http://www.ghostscript.com/article/65.html Reproducible: Always Steps to Reproduce: 1. 2. 3.
requires libpng-1.2.8 according to http://www.cs.wisc.edu/~ghost/doc/AFPL/get850.htm, which is not currently in portage. filed a ebuild req as Bug #74456 :)
*** Bug 75842 has been marked as a duplicate of this bug. ***
i'm stuck on this. everything builds fine until the ijs portion, then it dies on the man page of all things. /bin/sh ./libtool --mode=link i686-pc-linux-gnu-gcc -O2 -march=pentium3 -pipe -fomit-frame-pointer -funit-at-a-time -ftracer -fforce-addr -fweb -ffast-math -Wl,-O1 -Wl,--sort-common -o ijs_server_example ijs_server_example.o libijs.la i686-pc-linux-gnu-gcc -O2 -march=pentium3 -pipe -fomit-frame-pointer -funit-at-a-time -ftracer -fforce-addr -fweb -ffast-math -Wl,-O1 -Wl,--sort-common -o ijs_server_example ijs_server_example.o ./.libs/libijs.a make: *** No rule to make target `ijs-config.1', needed by `all-am'. Stop. !!! ERROR: app-text/ghostscript-afpl-8.50 failed. !!! Function src_compile, Line 90, Exitcode 2 (dies with sane flags as well) i'll post what i've done. it looks like CUPS isn't putting out a new patch, by the looks of their CVS, so i went and updated the one included w/ cups 1.1.23. it applies cleanly but hasn't been tested. i updated the DEPENDS and made sure all the sed lines still apply. i changed the search path fix to use ${PV} instead of a hardcoded version. commenting out the ijs lines results in a working build. i don't have a printer, so that suits me okay, but it's probably not the best for most people. ;]
Created attachment 47807 [details] incomplete gs-afpl 8.50 ebuild
Created attachment 47808 [details, diff] gs850-lib.patch updated CUPS patch. needs testing.
Created attachment 47810 [details] incomplete gs-afpl 8.50 ebuild take two... the complete incomplete ebuild.
Created attachment 47811 [details] incomplete gs-afpl 8.50 ebuild last time, i swear. just realized that ${PV} doesn't work well inside of sed. fixed.
>updated CUPS patch Does this patch work for cups 1.1.23 which was officially launched on 1-3 but has not yet shown up in Portage?
yes it should. it patches against ghostscript files, rather than cups files, and from the looks of the cups CVS, there had been no changes to the pstoraster directory for quite a while.
Dying with this error: T_H -DGX_COLOR_INDEX_TYPE="unsigned long long" -O2 -march=pentium4 -fomit-frame-pointer -pipe -s -I./obj -I./src -o ./obj/gdevcups.o -c pstoraster/gdevcups.c pstoraster/gdevcups.c: In function `cups_get_matrix': pstoraster/gdevcups.c:575: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c:575: error: (Each undeclared identifier is reported only once pstoraster/gdevcups.c:575: error: for each function it appears in.) pstoraster/gdevcups.c: In function `cups_get_space_params': pstoraster/gdevcups.c:1023: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c: In function `cups_open': pstoraster/gdevcups.c:2432: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c:2512:67: macro "gs_malloc" requires 4 arguments, but only 3 given pstoraster/gdevcups.c: In function `cups_print_pages': pstoraster/gdevcups.c:2512: error: `gs_malloc' undeclared (first use in this function) pstoraster/gdevcups.c:2522:54: macro "gs_malloc" requires 4 arguments, but only 3 given pstoraster/gdevcups.c:2537: error: too few arguments to function `gs_to_exit' pstoraster/gdevcups.c:2585:55: macro "gs_free" requires 5 arguments, but only 4 given pstoraster/gdevcups.c:2585: error: `gs_free' undeclared (first use in this function) pstoraster/gdevcups.c:2586:76: macro "gs_free" requires 5 arguments, but only 4 given pstoraster/gdevcups.c:2589: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c: In function `cups_put_params': pstoraster/gdevcups.c:2706: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c: In function `cups_set_color_info': pstoraster/gdevcups.c:3223: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c: In function `cups_sync_output': pstoraster/gdevcups.c:3355: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c: In function `cups_print_chunked': pstoraster/gdevcups.c:3388: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c:3405: warning: passing arg 1 of `gs_to_exit' makes pointer from integer without a cast pstoraster/gdevcups.c:3405: error: too few arguments to function `gs_to_exit' pstoraster/gdevcups.c: In function `cups_print_banded': pstoraster/gdevcups.c:3541: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c:3560: warning: passing arg 1 of `gs_to_exit' makes pointer from integer without a cast pstoraster/gdevcups.c:3560: error: too few arguments to function `gs_to_exit' pstoraster/gdevcups.c: In function `cups_print_planar': pstoraster/gdevcups.c:4144: error: `gs_stderr' undeclared (first use in this function) pstoraster/gdevcups.c:4145: warning: passing arg 1 of `gs_to_exit' makes pointer from integer without a cast pstoraster/gdevcups.c:4145: error: too few arguments to function `gs_to_exit' make: *** [obj/gdevcups.o] Error 1 !!! ERROR: app-text/ghostscript-afpl-8.50 failed. !!! Function src_compile, Line 89, Exitcode 2 !!! make failed !!! If you need support, post the topmost build error, NOT this status message.
