Re: [gentoo-user] no more googleearth in portage
Raffaele Belardi wrote: > I suppose it's due to Google's choice to support only Chrome, although I > missed the Gentoo > news bit if there was one. > > For Android there is the really good Open Street Map application, are there > any desktop > alternatives in Portage for non-Chrome users? I know OSM has a web interface > but I'd > prefer a standalone application. > > raffaele > > If it helps any, I still have it installed and I rescued the ebuild folder. I can send it to you if you want it, or anyone else, off list. Let me know if you or anyone else is interested. It's version 7.1.8.3036 but it works for me here. I wasn't even aware that it had been removed. I to like to look at things from a above viewpoint and I don't have a phone or anything that can use googleearth. I'm glad you posted because I wouldn't have noticed otherwise. Just curious, could this be in a overlay somewhere? Dale :-) :-)
Re: [gentoo-user] no more googleearth in portage
On Wed, 20 Sep 2017 01:59:14 -0500, Dale wrote:
> If it helps any, I still have it installed and I rescued the ebuild
> folder. I can send it to you if you want it, or anyone else, off list.
> Let me know if you or anyone else is interested. It's version
> 7.1.8.3036 but it works for me here.
>
> I wasn't even aware that it had been removed. I to like to look at
> things from a above viewpoint and I don't have a phone or anything that
> can use googleearth. I'm glad you posted because I wouldn't have
> noticed otherwise.
>
> Just curious, could this be in a overlay somewhere?
Of course...
% eix -R googleearth
* sci-geosciences/googleearth
Available versions: 5.1.3535.3218-r1^ms[2] (~)5.2.1.1329_beta^ms[2]
(~)6.0.1.2032_beta^ms[2] (~)7.1.4.1529^m[1] (~)7.1.8.3036^m[3] {+bundled-libs
(+)qt-bundled}
Homepage: https://earth.google.com/
Description: A 3D interface to the planet
[1] "bircoph" layman/bircoph
[2] "multilib-portage" layman/multilib-portage
[3] "voyageur" layman/voyageur
--
Neil Bothwick
WinErr 01C: Uncertainty error - Uncertainty may be inadequate.
pgpAMv5uC3GYK.pgp
Description: OpenPGP digital signature
[gentoo-user] Sendmail confused by network change
Dear Gentoo Users, On my new Gentoo laptop installation I recently installed Sendmail in order to receive messages from Cron on the root account. I noticed that when I connect my laptop to a different network than the one I connected to during booting, Sendmail does not know what to do with the Cron mail any more. For the purpose of clarity, let’s say the host name of this laptop is ‘hostname’. I did not configure the domain part of the host name because of the mobile nature of this machine. When I boot at home, Cron sends mail to root@hostname.homedomain. ‘homedomain’ is automatically added to all host names on my home network by the router. It can only be resolved inside the network; it is not a registered domain name. I can receive mail from Cron just fine. When I boot at work, Cron sends mail to root@hostname. Note that the domain name ‘workdomain’ is not added to the host name. I can still receive Cron mail. However, when I take the laptop home without rebooting and connect to the home network, Sendmail is unable to deliver the Cron mail for root@hostname.homedomain and sends notifications of this to root@hostname.homedomain, which somehow do seem to arrive without problems. The error message is “config error: mail loops back to me”. Based on what I can find about this error on the internet, it looks like Sendmail does not know where hostname.homedomain is and asks my router to resolve that. When it finds out it is localhost, it thinks something is wrong and does not deliver the mail. A possible solution is to register hostname.homedomain as an alias of hostname or localhost, but I would rather not do that, since hard coding domain names on a laptop seems kludgy to me. Does anyone know a more elegant solution? Some way to inform Sendmail about changes to the domain name, or configure it to check for these changes? I would rather not have to reboot. Restarting Sendmail is acceptable, I guess, but perhaps not the most elegant solution. Sincerely, Bas -- Sebastiaan L. Zoutendijk | slzoutend...@gmail.com
Re: [gentoo-user] Sendmail confused by network change
