| <<<Back 1 day (to 2016/08/30) | 20160831 |
tor8 | fredross-perry: you'll still need the annotation type enum to create new annotations | 10:24.30 |
deekej | HenryStiles: Thank you for the commit in regards to BZ #697032 | 10:26.45 |
| *commits | 10:26.58 |
| Small question - can I expect that these commits will make it to 9.20 release? | 10:27.23 |
chrisl | deekej: yes, the commits will be in the next release | 10:35.47 |
deekej | thank you all :) | 10:35.58 |
chrisl | deekej: I know this is a difficult question, but do you have any feel (or way to assess) if your users use the devices from our "contrib" directory? | 10:36.50 |
deekej | I will check the specfile, if we actually ship it | 10:39.32 |
| I don't know what previous maintainers had set up | 10:39.43 |
| still getting through that mess in Fedora :) | 10:39.57 |
chrisl | deekej: our problem is that many/most of those devices are ones we cannot test, and the original authors have long since moved on/lost interest. So, they are, essentially, unmaintained, and that makes us uneasy | 10:41.34 |
| FWIW, there is a list of the contrib devices at the top of the contrib/contrib.mak file | 10:42.08 |
deekej | OK, so looking into the specfile for current Fedora package - we do not ship those files at all | 10:45.08 |
| the only way people can get them is to download the source tarball and compile it/use it themselves | 10:45.33 |
chrisl | Really?!?!? Interesting..... | 10:45.51 |
deekej | yeah | 10:45.56 |
| sorry, I took over the ghostscript when I started to work at Red Hat | 10:46.13 |
| now I'm slowly putting my packages back into order | 10:46.26 |
chrisl | No, you're being very helpful. I'm just surprised (and pleased) at that. | 10:46.49 |
deekej | so, you do *not* want me to create a new sub-package for it, right? | 10:47.13 |
chrisl | Absolutely. As I said, those devices are untested and unmaintained, so we've genuinely no idea if they work | 10:47.47 |
deekej | I see. I know from my experience it's not easy to maintain legacy code | 10:48.13 |
| (I'm also maintainer of grub and tcsh... *sad sigh* :D ) | 10:48.30 |
| it might be possible that some other package use the contrib/ folder | 10:49.08 |
chrisl | How many Ghostscript packages are there? | 10:49.35 |
deekej | well, because of Fedora packaging policy, we have main ghostscript package with its subpackages | 10:50.09 |
| then there's also ghostscript-fonts, urw-fonts | 10:50.23 |
| lcms2 package, poppler-data package | 10:50.49 |
| on all of those packages the ghostscript compilation depends | 10:51.00 |
chrisl | Yeh, but only the package for the actual Ghostscript executable can use the contrib directory - the devices are built into the gs exe | 10:51.47 |
deekej | I guess the best way to find out if contrib/ folder is being used by some other packages, is to ask on fedora-devel mailing list | 10:51.59 |
| ah, ok | 10:52.21 |
chrisl | IIRC, there's two packages for Ghostscript binaries - one with X11 and one without: does that sound right? | 10:52.54 |
deekej | in Fedora, the gs exe itself is not being used exactly - the --enable-dynamic is being used with ./configure script, and we use the gsx instead | 10:53.46 |
| so some other applications can use the ghostscript as a "library" | 10:54.14 |
| yes | 10:54.19 |
chrisl | Yeh, so you either gs libgs.so with X11 support, or without | 10:54.38 |
deekej | we have *-gtk and *-X11 package | 10:54.47 |
chrisl | Okay, so three: noX, X and gtk | 10:55.10 |
deekej | yes, the *-core contains the libgs.so (and ligijs-*.so), then the *-gtk contains gsx binary, and *-x11 contains the X11.so | 10:56.15 |
chrisl | So, it's pretty much certain that you don't use those contrib devices. | 10:56.55 |
deekej | I guess it is :) | 10:57.18 |
| sorry, I will be right back | 10:57.26 |
chrisl | Coo, that sounds like we have fairly reasonable cause to remove contrib..... | 10:58.04 |
