Re: [gentoo-user] Unknown gcc-version needed by nvidia-cuda-toolkit?

[Manuel Mommertz]

Am Sonntag, 8. April 2018, 05:46:09 CEST schrieb tu...@posteo.de:
> ...I think, that gcc-7.3.0 matches   But what is the 'c' in '[cxx]' for?

If the cxx useflag is active, gcc will have support to compile c++

Greets
Manuel

[gentoo-user] Unknown gcc-version needed by nvidia-cuda-toolkit?

[tuxic]

Hi,

while doing the update of my system I got this:

Calculating dependencies... done!
 * Dependencies could not be completely resolved due to
 * the following required packages not being installed:
 * 
 *   https://gcc.gnu.org/
 Description: The GNU Compiler Collection

[I] sys-devel/gcc-config
 Available versions:  1.8-r1 (~)1.9.0-r1 (~)1.9.1 **
 Installed versions:  1.9.1(02:25:19 AM 12/17/2017)
 Homepage:https://gitweb.gentoo.org/proj/gcc-config.git/
 Description: Utility to manage compilers

Found 2 matches


...I think, that gcc-7.3.0 matches  

Re: [gentoo-user] Bootloader hangs without a keyboard

[mad.scientist.at.large]

The obvious solution is to leave a keyboard attached, OR make the computer 
think so with one of the adapters that let you use ps/2 peripherals on a 
machine that only has usb.  They always show up as a keyboard and mouse to the 
computer, even if only the adapter is plugged in.  They have them on ebay etc., 
from $1 up.
mad.scientist.at.large (a good madscientist)
--
God bless the rich, the greedy and the corrupt politicians they have put into 
office.   God bless them for helping me do the right thing by giving the rich 
my little pile of cash.  After all, the rich know what to do with money.


7. Apr 2018 17:18 by mar...@gmx.de :


> Hi list
>
> I have a bit of a weird issue.  I recently acquired a used server (Fujitsu 
> Primergy TX140 S1) via a friend of mine.  I have Gentoo installed on it now 
> and overall it works fine save for one perplexing issue [0]: The bootloader 
> hangs when booting without a keyboard attached.  I tried it with both GRUB 
> and 
> systemd-boot (which works nicely in combination with kernel-install); both 
> only work with a keyboard attached.  GRUB hangs when the countdown starts, 
> while systemd-boot hangs when the cursor first appears, but before the boot 
> entries are listed.  In both cases the cursor freezes.
>
> I'm curious if anybody has seen this issue before and/or has any idea how to 
> solve it.  The only resources I could find online that seem related are [1] 
> (different OS, no resolution, hints at BIOS), [2] (which provides a potential 
> solution, though I'm not sure whether I can actually implement it), and [3] 
> (the patch didn't help).  There was also a related systemd bug at [4].
>
> I updated the BIOS, but that helped as much as anticipated, that is, not at 
> all (it was at the penultimate version already).  I also tried unplugging the 
> display in case the BIOS was expecting a keyboard then.
>
> Any help is much appreciated.
>
> [0] Well, there's a second one, too, but I'll have to see if it shows up 
> again.
> [1] > 
> https://unix.stackexchange.com/questions/66696/nas4free-freebsd-wont-make-it-past-bootloader-without-keyboard
>  
> 
> [2] > 
> https://www.linuxquestions.org/questions/linux-newbie-8/os-cannot-be-loaded-without-keyboard-because-of-grub-805350
>  
> 
> [3] > https://bugs.archlinux.org/task/44923 
> 
> [4] > https://github.com/systemd/systemd/pull/3735 
> 
>
> Greetings
> -- 
> Marc Joliet
> --
> "People who think they know everything really annoy those of us who know we
> don't" - Bjarne Stroustrup

[gentoo-user] Bootloader hangs without a keyboard

[Marc Joliet]

Hi list

I have a bit of a weird issue.  I recently acquired a used server (Fujitsu 
Primergy TX140 S1) via a friend of mine.  I have Gentoo installed on it now 
and overall it works fine save for one perplexing issue [0]: The bootloader 
hangs when booting without a keyboard attached.  I tried it with both GRUB and 
systemd-boot (which works nicely in combination with kernel-install); both 
only work with a keyboard attached.  GRUB hangs when the countdown starts, 
while systemd-boot hangs when the cursor first appears, but before the boot 
entries are listed.  In both cases the cursor freezes.

