| <<<Back 1 day (to 2014/05/09) | 2014/05/10 |
henrys | mvrhel_laptop: cool stuff - I printed to my PCL printer from gsview and it produced nice hpgl/2 lines and beziers. | 00:51.26 |
mvrhel_laptop | henrys: oh cool | 04:05.31 |
| ok. I need to do some work to make "open with" work and also drag and drop work | 06:17.06 |
| maybe will get that going this weekend | 06:17.17 |
kens | I'm just pulling off the installer now | 06:17.19 |
mvrhel_laptop | ok. there are several things as I work with it that I find I need to fix. | 06:17.46 |
| so right now you have to open the file from the applicaiton | 06:18.10 |
kens | Well you did say it was a beta | 06:18.10 |
mvrhel_laptop | yes | 06:18.13 |
kens | Definitely need to ask before adding shortcuts to the desktop, I dislike apps that don't ask firtst. | 06:19.44 |
mvrhel_laptop | yes. I may drop that. | 06:20.02 |
| also it asks if you want to launch but that does not work for some reason | 06:20.14 |
kens | Unchecking 'launch the program', the check box stays checked, but the app doesn't launch.... | 06:20.20 |
| Oh right | 06:20.25 |
mvrhel_laptop | this is my first time fooling with this installer | 06:20.28 |
| need to go through the settings a bit more carefully | 06:20.49 |
kens | OK launching from the shortcut I get 'DllNotFoundExcepetion: MuPDF DLL not found' | 06:21.00 |
mvrhel_laptop | really? | 06:21.12 |
kens | Yep | 06:21.15 |
| app opens but 'File' is grayed out | 06:21.28 |
mvrhel_laptop | hehe | 06:21.32 |
| yes | 06:21.33 |
| if it does not have the mupdf dll you are dead in the water | 06:21.41 |
| so do me a favor | 06:21.52 |
| find where the application is located | 06:22.05 |
kens | "C:\Program Files (x86)\Artifex\gsview\" | 06:22.14 |
mvrhel_laptop | and see that there are 2 DLLs in there | 06:22.20 |
kens | Should NOT be in (x86) as its 64-bit | 06:22.25 |
mvrhel_laptop | are you on 32 bit machine? | 06:22.26 |
kens | No, 64 -bit | 06:22.32 |
mvrhel_laptop | right | 06:22.43 |
| should be in programs | 06:22.49 |
| hmmm | 06:22.53 |
kens | Should be, but isn't | 06:22.55 |
mvrhel_laptop | in any event, are there two DLLs there? | 06:23.11 |
kens | It is in (x86) and tehre are 2 DLLs there | 06:23.17 |
| Let me just copy to the right place | 06:23.30 |
mvrhel_laptop | still that is not right | 06:23.41 |
| are the dlls ending in 64 in there names? | 06:23.59 |
| their | 06:24.04 |
kens | Same error launching from program filesz | 06:24.06 |
mvrhel_laptop | kens: ok bummer. | 06:24.23 |
kens | Yeah I have a gsdll64.dll and mupdfnet64.dll | 06:24.27 |
| and 32 versions of each also | 06:24.44 |
mvrhel_laptop | ok so I wonder why it does not find it when it is right there | 06:24.49 |
| huh | 06:24.51 |
| the 32 are there too? | 06:24.56 |
kens | Where are the registry keyts ? | 06:25.01 |
| Yes, 32 in both | 06:25.06 |
mvrhel_laptop | no registry stuff yet | 06:25.09 |
kens | Oh.... | 06:25.14 |
| let me try command shell | 06:25.27 |
mvrhel_laptop | I am puzzled why the 32s are in there | 06:25.52 |
kens | Same again | 06:25.54 |
mvrhel_laptop | kens: ok dont spend any more time on it then | 06:26.12 |
kens | FWIW the GS dll seems to be OK | 06:26.16 |
mvrhel_laptop | I will try it on a few machines here this weekend | 06:26.28 |
kens | GPL Ghostscript GIT PRERELEASE 9.15 64 bit | 06:26.33 |
| OK let me know if I can do somethign to help | 06:26.45 |
| I've got guests this weekend so I won't be around much but I'll read the logs | 06:27.03 |
mvrhel_laptop | ok. np thanks kens | 06:27.11 |
kens | BTW What's the DLL supposed to be called ? | 06:27.29 |
mvrhel_laptop | oh kens | 06:27.34 |
| did you grab the one in the gsview directory? | 06:27.44 |
kens | Yes | 06:27.47 |
| As per your email | 06:27.55 |
mvrhel_laptop | ok. | 06:28.08 |
kens | That's from casper, is htere a better one ? | 06:28.19 |
mvrhel_laptop | becuase i see I had one in my home directory | 06:28.25 |
kens | Hmm, didn't notice that | 06:28.34 |
mvrhel_laptop | but you need the one in gsview/ | 06:28.35 |
kens | THe one I have is 26,135 kB | 06:29.04 |
| 26,761,451 bytes | 06:29.27 |
mvrhel_laptop | ok that looks right | 06:29.33 |
| wtf | 06:29.36 |
kens | No idea, its peculiar that it can find the GS dll ok, but not the MuPF one | 06:29.54 |
| Would suggest a naming problem to me | 06:30.07 |
mvrhel_laptop | yes. but even more odd to me is that you have the 32s in there | 06:30.09 |
kens | Ah well, don't know about that. | 06:30.21 |
| I assumed you wanted to install both | 06:30.28 |
mvrhel_laptop | no | 06:30.31 |
kens | I could delete and reinstall | 06:30.35 |
mvrhel_laptop | I doubt that will matter. I can't fathom how you have them | 06:30.50 |
| there are only 3 files in the installer | 06:31.07 |