On Wed, 20 Sep 2017 05:45:07 -0400, Bas Zoutendijk wrote: > > Dear Gentoo Users, > > On my new Gentoo laptop installation I recently installed Sendmail in > order to receive messages from Cron on the root account. I noticed that > when I connect my laptop to a different network than the one I connected > to during booting, Sendmail does not know what to do with the Cron mail > any more. > > For the purpose of clarity, let’s say the host name of this laptop is > ‘hostname’. I did not configure the domain part of the host name > because of the mobile nature of this machine. > > When I boot at home, Cron sends mail to root@hostname.homedomain. > ‘homedomain’ is automatically added to all host names on my home network > by the router. It can only be resolved inside the network; it is not a > registered domain name. I can receive mail from Cron just fine. > > When I boot at work, Cron sends mail to root@hostname. Note that the > domain name ‘workdomain’ is not added to the host name. I can still > receive Cron mail. However, when I take the laptop home without > rebooting and connect to the home network, Sendmail is unable to deliver > the Cron mail for root@hostname.homedomain and sends notifications of > this to root@hostname.homedomain, which somehow do seem to arrive > without problems. The error message is “config error: mail loops back > to me”. > > Based on what I can find about this error on the internet, it looks like > Sendmail does not know where hostname.homedomain is and asks my router > to resolve that. When it finds out it is localhost, it thinks something > is wrong and does not deliver the mail. A possible solution is to > register hostname.homedomain as an alias of hostname or localhost, but I > would rather not do that, since hard coding domain names on a laptop > seems kludgy to me. > > Does anyone know a more elegant solution? Some way to inform Sendmail > about changes to the domain name, or configure it to check for these > changes? I would rather not have to reboot. Restarting Sendmail is > acceptable, I guess, but perhaps not the most elegant solution. Restarting sendmail seems fine with me, if you want to have something that works everywhere, why not get a domain name from ddns or somewhere and use a full fqdn all the time -- you can put your home machine on another host in that domain and you will be good to go. -- Your life is like a penny. You're going to lose it. The question is: How do you spend it? John Covici cov...@ccs.covici.com
Re: [gentoo-user] no more googleearth in portage
On 09/20/2017 01:55 AM, Raffaele Belardi wrote: > I suppose it's due to Google's choice to support only Chrome https://bugs.gentoo.org/show_bug.cgi?id=624966
[gentoo-user] Re: Sendmail confused by network change
On 2017-09-20 11:45, Bas Zoutendijk wrote: > When I boot at home, Cron sends mail to root@hostname.homedomain. > ‘homedomain’ is automatically added to all host names on my home > network by the router. It can only be resolved inside the network; it > is not a registered domain name. I can receive mail from Cron just > fine. > > When I boot at work, Cron sends mail to root@hostname. Note that the > domain name ‘workdomain’ is not added to the host name. I can still > receive Cron mail. However, when I take the laptop home without > rebooting and connect to the home network, Sendmail is unable to > deliver the Cron mail for root@hostname.homedomain and sends > notifications of this to root@hostname.homedomain, which somehow do > seem to arrive without problems. The error message is “config error: > mail loops back to me”. According to crontab(5), you can configure where the mail is sent, by setting the MAILTO variable in the crontab file. [This is for cronie, I am not sure if other cron variants do this.] So, you could tell it to send to ; presumably this would solve the problem. Despite being one of the retro/traditionalist guys on this list, I have to say installing sendmail just for this purpose doesn't seem proportionate. Even if you don't want to look into the smallish MTA packages (nullmailer etc.), you can make /usr/sbin/sendmail to be a pretty trivial script (or a link to one) that just locks the destination mailbox and appends the input to it. Heck, procmail could do that. -- Please don't Cc: me privately on mailing lists and Usenet, if you also post the followup to the list or newsgroup. Do obvious transformation on domain to reply privately _only_ on Usenet.
Re: [gentoo-user] Change Readline mode for all users?
> On 19 Sep 2017, at 01:30, Simon Thelen wrote: > >> I find that my pager and editor are set in /etc/env.d/99pager and >> /etc/env.d/99editor respectively, but creating a >> /etc/env.d/99bashlineediting file containing "set -o vi" doesn't seem >> to work. > Either "set-editing-mode vi" in /etc/inputrc for all readline programs > or in /etc/bash/bashrc (for bash-only) Oops! This spelling breaks the "m" key - for some reason it doesn't work in lowercase mode. The correct comment directive is "set editing-mode vi" with a space after "set". This is also discussed on the Vim wiki: http://vim.wikia.com/wiki/Use_vi_shortcuts_in_terminal Stroller.