deekej | I'm glad I can help :) | 11:40.13 |
norbertj | hi guys, I have this postscript job with code 'currentfile 43198872 () /SubFileDecode filter /FlateDecode filter cvx exec' followed by the binary data. Is there a way (like pdfinflt.ps for pdf) to deflate this? | 13:11.38 |
| could not figure this out | 13:12.00 |
chrisl | norbertj: there isn't a predefined tool for doing that. | 13:17.17 |
kens | Put the binary data in a file and then execute this PostScript : | 13:19.05 |
| { | 13:19.05 |
| InFile Data readstring { | 13:19.06 |
| OutFile exch writestring | 13:19.06 |
| } { | 13:19.07 |
| OutFile exch writestring | 13:19.07 |
| exit | 13:19.08 |
| } ifelse | 13:19.08 |
| } loop | 13:19.09 |
| InFile closefile | 13:19.09 |
| OutFile closefile | 13:19.10 |
| OOps, IRC didn't like that, try: | 13:19.33 |
| "/InFile (/bugs/695417/font.a85.txt) (r) file /FlateDecode filter def | 13:19.34 |
| "/OutFile (/bugs/695417/font.ttf) (w) file def | 13:19.34 |
| "/Data 32768 string def | 13:19.34 |
| " | 13:19.34 |
| "{ | 13:19.34 |
| " InFile Data readstring { | 13:19.34 |
| " OutFile exch writestring | 13:19.35 |
| " } { | 13:19.35 |
| " OutFile exch writestring | 13:19.36 |
| " exit | 13:19.36 |
| " } ifelse | 13:19.37 |
| "} loop | 13:19.37 |
| " | 13:19.38 |
| "InFile closefile | 13:19.38 |
| "OutFile closefile | 13:19.39 |
| But remove the quotes | 13:19.41 |
chrisl | thinks "pastebin"..... | 13:19.55 |
norbertj | chrisl: thanks will try that | 13:28.57 |
chrisl | norbertj: I was thinking kens should pastebin the code. | 13:29.59 |
norbertj | chrisl: I already copy/paste/changed it, going to try. | 13:32.33 |
| kens: thanks, works like a charm. | 13:44.25 |
HenryStiles | hi norbertj | 13:47.03 |
| sure deekej, actually I think we could pull a lot more out. | 13:49.40 |
chrisl | HenryStiles: deekej checked the Fedora package(s) for us, and they don't build in the contrib devices, so that's probably good evidence that we could remove contrib - which would be nice. | 13:51.09 |
norbertj | hi HenryStiles, long time no hear | 13:52.30 |
deekej | sorry guys, I need to leave now, but I will be back tomorrow... :) Cheers | 13:53.02 |
norbertj | HenryStiles: I entered a bug 697088 (with a fix ;). Can you have a look? | 13:53.43 |
HenryStiles | chrisl: the only thing I've heard complaint about is opvp from till | 13:53.48 |
| chrisl: but that was a few years ago | 13:54.24 |
chrisl | HenryStiles: Till complained about removing pretty much any of them about two years ago | 13:55.03 |
HenryStiles | norbertj: are Oce folks coming to Orlando? I'll miss the lake michigan swim. | 13:55.33 |
| that looks reasonable norbertj I'll have a closer look and regression test it today sometime. | 13:57.07 |
norbertj | HenryStiles: I have no idea. Sepideh is on holiday, others (Remco?, Ronald = replacement for Marten) I don't know I'll ask. | 14:00.13 |
| HenryStiles: It's specifically the plft_MSL at the bottom of the patch, I wonder if that would be correct. | 14:01.31 |
| Or should it remain plgv_MSL = 0, plft_MSL=10? | 14:04.28 |
HenryStiles | I'll have a look in a few minutes, just got up having some coffee and food | 14:05.31 |
norbertj | kens: what if the part before the binary data is: currentfile 22740 () /SubFileDecode filter /FlateDecode filter /ReusableStreamDecode filter | 14:15.33 |
| it looks like this needs some tweaking | 14:15.48 |
kens | Just use FLateDecode | 14:15.49 |
| ReusableSTreams are a different thing | 14:16.01 |
| In PostScript they are all streams, but reusbaleStreams aren't compression filters | 14:16.26 |
norbertj | kens: thanks. Have to go home now. | 14:22.16 |
| bye | 14:22.19 |
fredross-perry | tor8: I'm ok with using annotation type enums inside classes, but I don't see the need to make them public. I'd want to do something like new InkAnnotation(), and have the class hide the fz_ details. | 17:31.10 |
| Forward 1 day (to 2016/09/01)>>> | |