kens | Me neither, my setup is slightly unusual, but the program files are in teh normal places | 06:31.17 |
| THat's weird then | 06:31.34 |
| I defintiely have 5 files here | 06:31.43 |
| 4 DLLs and an executable | 06:31.56 |
mvrhel_laptop | you are running win 7 | 06:32.00 |
| 64 bit | 06:32.02 |
kens | Indeed | 06:32.06 |
mvrhel_laptop | ok | 06:32.08 |
kens | Hmm, the MuPDF dlls are dated 8/05/2014, the GS ones and executable 09/05/2014 | 06:32.59 |
mvrhel_laptop | oh that is very wrong | 06:33.08 |
| actually all those dates are odd | 06:33.25 |
| ok. I give up for now | 06:33.31 |
kens | Yeh its late for you | 06:33.38 |
| Sleep on it for a bit :-) | 06:33.47 |
mvrhel_laptop | I will try on my son's windows 7 machine tomorrow if I have time | 06:33.54 |
kens | I can also try a Windows 8 VM | 06:34.05 |
mvrhel_laptop | 2 baseball games, a softball game and 2 dance recitals tomorrow | 06:34.11 |
kens | Wow, busy day | 06:34.17 |
mvrhel_laptop | so I suspect it will be hard to get much done | 06:34.21 |
kens | No problem, I have plenty to keep me busy | 06:34.33 |
| back to shadings I think | 06:34.43 |
mvrhel_laptop | ok well thanks for giving it a try kens. I will see what I can do | 06:34.47 |
| good night | 06:34.50 |
kens | night Michael | 06:34.59 |
| mvrhel_laptop (for the logs) tried installing on my WINdows 8 VM and had the same problem. Installed into (x86) and the DLL can't befound. The GS DLL does seem to load (assuming that's where you get the version from) | 06:44.24 |
| Oh, and I'm getting the 32 bit DLLs there as well | 06:44.47 |
Robin_Watts | Downloads appear to be working again. | 08:04.06 |
kens | Ah forgot to check, good to know. | 08:04.19 |
tor8 | Robin_Watts: ping | 10:44.11 |
| three fixes on tor/master ... one of them pretty serious | 10:45.39 |
Robin_Watts | looking now. | 11:07.20 |
| printf one looks good. D'Oh. | 11:09.38 |
| It's interesting the 'similarity' figures in the commit diff for the new cmaps. | 11:10.42 |
| that would seem to indicate that we can save lots of space by commoning up areas. | 11:10.57 |
| All look fine to me. | 11:12.11 |
tor8 | yeah. most of the JIS UTF8 ones look *really* similar | 11:12.49 |
| sadly the new CMaps combined with actually supporting 32-bit code points (instead of just skipping them) means the cmap data size has doubled | 11:13.23 |
| but it's the right thing to do... | 11:13.29 |
| with a bit of looking or merging data, we might be able to make either a superset CMap with all of the JIS encodings and use that for all of them | 11:14.24 |
| or a common subset, and usecmap include that and have overrides | 11:14.37 |
Robin_Watts | is there a downside to having a superset cmap? | 11:14.51 |
tor8 | I don't think it'll be a major drawback of encoding characters that we should be unable to encode with a given standard | 11:15.39 |
| but I suspect there may be minute code point differences | 11:15.46 |
| there are roughly 170 code points that differ in UniJIS-UTF8-H and UniJIS2004-UTF8-H | 11:18.29 |
| encoding that difference with usecmap directives and a common subset would be a lot smaller than having the two full CMaps | 11:19.02 |
Robin_Watts | yeah. just pondering if there is a tool we can (mis)use to spot such repetitions. | 11:20.22 |
| diff maybe. | 11:20.23 |
tor8 | given how 'messy' the input data is, round tripping the cmaps through our parser and pretty-printing the final ranges would probably be better for this | 11:21.22 |
Robin_Watts | then diffing that? | 11:21.44 |
tor8 | the diff between a -UCS2 and a -UTF16 are of this form: | 11:21.49 |
| -<00b4> <00b4> 645 | 11:21.56 |
| +<00b4> 645 | 11:22.01 |
| so yeah. cleaning them, then using diff would probably be good enough | 11:22.26 |
| Robin_Watts: on tor/cmapclean, make build/debug/cmapclean to create a script that pretty-prints a sorted canonical version of a CMap for use with mupdf | 12:37.24 |
| cmapclean2 to print a more compact version, but not as amenable to running diff | 12:37.44 |
| rigel ~/src/mupdf/resources/cmaps/japan $ diff UniJIS-UTF32-H UniJISX02132004-UTF32-H|diffstat | 12:38.12 |
| unknown | 463 +++++++++++++++++++++++++++++++--------------------------------- | 12:38.12 |
| 1 file changed, 230 insertions(+), 233 deletions(-) | 12:38.12 |
| if we figure out a way to use that, we can probably shave the sizes back down significantly | 12:38.35 |
ray_laptop | I found something that may be the root of my 4-bit grayscale differences. FINALLY | 16:24.50 |
| I have the image scaling correct, but the thresholding was "off". It seems like I need to invert the threshold array (or make the thresholding logic work the other way around) | 16:26.29 |
| I finally constructed a highly asymmetric threshold array and it makes it clear. | 16:27.41 |
| Forward 1 day (to 2014/05/11)>>> | |