Re: [gentoo-user] Sendmail confused by network change
Dear John, On Wed 20 Sep 2017 at 07:05:11 -0400, John Covici wrote: > Restarting sendmail seems fine with me, if you want to have something > that works everywhere, why not get a domain name from ddns or > somewhere and use a full fqdn all the time -- you can put your home > machine on another host in that domain and you will be good to go. Thank you for this suggestion. I added ‘.localdomain’ to my host name and now Sendmail uses this domain name on all networks. Cron mail arrives again at home. I did not want to try a ‘real’ domain name, as this could make the system accessible from outside the local network, if I understand correctly. Right now it is hidden behind the router and only has a local IP address. I would like to keep it that way, because then I do not have to worry about network security settings. Nevertheless you led me to a solution. Thank you, Bas -- Sebastiaan L. Zoutendijk | slzoutend...@gmail.com
Re: [gentoo-user] Re: Sendmail confused by network change
Dear Ian, On Wed 20 Sep 2017 at 08:23:33 -0700, Ian Zimmerman wrote: > Despite being one of the retro/traditionalist guys on this list, I have > to say installing sendmail just for this purpose doesn't seem > proportionate. Even if you don't want to look into the smallish MTA > packages (nullmailer etc.), you can make /usr/sbin/sendmail to be a > pretty trivial script (or a link to one) that just locks the destination > mailbox and appends the input to it. Heck, procmail could do that. I appreciate your opinion. I normally seek to avoid unnecessary complication and find the right tool for the job. Since this is my first Gentoo install, this is the first time I had to pick an MTA by myself. Not knowing much about MTAs I though Sendmail was a common/traditional choice and would therefore probably serve my need. I will check out nullmailer and other MTAs to see if they better fit my need without adding extra complexity. As I wrote John I was able to resolve the issue so I can take my time now. Thank you, Bas -- Sebastiaan L. Zoutendijk | slzoutend...@gmail.com
Re: [gentoo-user] Sendmail confused by network change
On Wednesday, September 20, 2017 6:50:54 PM CEST Bas Zoutendijk wrote: > Dear John, > > On Wed 20 Sep 2017 at 07:05:11 -0400, John Covici wrote: > > Restarting sendmail seems fine with me, if you want to have something > > that works everywhere, why not get a domain name from ddns or > > somewhere and use a full fqdn all the time -- you can put your home > > machine on another host in that domain and you will be good to go. > > Thank you for this suggestion. I added ‘.localdomain’ to my host name > and now Sendmail uses this domain name on all networks. Cron mail > arrives again at home. Glad it works now. > I did not want to try a ‘real’ domain name, as this could make the > system accessible from outside the local network, if I understand > correctly. To say it bluntly, you don't :) > Right now it is hidden behind the router and only has a > local IP address. I would like to keep it that way, because then I do > not have to worry about network security settings. Nevertheless you led > me to a solution. This is how it works. To clarify: The IP-address determines how it is accessible. If the IP is not accessible from the rest of the world, then you don't have to worry too much about securing your mailserver. The domain name is only an address-book entry. Unless it is fully entered into the public one, noone will see the IP you are using. And as mentioned above, even if they can, the IP is sitting behind a router and is most likely in a private range. -- Joost
Re: [gentoo-user] Sendmail confused by network change
On Wed 20 Sep 2017 at 19:02:52 +0200, J. Roeleveld wrote: > To clarify: > The IP-address determines how it is accessible. If the IP is not accessible > from the rest of the world, then you don't have to worry too much about > securing your mailserver. > > The domain name is only an address-book entry. Unless it is fully entered > into > the public one, noone will see the IP you are using. And as mentioned above, > even if they can, the IP is sitting behind a router and is most likely in a > private range. Good to know, thanks. Sincerely, Bas -- Sebastiaan L. Zoutendijk | slzoutend...@gmail.com
Re: [gentoo-user] scary (?) situation with binutils
On Mon, Sep 18 2017, allan gottlieb wrote: > On Mon, Sep 18 2017, Andreas K. Huettel wrote: > >> No need to do anything complicated. Just run >> >> emerge --depclean --ask >> >> and have it remove outdated binutils versions (it will keep the current one, >> and also then select it for usage automatically). >> >> Then all these problems are gone... > > Thank you! > allan For others who may have encountered the infinite-emerge-binutils problem, let me report that (as I am sure andreas knows) the depclean fixes that problem while purging the masked versions of binutils. Would that the depclean likely have fixed the infinite-emerge-binutils problem several months ago when it first appeared. I don't remember seeing it proposed. Once again, thank you! allan
Re: [gentoo-user] no more googleearth in portage
On 09/19/2017 10:55 PM, Raffaele Belardi wrote: > I suppose it's due to Google's choice to support only Chrome, although I > missed the Gentoo > news bit if there was one. > > For Android there is the really good Open Street Map application, are there > any desktop > alternatives in Portage for non-Chrome users? I know OSM has a web interface > but I'd > prefer a standalone application. Gentoo has never handled new versions of googleearth well. They gave up this time. The alt version from the voyageur overlay (found from zugainia.org site) worked with the new 7.3.0.3830 google earth with slight modifications.