Created attachment 48870 [details] gs-afpl 8.50 ebuild Added use flags: ijs jasper jbig2dec rinkj. Not sure about jbig2dec dependency (built-in or package needed?). Changed "epatch gs850-lib.patch" to "epatch ${FILESDIR}/gs850-lib.patch" so it looks in the 'files' directory for the patch. Still non-working with cups, and also non-working with rinkj.
Created attachment 48871 [details] patch to enable rinkj support to be placed 'files' dir - enable the rinkj device does not compile here
Created attachment 48876 [details] Errors from patch while emerging USE="-cups ijs jasper jbig2dec rinkj" emerge -u ghostscript-afpl
With just the normal use=cups it fails building like above, with ijs rinkj etc I get the above patch failure.
Created attachment 48891 [details] replacement rinkj patch
Created attachment 48981 [details] replacement (#2) for the rinkj patch now patches correctly for those who had trouble with the previous one although the rinkj driver still does not compile here the patch does enable the driver per the instructions provided by the afpl-8.50 source
Just want to add that I think the ebuild is relatively complete and usable with the exception that the cups flag cannot be used as the cups gs-lib patch, although correctly implemented in the same manner it is for 8.11 and 8.15 simply does not work for 8.50 as there are other incompatible changes in the 8.50 release. Also, the small change the rinkj.patch makes, simply allowing the build of the rinkj device driver doesn't compile here. Maybe it's my gcc version or whatever, YMMV. Suffice it to say that AFAIK, without a new cups patch, 8.50 with cups support is a no-go. But if you don't need cups support but you do need jasper or jbig2 then this ebuild should work fine for you.
Created attachment 49360 [details, diff] Update gdevcups.c to work with gs-8.50 Looks like the internal GS library has changed somewhat, ie. writing to stderr won't work anymore and the prototypes for gs_exit, gs_malloc and gs_free have changed. I've tried to fix up gdevcups.c to work with this release, replacing debugging output with calls to dpf (which seems to be the intended routine for this sort of thing) and adding the required argument to the aforementioned gs_{exit,malloc,free) calls. As far as I can tell the patch is correct (used other driver files for reference).
*** Bug 90530 has been marked as a duplicate of this bug. ***
The ebuild provided works like a charm on my system (x86 stable). I have compiled [ebuild R ] app-text/ghostscript-afpl-8.51 +X -cjk +cups +gtk -ijs -jasper -jbig2dec -rinkj 0 kB [1] using gs850-lib.patch and gdevcups.patch. I have made a little patch to the ebuild so two sed commands are indipendent of ghostscript version :-)
Created attachment 60777 [details, diff] ghostscript-afpl-8.51.ebuild.patch
What is the ebuild provided for ghostscript-afpl-8.51, mentioned on comment #20 and to which the patch on comment #21 would apply?
(In reply to comment #22) > What is the ebuild provided for ghostscript-afpl-8.51, mentioned on comment #20 > and to which the patch on comment #21 would apply? I have used the ebuild for ghostscript 8.50: if you watch at src_compile() there are 3 sed operations which use the version of the ebuild, "8.50". The solution is to copy the ebuild for ghostscript 8.51 and upgrade those lines. This operation is useless as I know the ghostscript version (with ${PV}), so I have suggested the patch in #21
gs-afpl 8.50 ebuild with the patches you guys send don't work with USE="ijs" and the gdevcups.c patch don't work using cups 1.1.23 and gs-8.51. Any one of you can write a patch for cups 1.1.13 that works on gs-8.51 ? I Don't have time right now (closing another bugs) to build a patch for it. If you submit the patch i will commit gs 8.51 in cvs. Thanks to all and sorry for the delay.
Hi, I want to generate CMYK files from ghostscript. Although the bitcmyk option is available in ESP ghostscript (the default on Gentoo as app-text/ghostscript), I'd prefer the ghostscript-afpl version having spotcmyk or even better ghostscript-gnu version having PhotoShop psdcmyk file driver and also tiff32nc driver. As there are problem thats the ESP version has most drivers, I think I'd like to have concurrently at least app-text/ghostscript and app-text/ghostscript-gnu. Yes, I use cups for printing locally using Foomatic drivers. How about using --with-gs=NAME name of the ghostscript executible [gs] and placing the tree and binary under some separate place so that at least skilled users could call directly "gs-afpl blah blah" and "gs blah" and "gs-gnu blah". Let's say some shell wrapper called /usr/bin/gs would cal eithe rof them based on some ENV variable. Most people would not realize, but those who would might happily use /usr/bin/gs-gnu etc. I hope you got the idea.
ghostscript 8.51 does not seem to be in the build tree...looking at the traffic here, was this abandoned after having almost been done?
