Just now the 1.0.1 version of splix got released. Reproducible: Always Steps to Reproduce:
Version 1.0.1-1 was released to fix a issue with the Samsung ML-2010 printers that was introduced in 1.0.1. Depends on cupsddk now, which is masked and needs a amd64 keyword. I did a test and cupsddk will compile on a amd64 box if anyone else is interested.
Unfortunately, /usr/libexec/cups/filter/rastertospl2 from version 1.0.1 (and 1.0.1-1) segfaults on my machine. I've already sent a mail to splix-devel, still waiting for a reply. Until this is solved, I won't add 1.0.1 (or 1.0.1-1) to Portage. I don't have a problem taking cupsddk from Sunrise and putting it into Portage, once splix is fixed.
Understood, thanks for your response.
Does rastertospl2 work for you, Greg?! You can try it this way: /usr/libexec/cups/filter/rastertospl2 1 root foo 1 1 /etc/issue
Yeah has been working fine for me on my amd64 box.
Please, upgrade ebuild of splix to 1.0.1-1 version as it should fix "internal errors" https://sourceforge.net/tracker/?func=detail&atid=874746&aid=1655236&group_id=175815
I already tried 1.0.1-1 on my local machine some days ago -- same results as with 1.0.1 :(
Created attachment 112268 [details, diff] fix segfault when color attribute is missing This fixes segfault when calling rastertospl2 directly from shell with $PPD unset or pointing to a ppd-file without QPDL Color attribute. splix author is informed via mail.
Created attachment 112271 [details, diff] fix samsung ml2010 resolution bug this is a diff of splix vesion 1.0.1-1 and 1.0.1. it can be integrated into 1.0.1 without the need to import cupsddk for 1.0.1-1.
please try the above patches and if they work stabilize splix 1.0.1.
Created attachment 112997 [details] splix-1.0.1-r1.ebuild Updated ebuild to include patches submitted by Sascha Lucas
Works great Updated the ebuild to include them. Hopefully this gets into portage sometime soon.
Comment on attachment 112997 [details] splix-1.0.1-r1.ebuild Updated ebuild to include patches.
Created attachment 114180 [details, diff] as-needed.patch Fix as-needed linking, the makefile was buggy
(In reply to comment #14) > Created an attachment (id=114180) [edit] > as-needed.patch > > Fix as-needed linking, the makefile was buggy Please explain your fix -- I don't understand why it should yield anything, as the Makefile variables are AFAIK already only *extended* using the += operator.
(In reply to comment #9) > Created an attachment (id=112271) [edit] > fix samsung ml2010 resolution bug > > this is a diff of splix vesion 1.0.1-1 and 1.0.1. it can be integrated into > 1.0.1 without the need to import cupsddk for 1.0.1-1. Sascha, this patch is already contained in 1.0.1-1 upstream. Also, what does it have to do with cupsddk?!
(In reply to comment #8) > Created an attachment (id=112268) [edit] > fix segfault when color attribute is missing > > This fixes segfault when calling rastertospl2 directly from shell with $PPD > unset or pointing to a ppd-file without QPDL Color attribute. splix author is > informed via mail. Ok, with this patch, /usr/libexec/cups/filter/rastertospl2 does not segfault anymore and the printer starts printing, but unfortunately spits out a blank page. Downgraded to 1.0.1_beta2 -- prints fine (aka "non blank" =)). The printer is a ML-1610 -- maybe there's more broken regarding non-color printers?!
It is incorrect to add the needed libraries to the ldflags(these are used only for auxiliary linker option and not for adding dependencies). And the needed libraries should come on the end of the invocation, the options in the beginning.
Wolfram: (In reply to comment #16) > (In reply to comment #9) > > Sascha, this patch is already contained in 1.0.1-1 upstream. yes I know, thats where I've it taken from. > Also, what does it have to do with cupsddk?! As Greg stats in comment #1 version 1.0.1-1 depends now on cupsddk. to save you some work (importing cupsddk) you can stabelize version 1.0.1 and have the same codebase as with v. 1.0.1-1 but without cupsddk.
(In reply to comment #17) > Ok, with this patch, /usr/libexec/cups/filter/rastertospl2 does not > segfault anymore and the printer starts printing, but unfortunately > spits out a blank page. > Downgraded to 1.0.1_beta2 -- prints fine (aka "non blank" =)). > The printer is a ML-1610 -- maybe there's more broken regarding > non-color printers?! I have the ML-2010, which is just a 1610 with 8MB RAM, isn't it? This works perfectly.
just to comment on this. In comment #2 Wolfram Schlich said he had no problems importing cupsddk from the Sunrise Overlay. I don't mind having all the patches if that is what it takes to get this working in portage. I do find it a bit much though when all we need is cupsddk versus 3 patches. Just my opinion on it.
(In reply to comment #17) > [...] > Ok, with this patch, /usr/libexec/cups/filter/rastertospl2 does not > segfault anymore and the printer starts printing, but unfortunately > spits out a blank page. > Downgraded to 1.0.1_beta2 -- prints fine (aka "non blank" =)). > The printer is a ML-1610 -- maybe there's more broken regarding > non-color printers?! Any news on this? splix-1.0.1_p1 is surely not going stable with my printer printing blank pages ;)
I could confirm that my ML-2010 works perfectly fine with all the patches here
my ml-2010 works also fine, please count me too. and please do the revbump with this patches here, at least for ~arch. Wolfram: what are you trying to print, wich does not work? the cups test page?
splix-1.0.1 does not compile for me. I have cups-1.2.11. Don't really have the time right now to test further but it could be as-needed related.
Gergan Penkov's as-needed.patch is proper. It's incorrect to add libraries to be linked in in LDFLAGS. That's what LIBADD is for. As per the dependency ordering of the linker, the libraries must come AFTER the object being linked. That being said. All the patches together here compile for me with cups 1.2.11. The current 1.0.1 ebuild does not compile and work for me.
I have an ML-2010, and my printer too prints blank pages. Before, rastertospl2 segfaulted.
(In reply to comment #27) > I have an ML-2010, and my printer too prints blank pages. Before, rastertospl2 > segfaulted. can you give a reproduceable example (postscript / pdf etc)? I would like to test it on my ML-2010. If there are blank pages too, then upstream needs to be informed.
1.0.1-1 in CVS as 1.0.1_p1. Thanks everybody!