I'm curious if anybody has seen this issue before and/or has any idea how to 
solve it.  The only resources I could find online that seem related are [1] 
(different OS, no resolution, hints at BIOS), [2] (which provides a potential 
solution, though I'm not sure whether I can actually implement it), and [3] 
(the patch didn't help).  There was also a related systemd bug at [4].

I updated the BIOS, but that helped as much as anticipated, that is, not at 
all (it was at the penultimate version already).  I also tried unplugging the 
display in case the BIOS was expecting a keyboard then.

Any help is much appreciated.

[0] Well, there's a second one, too, but I'll have to see if it shows up 
again.
[1] 
https://unix.stackexchange.com/questions/66696/nas4free-freebsd-wont-make-it-past-bootloader-without-keyboard
[2] 
https://www.linuxquestions.org/questions/linux-newbie-8/os-cannot-be-loaded-without-keyboard-because-of-grub-805350/
[3] https://bugs.archlinux.org/task/44923
[4] https://github.com/systemd/systemd/pull/3735

Greetings
-- 
Marc Joliet
--
"People who think they know everything really annoy those of us who know we
don't" - Bjarne Stroustrup


signature.asc
Description: This is a digitally signed message part.

[gentoo-user] [OT?] Cheap Chinese Pen Tablet (UC-Logic Derivative) Problems

[R0b0t1]

Hello list,

I can plug in a Huion H420 I recently purchased. It registers, but the
lower portion of the screen (~150px) is not able to be navigated to by
the pen. I've tried looking for bugs (some are mentioned on Ubuntu
forums) but can't figure out if any are applicable to my device or
even what they are. Enabling or disabling the proper driver doesn't
seem to change the operation of the pen tablet.

Perhaps more problematic, the tablet registers as a mouse. I don't
want it to be a mouse. Shouldn't it register as a unique HID class?
Can anyone explain how pen input works on Linux? On Windows, the pen
can do mouse actions, but it is unique and separate.

Cheers,
 R0b0t1

Re: [gentoo-user] Mutt database option

[Floyd Anderson]

On Sat, 07 Apr 2018 14:42:49 +0100
Mick  wrote:

On Saturday, 7 April 2018 14:35:27 BST Floyd Anderson wrote:

Hi Mick,

On Sat, 07 Apr 2018 11:21:23 +0100

Mick  wrote:
>So far I had been using gdbm, but I now see that emerge also added lmdb.

Same here, so I gave lmdb a try as hcache backend.

>Which one is best to use? What have you chosen?

I assume you mean for speed? I don’t know and it may become very
academic to answer this. But you can find some none Mutt-specific
benchmark results on NeoMutt’s website [1].

Note, the mentioned benchmark page say:

“[…] you’ll need a reasonable large number of
messages – >50k – to see anything interesting”

Using lmdb as backend, I do not realise any differences over gdbm within
Mutt respectively NeoMutt and I doubt one really can (without measuring
it exactly – which I haven’t done yet).


References:
  [1] 


Thanks Floyd, good information.

I also switched to lmdb now and updated my use flags accordingly for mutt.  I
see neomutt gaining traction, but I am still running mutt here.  Is there a
benefit from switching?


I think yes but I’m also using both here. Mutt for testing different 
behaviour (sometimes issues) of NeoMutt and NeoMutt as as my day-to-day 
mail client workhorse.


The main reason for my switch to NeoMutt was that I’ve had no luck with 
colourisation in Mutt (nearly two years ago). TBH, afterwards I realised 
that the problem was sitting in front of the screen and used 
sys-libs/slang instead of sys-libs/ncurses), so no reason for a switch.


But I like the faster development/release cycle, the goal to clean up 
the 20 years old code base and some features [1] of NeoMutt, e.g. 
Lua-scripting, sidebar; which sometimes find their way into Mutt. I have 
no experiences contributing patches to Mutt but regarded to NeoMutt, it 
meet my expectations and that is fun.


