that the program is trying to call a function named 'Hebrew' without defining such a function. All of that is from the error message | 07:53.44 |
| | <KenSharp> As I said on the bug report, you hadn't attached the PostScript program (file) so I wasn't able to test it. Now that you have I will look into it. | 07:54.33 |
| | <KenSharp> We don't do telephoen support, not even for commercial customers. I don't see who you sending me pictures of your desktop will help. Now that you have supplied the file I will look at it. | 07:55.16 |
| | <KenSharp> I see Ray has closed the bug report along with a more complete description of the problem with the file, which is not a valid PostScript program because it has somehow been corrupted.. | 08:01.27 |
| | <KenSharp> Given that this occurs only for the word Hebrew, and that this is the insertion of new characters, it seems likely that this is the result of some kind of search and replace, rather than the more usual cause of sending files between Windows and Linux, which would result in CR/LF translations. Whether the OpenText ExStream product corrupted the font when converting it from TrueType, or the damage was done by some later edi | 08:04.58 |
| | <RobertLM78> getting an error: GPL Ghostscript 9.26: ERROR: A pdfmark destination page 21 points beyond the last page 1. | 11:24.07 |
| | <RobertLM78> only the pdf has over 500 pages............. wth? | 11:24.23 |
| | <RobertLM78> calling with the usual '$ gs -sDEVICE=pdfwrite -q -dBATCH -dNOPAUSE -sOutputFile=output.pdf -dPDFSETTINGS=/prepress index.info -f input.pdf' | 11:25.22 |
| | <KenSharp> RobertLM78 well that suggests that you have an annotation with a Dest which references a page outside the range of pages in the file. | 11:35.00 |
| | <KenSharp> However, 9.26 is old so it may be a bug. | 11:35.08 |
| | <KenSharp> Without seeing the PDF file I can't comment | 11:35.14 |
| | <RobertLM78> I might have something wrong in my index.info file... just tried a different one from before and it worked | 11:35.57 |
| | <RobertLM78> wish i could see what the hell i was doing wrong XD | 11:36.11 |
| | <KenSharp> I presume index.info is a PostScript program ? | 11:36.33 |
| | <KenSharp> Which includes pdfmark operations ? | 11:36.40 |
| | <RobertLM78> Yes.. for ghostscript on a pdf file | 11:36.53 |
| | <KenSharp> Well the obvious thing to do is to binary chop the index.info file. Cut it in half and see if the error goes away, if it did then the error is in the first half, if it didn't cut the remaineder in half, and so on | 11:37.32 |
| | <RobertLM78> sorta what i've been doing ... seems to fail as soon as it hits the nested parts | 11:38.04 |
| | <KenSharp> Once you have a pdfmark which exhibits the problem for certain, you can put that here | 11:38.16 |
| | <RobertLM78> okay - thanks :) | 11:38.34 |
| | <KenSharp> It's hard for me to offer much help, without seeing the files | 11:38.35 |
| | <RobertLM78> http://cini.classiccmp.org/pdf/Commodore/C64%20Programmer's%20Reference%20Guide.pdf | 11:39.08 |
| | <KenSharp> That's a blast from the past ๐ | 11:39.35 |
| | <RobertLM78> :D | 11:40.32 |
| | <KenSharp> Looks to be a scan of the original documentation, each page is an image, which makes it kind of big. There are two outlines already, Page 1 and Page 2 | 11:41.04 |
| | <RobertLM78> so here's the start of my index.info file for this pdf (i've never had trouble indexing a partial file): | 11:41.28 |
| | <RobertLM78> https://pastebin.com/pwKEmYbY | 11:41.28 |
| | <RobertLM78> Yeah i saw that :( | 11:41.40 |
| | <KenSharp> Let me give this a quick try | 11:41.54 |
| | <RobertLM78> However it looks like they get pushed to the bottom after a successful indexing | 11:42.01 |
| | <RobertLM78> thanks :D | 11:42.14 |
| | <KenSharp> They will be processed after the pdfmarks in your PostScript file, so kind of yeah | 11:42.21 |
