| <<<Back 1 day (to 2018/05/22) | 20180523 |
chrisl | kens: I guess almost everyone you know has pointed you at this: http://www.computerhistory.org/_static/atchm/the-eudora-email-client-source-code/ | 08:15.49 |
kens | I think it came up on the Register at some point :-) | 08:16.13 |
chrisl | Well, it only went public yesterday, so..... | 08:16.45 |
kens | Maybe I'm thinking of something else then, could have sworn I heard of it before | 08:17.01 |
| Hmm, well maybe I'll have a bash at rebuilding it, might be an interesting project for weekends | 08:19.52 |
| chrisl you about ? | 12:43.45 |
chrisl | kens: yeh | 13:04.46 |
kens | The problem with the font is it has CIDSystemInfo dictionary which contains neither Registry nor Ordering | 13:05.08 |
| So it throws an undefined error in \buildfont11 | 13:05.25 |
| Do you recall changing that at all at any point ? | 13:05.37 |
chrisl | Correct, I'd think | 13:05.38 |
kens | Yeah thohse are required entries | 13:05.46 |
| Allegedly 9.22 didn't throw an error though | 13:06.15 |
| Ha ha no that's just b*llocks | 13:07.03 |
| It does throw an error | 13:07.08 |
chrisl | I was about to say.... | 13:07.10 |
kens | Oh apparently its 9.05 that doesn't, I just can't read | 13:07.30 |
| And that is indeed correct | 13:07.50 |
| I suspect we now use the CIDSystemInfo, which is why we complain if its missing | 13:08.24 |
chrisl | I think we'd only use it if we're doing subsitiution | 13:08.56 |
kens | That was what I meant, yes. But we definitely complain (now) if the information isn't there | 13:09.21 |
chrisl | Is the CIDFont not embedded? | 13:09.39 |
kens | I could pull out the old 9.05 code and see what it does but I'm disinclined to | 13:09.44 |
| The CIDFont is embedded, its atype 42, and its present | 13:09.54 |
| I disassembled it earlier and it looks OK | 13:10.05 |
| I suspect that if we ignored the CIDSystemInfor then it would 'probably' work | 13:10.22 |
chrisl | Unless code was added to validate CIDSystemInfo against the CMap, I can't imagine what else we'd have done | 13:11.09 |
| If you want to throw the file over to me, I can bisect it pretty quickly | 13:11.30 |
kens | No me neither, I was just wondering if you rememberd chaning it, but its so long ago it could be anything really | 13:11.41 |
| Its hte file attached to Bug 699294 | 13:11.57 |
chrisl | Oh, I thought you had a cut down one | 13:12.09 |
kens | I'm using the file as is, but you can check it more quickly by just running -dFirstPage=3 | 13:12.16 |
| I don't want to reduce the file (yet) because what I'm really interested in is why the following pages don't work either. The broken font I can understand. | 13:13.09 |
| coffee..... | 13:14.44 |
chrisl | kens: http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=8eb4118573d2d6959f8578a10f9d76ce9d802799 | 13:34.32 |
kens | :-) | 13:34.47 |
| Now if only I'd said *why* I wanted that..... | 13:35.17 |
chrisl | I'd have to guess for pdfwrite | 13:36.06 |
kens | I'm assuming so, presumably so we can write it out at the other end | 13:36.30 |
chrisl | Probably better than hard coding values <sigh> | 13:37.00 |
kens | Well that was pretty poor, yes | 13:37.12 |
| So it looks like we gort away with it previously because of those hard coded values | 13:37.27 |
chrisl | It's easy enough to fix, if you want me to do it | 13:37.37 |
kens | That meant that the CIDSystemInfo in the PDF file wasn't being checked | 13:37.42 |
| Hmm, how were you planning to fix it ? | 13:38.07 |
| Inserting substitutes for the missing values ? | 13:38.20 |
chrisl | Check if the entry exists, if so use it, if not fake it | 13:38.27 |
kens | I tried that, it still fell over. | 13:38.40 |
| But I might have made a mistake | 13:38.47 |
| Actually, I see that I did | 13:39.01 |
chrisl | I'd also probably opt for Identity rather than Japan1 | 13:39.10 |
kens | Identitty was what I used, Japan1 is just daft | 13:39.29 |
| I suppose we should 'fix' broken CIDFonts as well :-( | 13:39.50 |
chrisl | Fix the CIDFont, but print a "Warning: your CIDFont is buggered" message a million times before carrying on | 13:40.45 |
kens | A warnign is probably a good idea | 13:40.58 |
| FWIW fixing the cidsysteminfor lets acrobat open it | 13:41.12 |
| Ghostscript still throws errors (so does Acrobat to be fair) | 13:41.44 |
chrisl | Related errors? | 13:42.23 |
kens | Well it still says it can't process the font stream | 13:42.34 |
| And Acrobat mumbles something about fonts too | 13:42.44 |
| Acrobat says Bad font object or font descriptor object | 13:43.08 |
chrisl | Odd | 13:43.28 |
kens | Gs continues to throw errors on every page after the fotn error | 13:43.30 |
| Oh, there's more than one font with the same problem | 13:44.17 |
| about 6, that'll be why the other pages barf too I expect | 13:45.06 |
| Yeah fixing all those makes Acrobat happy | 13:45.26 |
| and Ghostscript too | 13:45.44 |
| I guess I shoudl fix it, since its technically me that broke it :-) | 13:46.00 |
chrisl | Yeh, with faked up entries, the file runs without errors/warnings | 13:47.00 |
