Re: [gentoo-user] Nvidia driver surprise
> On 2018-10-28, at 10:42, Davyd McColl wrote: > > Or perhaps go team red. Personally, I've had enough of team green's drivers. > My 660ti is still working fine, but if it dies, I think I'm going red. Not > saying the grass is 100% greener there (heh) but at least you can't be held > hostage with binary blobs. And they're cheaper. /2c > > -d We're not in good shape on either side here. I would use AMD if they had better performance on Linux with the open source driver. In terms of the long term, I don't see either open source driver supporting every device for as long as the drivers are supported. At some point there will probably be a split and then eventually those legacy drivers will be removed from the kernel due to lack of maintainers. Andrew
Re: [gentoo-user] Nvidia driver surprise
Apologies - green is nvidia and red is amd. No special code - that's just their primary marketing colors. -d On October 28, 2018 17:03:32 Philip Webb wrote: 181028 Davyd McColl wrote: 181028 Philip Webb : Perhaps I should buy a more upto-date graphics card. Or perhaps go team red. Personally, I've had enough of team green's drivers. My 660ti is still working fine, but if it dies, I think I'm going red. you can't be held hostage with binary blobs. And they're cheaper. Sorry, I don't understand your red/green code. Upgrade to the 1050 or so and you won’t have to upgrade for years. I am still on my GTX 980 since 2015. I looked at my local store (Canada Computers, College St, Toronto) & they offer an Asus GeForce GT1030 LP for CAD 120 : it runs at 1228 MHz & has 2 GB memory @ 3004 MHz with 64-bit i/face. My mobo is an Asus M5A97 AMD, which I bought 3 years ago . Does anyone see a problem using this card with Nvidia & my mobo ? -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Nvidia driver surprise
181028 Davyd McColl wrote: >> 181028 Philip Webb : Perhaps I should buy a more upto-date graphics card. > Or perhaps go team red. Personally, I've had enough of team green's drivers. > My 660ti is still working fine, but if it dies, I think I'm going red. > you can't be held hostage with binary blobs. And they're cheaper. Sorry, I don't understand your red/green code. >> Upgrade to the 1050 or so and you won’t have to upgrade for years. >> I am still on my GTX 980 since 2015. I looked at my local store (Canada Computers, College St, Toronto) & they offer an Asus GeForce GT1030 LP for CAD 120 : it runs at 1228 MHz & has 2 GB memory @ 3004 MHz with 64-bit i/face. My mobo is an Asus M5A97 AMD, which I bought 3 years ago . Does anyone see a problem using this card with Nvidia & my mobo ? -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Nvidia driver surprise
Or perhaps go team red. Personally, I've had enough of team green's drivers. My 660ti is still working fine, but if it dies, I think I'm going red. Not saying the grass is 100% greener there (heh) but at least you can't be held hostage with binary blobs. And they're cheaper. /2c -d On October 28, 2018 4:28:37 PM Andrew Udvare wrote: On Oct 28, 2018, at 09:53, Philip Webb wrote: Perhaps I should buy a more upto-date graphics card. Does anyone have comments or suggestions ? Upgrade to the 1050 or so and you won’t have to upgrade for years. I am still on my GTX 980 since 2015.
[gentoo-user] Re: Nvidia driver surprise
On 28/10/2018 15:53, Philip Webb wrote: I updated to xorg-server-1.20.3 yesterday & ran into trouble. It wanted to update to nvidia-drivers-396.54 , which doesn't support my graphics card ( Asus GT610 : bought 150914 ) : there was an error msg in Xorg.0.log to this effect. I tried 390.67 (which I had been using before yesterday) & updating to 390.87 , but got errors with the latter "failed to initialise kernel module" (Xorg) & "nvidia : unknown symbol backlight_device_register" (kernel) ; 390.67 won't emerge with xorg-server-1.20.3 . So I'm back using Nouveau, which seems to work adequately for most things, but won't run 3D screensavers. Perhaps I should buy a more upto-date graphics card. Does anyone have comments or suggestions ? The nvidia legacy GPU driver branch is a PITA to follow. You need to somehow magically know yourself which x.org versions they support and package.mask newer versions of xorg-server yourself. Nvidia does not seem to document the xorg versions they support. At least I haven't found the place they document it, if they do. Right now, I would try to mask >=x11-base/xorg-server-1.20, do a: emerge -auDN @world then re-emerge the legacy nvidia driver and see if that helps.
Re: [gentoo-user] Nvidia driver surprise
> On Oct 28, 2018, at 09:53, Philip Webb wrote: > > Perhaps I should buy a more upto-date graphics card. > > Does anyone have comments or suggestions ? Upgrade to the 1050 or so and you won’t have to upgrade for years. I am still on my GTX 980 since 2015.