| | <Meepleschubser> Hi, I've been using GS for many years to convert PDF to Tiff and it also resizes it. But with the new PDF interpreter it doesn't seem to work anymore, I always get empty PDF files. Unless I remove FitPage. But then the content is cut off. | 11:42.46 |
| | <RobertLM78> here's an index from another book that worked fine for the original book and for this book if i strip off the extra pages: | 11:43.05 |
| | <RobertLM78> https://pastebin.com/qeZHVJbT | 11:43.06 |
| | <KenSharp> @Meepleschubser your best bet is to report a bug and attach an example file | 11:43.17 |
| | <Meepleschubser> Ok. I will do that. Thanks | 11:43.50 |
| | <KenSharp> NP In passing let me note that we've done a lot of work on the new PDF interpreter and there's a very good chance the problem is already fixed. We should be doing a new release within a month more or less | 11:44.33 |
| | <KenSharp> If you can build GS from source yourself, and want to give it a try please feel free. | 11:44.48 |
| | <KenSharp> here will be a release candidate to try next week. | 11:45.00 |
| | <KenSharp> Otherwise | 11:45.08 |
| | <KenSharp> RobertLM78 so I can duplicate the problem, I'm getting 5 errors | 11:46.22 |
| | <RobertLM78> lol.. well that would help.. i was missing a closing parenthesis on line 7 XD | 11:48.01 |
| | <KenSharp> Aha, well that's not ideal certainly | 11:48.18 |
| | <KenSharp> Yes (Screen Display Codes... is what I just narrowed it to | 11:48.57 |
| | <RobertLM78> Still having trouble on the next nested group though :( | 11:49.01 |
| | <RobertLM78> you can do double nested as the next group has correct? | 11:49.41 |
| | <KenSharp> This is throwing an error for me : | 11:49.44 |
| | <KenSharp> [/Page 49 /Title ((Introduction) /OUT pdfmark | 11:49.45 |
| | <RobertLM78> i wonder if that initial count is wrong | 11:49.50 |
| | <KenSharp> The /Count in Outlines is awfully confusing I'm afraid, bad specification on Adobe's part | 11:50.22 |
| | <RobertLM78> Lmao... XDD well that's embarrassing... i was starting to think that i had just made some silly error | 11:50.27 |
| | <RobertLM78> yeah.. it gets a bit ugly | 11:50.42 |
| | <Meepleschubser> A colleague just found out if I use EPSFitPage instead of FitPage it works again. | 11:51.03 |
| | <KenSharp> Hmm well that's surprising | 11:51.18 |
| | <KenSharp> Unless you are using an EPS file of course ๐ | 11:51.28 |
| | <RobertLM78> thanks @Ken 8-) | 11:52.03 |
| | <RobertLM78> .sometimes it's good to get another set of eyes on things | 11:52.11 |
| | <Meepleschubser> No. only PDF to tif | 11:52.13 |
| | <KenSharp> @Meepleschubser Well it is possible that whatever the problem was has been fixed, and may only have affected on of the controls anyway. If you could test the release candidate next week it would be helpful | 11:52.50 |
| | <KenSharp> RobertLM78 it 'looks like' it doesn't like the /Count from the page 20 reference. Almost like it's trying to continue that group | 11:53.39 |
| | <KenSharp> I may have to start a debugger | 11:54.11 |
| | <RobertLM78> Seems to be working now... just gotta watch those parentheses :D | 11:54.20 |
| | <KenSharp> Oh yeah.... | 11:54.32 |
| | <KenSharp> I see Intr=odcution | 11:54.53 |
| | <KenSharp> RobertLM78 there does seem to be a different problem with page 514 though. Wrong stream lengths | 11:55.46 |
| | <RobertLM78> heheh - yeah XD | 11:55.48 |
| | <KenSharp> And also on page 1 oddly | 11:56.05 |
| | <Meepleschubser> Release Candidate next week here in Discord? Or why can i get the Version (for windows) | 11:56.06 |
| | <RobertLM78> must be what this is about: | 11:56.45 |
| | <RobertLM78> **** Error: stream Length incorrect. | 11:56.46 |
| | <RobertLM78> Output may be incorrect. | 11:56.47 |
| | <RobertLM78> that's an issue with the pdf itself though, no? | 11:57.16 |