kens | That's what I see yes | 13:47.11 |
| The reason the later pages fail is the same reaon, just different fonts | 13:47.22 |
SchleimKeim | Hello. Quick questions: is there a url where i can get the windows installers for older (< 9) Versions of ghostscript? | 13:58.58 |
kens | Yes, from our website | 13:59.09 |
| http://www.ghostscript.com | 13:59.14 |
SchleimKeim | kens: in that case i'm too dumb to find it. i'll check again | 13:59.50 |
kens | No hold on | 13:59.56 |
| the website is having a bad hari day | 14:00.02 |
| The link is at the bottom of this page: | 14:00.16 |
| https://www.ghostscript.com/download/gsdnld.html | 14:00.17 |
| But its broken | 14:00.21 |
SchleimKeim | kens: now i found the tiny link at the bottom :D | 14:00.37 |
kens | chrisl is that down to you ? | 14:00.38 |
SchleimKeim | but yeah, its broken | 14:00.48 |
kens | http://downloads.ghostscript.com/public/ | 14:00.54 |
SchleimKeim | yeah it seems down at the moment. well that's no problem. i'll just check the link again in a few days | 14:02.12 |
| i don't need it right now, i just wanted to make sure that older versions are still available :) | 14:02.27 |
kens | Its supposed to be there, but to be honest its not something that gets a whole lot of attention | 14:02.33 |
| You could (of course) always pull form teh tag in the Git repository and build from that :-) | 14:02.54 |
chrisl | We're only keeping about 4 versions back now. Why on earth would anyone want older versions....? | 14:06.10 |
SchleimKeim | 4 Versions back would be more than enough | 14:06.41 |
kens | has no clue | 14:06.45 |
| 4 version back won't take you back before 8.x | 14:06.59 |
SchleimKeim | kens: of course i COULD build it myself. the problem is that i really lack C++ experience, and building it would take me ages | 14:07.02 |
kens | Its C not C++ :-) | 14:07.12 |
SchleimKeim | kens: perfectly fine. version 8.x should suffice | 14:07.15 |
chrisl | Four versions back would get you 9.19 | 14:07.22 |
kens | And all you do is load it into Visal Studio and press teh 'build' button | 14:07.24 |
| chrisl even if its only 4 versions, the link should really work | 14:08.41 |
chrisl | Well, it did until a very short time ago | 14:09.07 |
kens | I thought it did too, but not today | 14:09.17 |
chrisl | We didn't renew the hosting | 14:09.30 |
kens | Ah! | 14:09.35 |
SchleimKeim | haha ^^ | 14:09.37 |
| chrisl: to answer your 'why': i need to insert some special comments into postscript files to further parse them and take them apart. converting certain PDF Files with gs9.X sometimes results in PS that can't be taken appart that easily. with 8.x it works fine | 14:11.03 |
chrisl | I suppose "fine" is a matter of perspective.... non-scalable Postscript isn't "fine" to me | 14:12.04 |
SchleimKeim | of course its a matter of perspective | 14:12.24 |
kens | No fonts, everything degenerated to rectangles, lack of colour space preservation, icky | 14:12.30 |
SchleimKeim | i'm not saying 9.x isn't better. i really just need it for certain PDFs | 14:13.45 |
| (cloning git://git.ghostscript.com/ghostpdl.git as we speak) | 14:14.17 |
kens | Over time the likelihood of those PDFs continuing to work in 8.x decreases | 14:14.23 |
SchleimKeim | i know | 14:15.06 |
chrisl | I've updated the link to point to github | 14:15.30 |
kens | Yes that's better | 14:16.15 |
| And goes back to 9.18 | 14:16.31 |
| Thanks chrisl | 14:16.37 |
SchleimKeim | cool, thx. in that case i hope i find the 8.x setups somewhere on my network | 14:18.19 |
chrisl | SchleimKeim: https://sourceforge.net/projects/ghostscript/files/GPL%20Ghostscript/8.71/ | 14:21.18 |
SchleimKeim | yaaay! :D | 14:21.53 |
chrisl | I'm wary of the alleged 64 bit binary.... | 14:22.26 |
| Also, of sourceforge's activities..... which is why we dropped it | 14:23.07 |
SchleimKeim | good decision | 14:23.19 |
| although i recently read somewhere that they got a new owner who intends to make it 'what it once was' | 14:24.02 |
| but IMHO that ship settled long time ago | 14:24.16 |
chrisl | SchleimKeim: BTW, have you looked at the ps2write output recently? | 14:25.47 |
SchleimKeim | as in 'do you know that it's deprecated' ? | 14:29.10 |
| i do know :) | 14:29.24 |
kens | No pswrite is deprecated, ps2write isn't | 14:29.41 |
| I think chrisl was hinting that you might well be able ot use ps2write, which produces much better output. Its also capable of injecting PostScript into the output on a per document or per page basis | 14:30.27 |
SchleimKeim | sorry, i'm a little sleep deprived. but yes i know that the devices changed after version 9 | 14:30.32 |
chrisl | Originally, ps2write's output was all compressed and wasn't DSC compliant - it now defaults to not compressed, and produces DSC compliant output | 14:30.45 |
| FWIW, pswrite survived up to 9.07 | 14:34.02 |
SchleimKeim | well, i collected lots of older versions now. thanks for the help and have a nice day! | 14:52.35 |
| Forward 1 day (to 2018/05/24)>>> | |