[gentoo-user] Nvidia driver surprise
I updated to xorg-server-1.20.3 yesterday & ran into trouble. It wanted to update to nvidia-drivers-396.54 , which doesn't support my graphics card ( Asus GT610 : bought 150914 ) : there was an error msg in Xorg.0.log to this effect. I tried 390.67 (which I had been using before yesterday) & updating to 390.87 , but got errors with the latter "failed to initialise kernel module" (Xorg) & "nvidia : unknown symbol backlight_device_register" (kernel) ; 390.67 won't emerge with xorg-server-1.20.3 . So I'm back using Nouveau, which seems to work adequately for most things, but won't run 3D screensavers. Perhaps I should buy a more upto-date graphics card. Does anyone have comments or suggestions ? -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
[gentoo-user] Re: Why doesn't revdep-rebuild catch undefined symbol errors?
On 28/10/2018 12:36, Andrew Savchenko wrote: revdep-rebuild catches libraries from removed packages (including removed older versions) still in use by other packages. Though with proper subslot dependencies revdep-rebuild is rarely needed. If for some reason library containing required symbol was forcefully removed, revdep-rebuild cannot magically recreate that symbol. Ah, alright. I was thinking of this tool as a detector for things that need to be rebuild. So it seems the issue I ran across has no tool to detect it (even though it seems it's easy to detect.)
Re: [gentoo-user] File conflict
Hi, On Sun, 9 Sep 2018 19:46:38 +0200 tu...@posteo.de wrote: > Hi, > > I tried to install platformio and got a file conflict: > > dev-embedded/sunxi-tools > > installs /usr/bin/pio > > as dev-embedded/platformio-3.6.0 > > tries also. > > The installation of platformio was aborted due to > the file conflict. Report this to bugzilla and CC both package maintainers. Best regards, Andrew Savchenko pgpNE0rYBq_dx.pgp Description: PGP signature
Re: [gentoo-user] app-shells/bash slotted?
On Sat, 06 Oct 2018 03:36:36 -0400 John Covici wrote: > Hi. I noticed when preparing to do my world update today that > app-shells/bash is now slotted. I wonder what the purpose of this is, > there seems to be no way to select the different versions and > upgrading bash broke dracut -- unless they fixed it recently. > > Any thoughts? Different bash versions may pose different behaviour on the same code. Since bash is heavily used within ebuilds and eclasses, we may have to use bash slot different from user installed in future. The same need may arises for users with complicated bash scripts which may/will not work correctly with other bash versions than a specific one. Best regards, Andrew Savchenko pgpkthPThEttM.pgp Description: PGP signature
Re: [gentoo-user] Why doesn't revdep-rebuild catch undefined symbol errors?
On Thu, 25 Oct 2018 09:30:00 +0300 Nikos Chantziaras wrote: > Trying to rebuild media-sound/pavucontrol or any other package that uses > dev-cpp/cairomm (like pulseeffects) will fail with: > > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../x86_64-pc-linux-gnu/bin/ld: > > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../lib64/libcairomm-1.0.so: > undefined reference to `cairo_script_create' > > and a bunch of other "undefined reference" errors in libcairomm-1.0.so. > I don't know what broke cairomm. But I thought revdep-rebuild would > catch this. It doesn't. > > Now this is a rebuild of pavucontrol, so if I run: > >ldd -r /usr/bin/pavucontrol > > I will get: > >undefined symbol: cairo_script_create_for_stream > (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_set_mode (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_surface_create > (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_from_recording_surface > (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_create (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_surface_create_for_target > (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_get_mode (/usr/lib64/libcairomm-1.0.so.1) >undefined symbol: cairo_script_write_comment > (/usr/lib64/libcairomm-1.0.so.1) > > If I rebuild dev-cpp/cairomm manually (emerge -1 dev-cpp/cairomm), the > error goes away, and the affected packages can now be emerged successfully. > > Shouldn't revdep-rebuild catch this problem? Shouldn't it check > installed binaries for "undefined symbol" errors? revdep-rebuild catches libraries from removed packages (including removed older versions) still in use by other packages. Though with proper subslot dependencies revdep-rebuild is rarely needed. If for some reason library containing required symbol was forcefully removed, revdep-rebuild cannot magically recreate that symbol. Best regards, Andrew Savchenko pgpAJrxuaNyXT.pgp Description: PGP signature
Re: [gentoo-user] Inkscape failed to compile...(update from this morning)
On Sunday, 28 October 2018 09:05:59 GMT tu...@posteo.de wrote: > oh YEAH...sometimes the """problem""" seems to imply > a vast complicate solutionand one do not even > assume that there is something that straight forward > as > > emerge --deselect cairomm > > . > GREAT! :) > > Cheers! > Meino Glad to be of service. :) -- Regards, Peter.