Back to the topic, with a another example. As far as I can tell you 
cannot change the hcache backend without recompiling Mutt where NeoMutt 
implements the ‘$header_cache_backend’ configuration variable for.


But in the end I can only say forget all things above, I’m only more 
familiar with NeoMutt than with Mutt.



References:
 [1] 



--
Regards,
floyd

Re: [gentoo-user] Mutt database option

[Mick]

On Saturday, 7 April 2018 14:35:27 BST Floyd Anderson wrote:
> Hi Mick,
> 
> On Sat, 07 Apr 2018 11:21:23 +0100
> 
> Mick  wrote:
> >So far I had been using gdbm, but I now see that emerge also added lmdb.
> 
> Same here, so I gave lmdb a try as hcache backend.
> 
> >Which one is best to use? What have you chosen?
> 
> I assume you mean for speed? I don’t know and it may become very
> academic to answer this. But you can find some none Mutt-specific
> benchmark results on NeoMutt’s website [1].
> 
> Note, the mentioned benchmark page say:
> 
> “[…] you’ll need a reasonable large number of
> messages – >50k – to see anything interesting”
> 
> Using lmdb as backend, I do not realise any differences over gdbm within
> Mutt respectively NeoMutt and I doubt one really can (without measuring
> it exactly – which I haven’t done yet).
> 
> 
> References:
>   [1] 

Thanks Floyd, good information.

I also switched to lmdb now and updated my use flags accordingly for mutt.  I 
see neomutt gaining traction, but I am still running mutt here.  Is there a 
benefit from switching?

-- 
Regards,
Mick

signature.asc
Description: This is a digitally signed message part.

Re: [gentoo-user] Mutt database option

[Floyd Anderson]

Hi Mick,

On Sat, 07 Apr 2018 11:21:23 +0100
Mick  wrote:

So far I had been using gdbm, but I now see that emerge also added lmdb.


Same here, so I gave lmdb a try as hcache backend.


Which one is best to use? What have you chosen?


I assume you mean for speed? I don’t know and it may become very 
academic to answer this. But you can find some none Mutt-specific 
benchmark results on NeoMutt’s website [1].


Note, the mentioned benchmark page say:

   “[…] you’ll need a reasonable large number of
   messages – >50k – to see anything interesting”

Using lmdb as backend, I do not realise any differences over gdbm within 
Mutt respectively NeoMutt and I doubt one really can (without measuring 
it exactly – which I haven’t done yet).



References:
 [1] 


--
Regards,
floyd

Re: [gentoo-user] [OT] What is the best open-source VPN server for Linux?

[R0b0t1]

On Saturday, April 7, 2018, Mick  wrote:
> On Friday, 6 April 2018 18:55:18 BST gevisz wrote:
>> 2018-04-06 2:10 GMT+03:00 Grant Taylor :
>
>> > I'd encourage your friend to check out the VPN capabilities built into
>> > Windows.  He may need to install / configure (R)RAS to enable the
>> > features.
>>
>> Thank you for your advice. He is currently trying to set up RAS with SSTP
>> but RAS client so far cannot log into the server, while a third party VPN
>> just works (until the remote computer hangs for so far unknown reason
that
>> even may not be connected with the VPN server).
>>
>> We will continue to experiment to find the reason.
>
> Typical problems incurred with SSTP are relating to username
authentication
> and TLS certificate selection/configuration.
>
> SSTP authenticates OS users, not devices/PCs.  So use the *same* username
and
> passwd on all the OS login, SSTP VPN & RRAS wizards.
>
> The TLS server certificate has to contain a DN which will resolve to the
IP of
> the server in question, or better use the IP address both in the CN and
the
> X509v3 Subject Alternative Name fields.
>
> In addition, the SSTP certificate binding has to use the same TLS
certificate
> with that selected for RRAS and this is not always obvious (for SSTP at
> least).  You can use MSWindow's 'netsh ras show sstp-ssl-cert' command to
show
> the TLS certificate in use by SSTP and compare this with the RRAS
certificate
> selection.
>
> It is a bit of a faff, but that's what you get with SSTP.  The benefit of
it
> is that it is integrated with MSWindows authentication mechanisms and
network
> stack, allowing easy enterprise wide configuration and management.  For
your
> friend's one off VPN set up, OpenVPN, or SoftEther VPN is probably a
better
> MSWindows based option:
>

Companies which need user management tend to just set up an intranet and
provide VPN access to it which is likely not going to be a Microsoft
technology. There is no benefit to integrating OS authentication with your
transport security. If you contacted a Windows-focused business for your
administration they may set such a system up, but only because they don't
know any better.