| | <KenSharp> I'm not sure if we announce the Release Candidate here, I guess we probably should (I'm not responsible for making the releases). We do announce them on gs-devel. When it's available you can pick it up from the Ghostscript web site on the Downloads page | 11:57.21 |
| | <KenSharp> Yes, it seems to be. It looks like it is non-fatal though | 11:57.36 |
| | <KenSharp> At least for the new PDF interpreter, not sure about the one you are using | 11:57.55 |
| | <KenSharp> Yes it seems to recover well enough | 11:58.39 |
| | <RobertLM78> Alright... starting to come along now.. all the way to chapter 4 now... it's a bit arduous and it's probably best to find a "pure" and unindexed pdf, but this works... and even though it is a slog, i sort of prefer to just do the indexing myself... unless it's been indexed by someone who does proper full indexing | 12:12.05 |
| | <KenSharp> I notice the original file has been edited several times | 12:12.26 |
| | <RobertLM78> interesting | 12:12.35 |
| | <RobertLM78> people probably passing it around making changes and passing it on some more..... 40 year old machine now | 12:13.15 |
| | <KenSharp> Yes I remember it well, not that I had one | 12:13.30 |
| | <RobertLM78> i cut my teeth with a TI-99 and i thought it would be fun to check out another 8-bit era machine | 12:13.46 |
| | <KenSharp> OK lunch time. If anyone has more questions feel free to post them and I'll see them later | 12:15.12 |
| | <RobertLM78> thanks again Ken B) | 12:16.09 |
| | <Till Kamppeter> Hi, I am not able to build Ghostscript any more on Ubuntu Jammy, I get | 13:30.51 |
| | <Till Kamppeter> ``` | 13:30.52 |
| | <Till Kamppeter> /usr/bin/ld: ./obj/tif_zip.o: in function `ZIPCleanup': | 13:30.53 |
| | <Till Kamppeter> tif_zip.c:(.text+0x106): undefined reference to `libdeflate_free_decompressor' | 13:30.54 |
| | <Till Kamppeter> /usr/bin/ld: tif_zip.c:(.text+0x118): undefined reference to `libdeflate_free_compressor' | 13:30.56 |
| | <Till Kamppeter> ``` | 13:30.57 |
| | <Till Kamppeter> Is there some new dependency? | 13:30.58 |
| | <KenSharp> @Till Kamppeter Not that I am aware of. We build Ghostscript on Linux machines with every commit to the repository and that all seems to be working as expected | 13:31.46 |
| | <KenSharp> Are you building using system shared libraries or using teh GS-supplied 3rd party sources ? | 13:32.06 |
| | <Till Kamppeter> @KenSharp Any hint what can be wrong in my case? | 13:32.24 |
| | <KenSharp> That looks to me like a change in libtiff | 13:32.30 |
| | <KenSharp> Yeah tif_zip is definiitely part of libtiff | 13:33.33 |
| | <Till Kamppeter> @KenSharp Is Ghostscript using always my system's libtiff or its own libtiff (I simply ran ./configure)? | 13:34.23 |
| | <KenSharp> I'd expect, if you ran ./configure, it would be using the GS libtiff and friends, especially is you haven't deleted any of the sub-directories. | 13:34.56 |
| | <KenSharp> In fact it looks like it can't find zlib | 13:35.05 |
| | <KenSharp> I think libdeflate_free_decompressor is part of zlib and it seems like it can't find that reference | 13:35.29 |
| | <KenSharp> Let me check what version of Ubuntu I have here | 13:35.53 |
| | <KenSharp> I know it built OK on that yesterday | 13:36.04 |
| | <KenSharp> I'm going to need to reboot, back in a minute | 13:36.49 |
| | <Till Kamppeter> I have the newest one, 22.04. | 13:36.59 |
| | <Till Kamppeter> ``` | 13:42.21 |
| | <Till Kamppeter> till@till-x1nano:~/ghostscript/x$ dpkg -S libz.so.1.2.11 | 13:42.22 |
| | <Till Kamppeter> zlib1g:amd64: /lib/x86_64-linux-gnu/libz.so.1.2.11 | 13:42.23 |
| | <Till Kamppeter> till@till-x1nano:~/ghostscript/x$ dpkg -l zlib1g | 13:42.25 |
| | <Till Kamppeter> Desired=Unknown/Install/Remove/Purge/Hold | 13:42.26 |
| | <Till Kamppeter> | Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend | 13:42.27 |
| | <Till Kamppeter> |/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad) | 13:42.29 |