Re: [gentoo-user] Inkscape failed to compile...(update from this morning)
On 10/28 08:49, Peter Humphrey wrote: > On Sunday, 28 October 2018 06:46:37 GMT tu...@posteo.de wrote: > > On 10/28 04:13, Adam Carter wrote: > > > On Sun, Oct 28, 2018 at 4:08 PM Adam Carter wrote: > > > > On Sun, Oct 28, 2018 at 3:41 PM wrote: > > > >> Hi, > > > > > > > >> while syncing/updateing my system, inkscape failed to compile: > > > > Nikos posted about cairo issue recently that revdep-rebuild didnt fix. I > > > > had a similar issue with different software. In my case; > > > > > > > > emerge -1 x11-libs/cairo dev-cpp/cairomm dev-python/pycairo > > > > > > > > Fixed it. Not sure which package was the problem. > > > > > > dev-cpp/cairomm seems like the most likely candidate as its files are > > > mentioned in Nikos' and your build errors, so try rebuilding that one > > > first. > > Hi, > > > > reemergeing cairomm fixes the problem! > > Many thanks for that solution! :) > > > > Unfortunately I emerged and forget the '-1'...is there a > > way to fix that? > > emerge --deselect cairomm. > > -- > Regards, > Peter. Hi Peter, oh YEAH...sometimes the """problem""" seems to imply a vast complicate solutionand one do not even assume that there is something that straight forward as emerge --deselect cairomm . GREAT! :) Cheers! Meino
Re: [gentoo-user] Inkscape failed to compile...(update from this morning)
On Sunday, 28 October 2018 06:46:37 GMT tu...@posteo.de wrote: > On 10/28 04:13, Adam Carter wrote: > > On Sun, Oct 28, 2018 at 4:08 PM Adam Carter wrote: > > > On Sun, Oct 28, 2018 at 3:41 PM wrote: > > >> Hi, > > > > > >> while syncing/updateing my system, inkscape failed to compile: > > > Nikos posted about cairo issue recently that revdep-rebuild didnt fix. I > > > had a similar issue with different software. In my case; > > > > > > emerge -1 x11-libs/cairo dev-cpp/cairomm dev-python/pycairo > > > > > > Fixed it. Not sure which package was the problem. > > > > dev-cpp/cairomm seems like the most likely candidate as its files are > > mentioned in Nikos' and your build errors, so try rebuilding that one > > first. > Hi, > > reemergeing cairomm fixes the problem! > Many thanks for that solution! :) > > Unfortunately I emerged and forget the '-1'...is there a > way to fix that? emerge --deselect cairomm. -- Regards, Peter.
Re: [gentoo-user] Inkscape failed to compile...(update from this morning)
On 10/28 08:05, David Haller wrote: > Hello, > > On Sun, 28 Oct 2018, tu...@posteo.de wrote: > [..] > >reemergeing cairomm fixes the problem! > > No wonder. The error was e.g.: > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../lib64/libcairomm-1.0.so: > undefined reference to `cairo_script_create' > > $ qfile > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../lib64/libcairomm-1.0.so > dev-cpp/cairomm (/usr/lib64/libcairomm-1.0.so) > > So re-emerging dev-cpp/cairomm should be obvious ;) > > >Unfortunately I emerged and forget the '-1'...is there a > >way to fix that? > > # sed -i.orig '/^dev-cpp/cairomm$/d' /var/lib/portage/world > > Check with > > # diff -u /var/lib/portage/world.orig /var/lib/portage/world > > that only that one entry is gone. Then you can remove the .orig. > > Or just edit that world file with the editor of your choice. > > HTH, > -dnh > > -- > I don't care about people reading in OE. It just gets bad if they > send postings, viruses, and complaints about non-existing attachements. > Hi dnh, thanks for your help! I know of the entry in the world file but was unsure, whether somewhere else were traces of the cairomm installation... Cheers! Meino
Re: [gentoo-user] Inkscape failed to compile...(update from this morning)
Hello, On Sun, 28 Oct 2018, tu...@posteo.de wrote: [..] >reemergeing cairomm fixes the problem! No wonder. The error was e.g.: /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../lib64/libcairomm-1.0.so: undefined reference to `cairo_script_create' $ qfile /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../lib64/libcairomm-1.0.so dev-cpp/cairomm (/usr/lib64/libcairomm-1.0.so) So re-emerging dev-cpp/cairomm should be obvious ;) >Unfortunately I emerged and forget the '-1'...is there a >way to fix that? # sed -i.orig '/^dev-cpp/cairomm$/d' /var/lib/portage/world Check with # diff -u /var/lib/portage/world.orig /var/lib/portage/world that only that one entry is gone. Then you can remove the .orig. Or just edit that world file with the editor of your choice. HTH, -dnh -- I don't care about people reading in OE. It just gets bad if they send postings, viruses, and complaints about non-existing attachements.
Re: [gentoo-user] Inkscape failed to compile...(update from this morning)
On 10/28 04:13, Adam Carter wrote: > On Sun, Oct 28, 2018 at 4:08 PM Adam Carter wrote: > > > On Sun, Oct 28, 2018 at 3:41 PM wrote: > > > >> > >> Hi, > >> > >> while syncing/updateing my system, inkscape failed to compile: > >> > > > > Nikos posted about cairo issue recently that revdep-rebuild didnt fix. I > > had a similar issue with different software. In my case; > > > > emerge -1 x11-libs/cairo dev-cpp/cairomm dev-python/pycairo > > > > Fixed it. Not sure which package was the problem. > > > > dev-cpp/cairomm seems like the most likely candidate as its files are > mentioned in Nikos' and your build errors, so try rebuilding that one first. Hi, reemergeing cairomm fixes the problem! Many thanks for that solution! :) Unfortunately I emerged and forget the '-1'...is there a way to fix that? Cheers! Meino