Evaluating Microsoft software should be done extremely carefully. It is
very easy to waste time, ignoring other concerns. You may get something
working but it will not be easy to administrate or scale.

Microsoft's current revenue may be largely from customers using the sunk
cost fallacy.

Cheers,
R0b0t1

Re: [gentoo-user] [OT] What is the best open-source VPN server for Linux?

[Mick]

On Friday, 6 April 2018 18:55:18 BST gevisz wrote:
> 2018-04-06 2:10 GMT+03:00 Grant Taylor :

> > I'd encourage your friend to check out the VPN capabilities built into
> > Windows.  He may need to install / configure (R)RAS to enable the
> > features.
> 
> Thank you for your advice. He is currently trying to set up RAS with SSTP
> but RAS client so far cannot log into the server, while a third party VPN
> just works (until the remote computer hangs for so far unknown reason that
> even may not be connected with the VPN server).
> 
> We will continue to experiment to find the reason.

Typical problems incurred with SSTP are relating to username authentication 
and TLS certificate selection/configuration.

SSTP authenticates OS users, not devices/PCs.  So use the *same* username and 
passwd on all the OS login, SSTP VPN & RRAS wizards.

The TLS server certificate has to contain a DN which will resolve to the IP of 
the server in question, or better use the IP address both in the CN and the 
X509v3 Subject Alternative Name fields.

In addition, the SSTP certificate binding has to use the same TLS certificate 
with that selected for RRAS and this is not always obvious (for SSTP at 
least).  You can use MSWindow's 'netsh ras show sstp-ssl-cert' command to show 
the TLS certificate in use by SSTP and compare this with the RRAS certificate 
selection. 

It is a bit of a faff, but that's what you get with SSTP.  The benefit of it 
is that it is integrated with MSWindows authentication mechanisms and network 
stack, allowing easy enterprise wide configuration and management.  For your 
friend's one off VPN set up, OpenVPN, or SoftEther VPN is probably a better 
MSWindows based option:

http://www.softether.org/
https://github.com/SoftEtherVPN/SoftEtherVPN

-- 
Regards,
Mick

signature.asc
Description: This is a digitally signed message part.

[gentoo-user] Mutt database option

[Mick]

Hi All,

A mutt update today complained about which database to use for hcache:

!!! Problem resolving dependencies for mail-client/mutt from @selected  
.. done!

!!! The ebuild selected to satisfy "mail-client/mutt" has unmet requirements.
- mail-client/mutt-1.9.4-r1::gentoo USE="crypt gdbm gpg hcache imap lmdb mbox 
nls nntp pop sasl smime smtp ssl -berkdb -debug -doc -gnutls -gpgme -idn -
kerberos -libressl -notmuch -pgp_classic (-prefix) -qdbm (-selinux) -slang -
smime_classic -tokyocabinet -vanilla" ABI_X86="(64)"

  The following REQUIRED_USE flag constraints are unsatisfied:
hcache? ( exactly-one-of ( berkdb gdbm lmdb qdbm tokyocabinet ) )
  
  The above constraints are a subset of the following complete expression:
hcache? ( exactly-one-of ( berkdb gdbm lmdb qdbm tokyocabinet ) ) imap? ( 
ssl ) pop? ( ssl ) nntp? ( ssl ) smime? ( ssl !gnutls ) smime_classic? ( ssl !
gnutls ) smtp? ( ssl ) sasl? ( any-of ( imap pop smtp nntp ) ) kerberos? ( 
any-of ( imap pop smtp nntp ) )

So far I had been using gdbm, but I now see that emerge also added lmdb.  
Which one is best to use?  What have you chosen?

-- 
Regards,
Mick

signature.asc
Description: This is a digitally signed message part.