Look at my last comment. This package isn't a easy package to maintain. I have build an updated versions of all ghostscripts available in portage (esp,gnu,afpl) but this ebuilds are not ready yet to hit portage. If you guys want to help me with this test the ebuilds here and start fixing it. Thanks all for the reports, soon i will upload the latest versions to portage.
I'm ready to test this again, I need the tiffsep device which arrived in 8.3+ and is more complete in 8.50+, and future Scribus 1.3.1 users will really like that :). Which ebuilds/patches do I need to grab, or do you have URLs?
I ran into the same problems as Luis ie. gdevcups.patch didn't apply. Appearently this was caused by the fact that the level of subdirs of the original versus the patched files in the patch differed. I'm attaching a fixed version of the patch which applies cleanly here. I'm also attaching a consolidated ebuild (with the sed-modification). This works perfectly on my system.
Created attachment 69551 [details, diff] Fixed gdevcups.patch
Created attachment 69552 [details] gs-8.51.ebuild (Updated and consolidated)
Felix, your ebuild still depends on ghostscript version. Have you looked at the post #21?
Created attachment 70117 [details] patched ebuild
I have patched Felix' ebuild from comment #31 with the patch from comment #21 and verified the result manualy. This patch changes two sed lines slitely to use {PV} instead of hardcoded versions. The ebuild then builds cleanly on my x86 box: [ebuild R ] app-text/ghostscript-afpl-8.51 +X -cjk +cups -gtk -ijs -jasper -jbig2dec -rinkj 0 kB [1] The ebuild needs gs850-lib.patch and gdevcups.patch in the files dir. I can not test the new ghostscript any further atm. cause I have to apply another patch to make gs-afpl support my printer (Brother HL-5140 laserprinter). I will do this tomorrow and report back. If it works (as it does in afpl 8.15) should I post the patch for my printer here (or in a new wish) or should it not go into portage?
Cooool! It compiles, installs and seems to run ok. With the new devices, Scribus 1.3.1+ can now show the spot colour print preview :) Well done, and thanks!
(In reply to comment #35) Great to hear that it works for you Craig and indead a "well done" from my side too for the people who did the patches. Can you tell me if I need any of the USE flags (I only have X and cups atm) to benefit from the new ghostscript-afpl devices etc. in Scribus (Sorry, I am pretty much a ghostscript noob). And please forgive me the OT - I promise I will not extend this topic to much here in the bugger (or maybe you can answer via E-Mail instead?).
(In reply to comment #36) > (In reply to comment #35) > > Can you tell me if I need any of the USE flags (I only have X and cups atm) to > benefit from the new ghostscript-afpl devices etc. in Scribus (Sorry, I am > pretty much a ghostscript noob). And please forgive me the OT - I promise I > will not extend this topic to much here in the bugger (or maybe you can answer > via E-Mail instead?). > Mine were: [ebuild U ] app-text/ghostscript-afpl-8.51 [8.15] +X -cjk +cups +gtk -ijs -jasper -jbig2dec -rinkj 0 kB [1] So thats basically the same. You also need Scribus 1.3.1, and thats it other than setting up your spot colours in Scribus.
I have improved the rinkj patch: now the sources compile cleanly. I have used the following use flags: [ebuild R ] app-text/ghostscript-afpl-8.51 +X -cjk +cups +gtk +ijs +jasper +jbig2dec +rinkj 0 kB [2]
Created attachment 70140 [details, diff] rinkj.patch
When I wanted to slightly modify the ebuild to work with my printer I came across some sed lines in the ebuild, that look a little strange for me as a sed beginner: sed -i -e 's:/usr/local:/usr:' pstoraster sed -i -e "s:pstopcl6:pstopxl:" cups.mak sed -i -e 's:DEVICE_DEVS7=$(DD)faxg3.dev $(DD)faxg32d.dev $(DD)faxg4.dev :DEVICE_DEVS7=$(DD)faxg3.dev $(DD)faxg32d.dev $(DD)faxg4.dev $(DD)cfax.dev:' ${S}/Makefile.in 1. Question: I learned that sed -e is used to combine more then one sed command - but in all this examples it's just one sed command and the -e is not needed here IMHO. 2. What is the -i for? I couldn't find it in man sed. 3. What is the difference (if any) between sed 's:oldstring:newstring:' file sed "s:oldstring:newstring:" file and is there something like an "ebuild nettiquette" that syntax and symantic of e.g. sed is prefered or regarded as "standard"?
Created attachment 70674 [details, diff] ghostscript-afpl 8.52 ebuild http://sourceforge.net/project/shownotes.php?release_id=361878 Also needs gs852-lib.patch which I will attach in a mo.
Created attachment 70675 [details, diff] lib patch for gs 8.52
Created attachment 70676 [details, diff] ghostscript-afpl-8.52.ebuild with ijs fixed
Comment on attachment 70676 [details, diff] ghostscript-afpl-8.52.ebuild with ijs fixed Fix ijs compilation - ijs/configure is totally broken in the release
Thanks, I added the new ebuild to portage
Thanks to all for this one!
