| | <<<Back 1 day (to 2016/03/01) | 20160302 |
| kent3116 | Hi - I'm getting a Link 1000 Internal error during IMAGE::BuildImage | 07:05.46 |
| | followed by two C1001 An internal error has occurred in the Compiler bio.c Line 1 | 07:06.29 |
| | If I omit all occurances of fz_try(ctx) ... fx_catch(ctx) ... fx_always(ctx) then they don't occur | 07:09.15 |
| | whoops - I meant fz_try(ctx) ... fz_catch( ctx) ... fz_always(ctx) | 07:15.06 |
| | Okay - ignore the fz_try . That was just misleading due to it hitting a fz_rethrow_message as a result of commenting out the fz_catch :) | 07:40.34 |
| | After a bit more digging - the error occurs as soon as it hits fz_pre_scale(fz_rotate(&ctm, rotation), zoom, zoom); | 07:40.58 |
| malc_ | tor8: warning: assuming signed overflow does not occur when assuming that (X - c) > X is always false [-Wstrict-overflow] (for the logs) | 07:48.08 |
| | bad bad bad | 07:48.11 |
| kent3116 | Okay - please ignore above. FYI: It was a compiler optimisation set to 'Full Optimisation'. Changing to just 'Fastest' resolved the issue, so it appears it was linked to Minimium Size (/O1) flag. Made my project 1Mb larger, but at least it compiles | 08:04.12 |
| tor8 | kent3116: beware of optimizing compilers! :) | 10:03.06 |
| aditsu_ | hi, does gs automatically reduce the image resolution/quality? how can I stop that? I'm converting files with -sDEVICE=pdfwrite | 12:21.59 |
| kens | aditsu_ : No it doesn't, the controls for ths are documented in the ps2pdf.htm file in the doc directory | 13:03.48 |
| aditsu_ | kens: it seems that it does (reduce quality), I found this in the meantime: http://superuser.com/questions/360216/use-ghostscript-but-tell-it-to-not-reprocess-images | 13:05.51 |
| | looks like I can ask it to use lossless compression, but that increases the file size | 13:06.24 |
| kens | aditsu_ : No, it doesn't unless asked to. Of course if your original image is a JPEG and you compress as a JPEG you will experience the usual prolems fop requantising | 13:06.43 |
| aditsu_ | I'm converting pdf files with images inside, most of them are in jpeg format | 13:07.21 |
| kens | All images are decompressed, that's hte nature of hte wat Ghostscript works. If you reapply JPEG (which is a lossy compression system) then you will, obviously, lose more fidelity | 13:08.04 |
| aditsu_ | I'd like to preserve the original encoded images (just copy the same bytes) | 13:08.06 |
| kens | You can't. | 13:08.11 |
| | There's an open enhancement request for that, for JPEG, one day I may even get the time to address it, but I wouldn't be holding my breath | 13:09.09 |
| aditsu_ | also, it seems to use jpeg compression by default (which is fine, but I'm just saying I didn't explicitly request it) | 13:09.47 |
| kens | |That's quite a different questoin to 'does Ghostscript automatically reduce the image resolution/quality' | 13:10.35 |
| | Any processing of a JPEG image, followed by resaving as a JPEG will result in losses | 13:10.52 |
| | Moral of the story: Don't use lossy compression, or if you must make sure you don;t use it until the last step of your workflow | 13:11.20 |
| aditsu_ | ok, fine, as you say; how can I control the jpeg compression/quality rate when using pdfwrite? -dJPEGQ doesn't seem to work here | 13:12.40 |
| kens | Read teh documentation, its compliocated | 13:14.20 |
| | -dJPEGQ only affects the JPEG output devices | 13:14.33 |
| | Read ps2pdf.htm and look for the ColorImageFilter | 13:14.55 |
| | And also ColorImageDict | 13:15.28 |
| | Oh, and you will need to be using the current master code from Git if I remember correctly. I thnk I fixed a bug there that was preventing the JPEG dictionary having any effect. | 13:16.18 |
| aditsu_ | x_x | 13:17.31 |
| kens | THat may not be correct, I can't see the commit anywhere at the moment so I may have decided it was OK | 13:18.17 |
| | Ah no, its OK, it was only for files in PostScript using DCTEncode | 13:18.54 |
| | pdfwrite seems to have been immune | 13:19.03 |
| | Aha, in fact it is correct, ths commit http://git.ghostscript.com/?p=ghostpdl.git;a=commit;h=a7655b5d2fc42217eac71efd01f22fe3cca33d4a | 13:19.28 |
| | "This causes a *very* large number of diffs, because the pdfwrite (and | 13:19.41 |
| | ps2write) devices now actually use the parameters defined for these values | 13:19.41 |
| | resulting in small differences in DCT (JPEG) output." | 13:19.41 |
| aditsu_ | it would be nice to make it less complicated; something like PDFSETTINGS but without changing the dpi | 13:22.44 |
| kens | This is the way Adobe defined the settings, please address complaints to Adobe :-) | 13:23.12 |
| aditsu_ | I don't think Adobe is stopping you from defining additional options for gs | 13:24.18 |
| kens | When we have ot emulate Adobe, why would we add a second method ? | 13:25.00 |
| aditsu_ | I don't know, to help your users? :) | 13:25.51 |
| kens | They havce a mtehod, its defined. | 13:26.00 |
| | Our payiong customers seem happy | 13:26.10 |
| aditsu_ | ok, well, thanks for the help | 13:29.22 |
| tkamppeter | Hi, who works on the libopenjpeg fork included in Ghostscript? | 17:56.25 |
| HenryStiles | tkamppeter: probably best to talk to chrisl about this next week. We are all getting ready to travel to a meeting in san francisco | 18:09.17 |
| | tkamppeter: note you need git log --follow to look at file log history now, well at least history previous to the big directory reorganization last summer | 18:15.47 |
| RobinWattsLenovo | HenryStiles: What time are you arriving? | 23:27.55 |
| RobinWattsLenovo | runs | 23:28.03 |
| | Forward 1 day (to 2016/03/03)>>> | |
IRC Logs