| | <Till Kamppeter> ||/ Name Version Architecture Description | 13:42.30 |
| | <Till Kamppeter> +++-==============-======================-============-========================> | 13:42.31 |
| | <Till Kamppeter> ii zlib1g:amd64 1:1.2.11.dfsg-2ubuntu7 amd64 compression library - ru> | 13:42.32 |
| | <Till Kamppeter> till@till-x1nano:~/ghostscript/x$ | 13:42.34 |
| | <Till Kamppeter> ``` | 13:42.35 |
| | <KenSharp> Yeah I'm only on 21-20 Impish Indri | 13:43.00 |
| | <KenSharp> I'm afraid @chrisl is on vacation today, and the Ghostscript build system is arcane enough that I'm not really able to help. I'd suggest checking back on Monday @Till Kamppeter unless you find a solution before that | 13:43.54 |
| | <KenSharp> Hmm is that the Ubuntu package for Ghostscript ? | 13:45.13 |
| | <KenSharp> You might do better to clone our Git repository and try building from that | 13:45.26 |
| | <Till Kamppeter> OK, will move everything out to Monday, and then solve with @chrisl to match GS 9.56 with Ubuntu 22.04. I anyway also need to match my new Apple Raster support with @chrisl, It seems that he is prioritising rthat the ยจcups" output device uses the included CUPS library and for the Apple Raster I have to force the use of the system's one. | 13:47.38 |
| | <KenSharp> I don't think Chris intends to 'prioritise' the use of the supplied one, but we need to make it possible to use the supplied one, which your change to the makefile broke. Which meant we couldn't build the cups device on Windows | 13:48.36 |
| | <KenSharp> I thought Chris had written this so that the cups device could use the system cups libraries or the bundled one, but this is absolutely not my field, and you really do need to discuss it with him. | 13:49.16 |
| | <KenSharp> I'd guess he'll make a new VM and sintall the latest Ubuntu to try and replicate the build problem | 13:50.10 |
| | <Robin_Watts> @chrisl The solution to the release build issues came to me last night. | 14:29.56 |
| | <Robin_Watts> (or this morning, rather). | 14:30.05 |
| | <KenSharp> @Robin_Watts @chrisl is on vacation today so it can easily wait until Monday | 14:30.23 |
| | <Robin_Watts> For the first job, we specify "ref". For the second job, we need to specify "rev", but NOT "ref" | 14:30.35 |
| | <KenSharp> Or you can stick it here for an aide-memoire ๐ | 14:30.37 |
| | <Robin_Watts> Ah, well,in that case, I'll kick the tests off again | 14:30.59 |
| | <KenSharp> Please go ahead | 14:31.06 |
| | <KenSharp> Cluster is quiet right now | 14:31.12 |
| | <KenSharp> Yeah I'm only on 21-10 Impish Indri | 14:58.01 |
| | <RayJohnston> @Pritam I attached the fixed file for your benefit. As described, fixing it was pretty simple, but that would NOT work on files with binary contents which is implied by the DSC comment in the header: | 15:00.25 |
| | <RayJohnston> %%DocumentData: Clean8Bit | 15:00.26 |
| | <RayJohnston> @Pritam I attached the fixed file to the bug for your benefit. As described, fixing it was pretty simple, but that would NOT work on files with binary contents which is implied by the DSC comment in the header: | 15:00.45 |
| | <RayJohnston> %%DocumentData: Clean8Bit | 15:00.45 |
| | <KenSharp> Well the **real** problem is the OpenType ExStream converter I think. If you look at arial.ttf, the Copyright notice has ...All Rights Reserved.\n\nHebrew OpenType.... | 15:04.39 |
| | <KenSharp> That's perfectly legal in a TrueType font, but you can't just copy the legal notice verbatim to a PostScript comment, for precisely this reason, it might have line endings in it, and that terminates the PostScript comment. | 15:05.14 |
| | <KenSharp> I'd say this ought to be sent back to OpenType 'Your application has a bug' | 15:05.28 |
| | <KenSharp> Err That should be OpenText, not OpenType.... | 15:06.33 |
| | <<<Back 1 day (to 2022/02/24) | Forward 1 day (to 2022/02/26)>>> | |
Log of #ghostscript at irc.freenode.net.