[gentoo-user] Multiple network interfaces and openrc
When I add multiple net.* services to a single runlevel (basic example: both a net.en* and a net.wl* service in default runlevel), it has a surprising and undesirable effect: when I bring one of them down by stopping the service, dnsmasq also gets stopped. It is as if openrc thinks dnsmasq "depends" in the openrc sense on _both_ the net.* services, instead of just one. Is there a way to hack the dependency machinery to teach openrc the right way? -- Please don't Cc: me privately on mailing lists and Usenet, if you also post the followup to the list or newsgroup. Do obvious transformation on domain to reply privately _only_ on Usenet.
Re: [gentoo-user] no more googleearth in portage
On Wednesday, September 20, 2017, Raffaele Belardi wrote: > I suppose it's due to Google's choice to support only Chrome, although I missed the Gentoo > news bit if there was one. > I'm not sure that is right, see here: https://www.google.com/earth/ download/gep/agree.html. It looks like it is still "supported," but development may have slowed. > For Android there is the really good Open Street Map application, are there any desktop > alternatives in Portage for non-Chrome users? I know OSM has a web interface but I'd > prefer a standalone application. > Take a look at QGIS.[1] There is also NASA's WorldWind.[2] GIS is the keyword you should be using. It looks like the competitive non-Google Earth software projects are mainly targeted at developers. However, it most places I see that still talk about Google Earth tend to use it in a "developer" role (typically importing GIS data, not software development). The best forum topic I could find[3] states (as does the previously referenced bug) Google Earth was removed to sporadic crashes. This may not have been fixed. Regardless, I would invite you to not be afraid of installing packages into your $HOME, e.g. $HOME/.local or another non-root location. This is typically the first step before creating an ebuild. Some users seem to install Google Earth with overlays. Overlays should be evaluated with respect to their maintainers, but in general I have only seen them cause problems as they will invariably fall out of date. For sufficiently popular software packages this might not be an issue. Cheers, R0b0t1 [1]: http://www.qgis.org/en/site/ [2]: https://worldwind.arc.nasa.gov/ [3]: https://forums.gentoo.org/viewtopic-t-980876.html) P.S.: If you can find the time, please tell whomever is responsible that the STM32 HAL libraries look like they were designed by poorly supervised interns.
[gentoo-user] Re: Multiple network interfaces and openrc
On 2017-09-20, Ian Zimmerman wrote: > When I add multiple net.* services to a single runlevel (basic example: > both a net.en* and a net.wl* service in default runlevel), it has a > surprising and undesirable effect: when I bring one of them down by > stopping the service, dnsmasq also gets stopped. > > It is as if openrc thinks dnsmasq "depends" in the openrc sense on > _both_ the net.* services, instead of just one. Is there a way to hack > the dependency machinery to teach openrc the right way? >From my /etc/rc.conf: # Do we allow any started service in the runlevel to satisfy the dependency # or do we want all of them regardless of state? For example, if net.eth0 # and net.eth1 are in the default runlevel then with rc_depend_strict="NO" # both will be started, but services that depend on 'net' will work if either # one comes up. With rc_depend_strict="YES" we would require them both to # come up. rc_depend_strict="NO" -- Grant Edwards grant.b.edwardsYow! for ARTIFICIAL at FLAVORING!! gmail.com
Re: [gentoo-user] Re: Multiple network interfaces and openrc
On 09/20/2017 04:28 PM, Grant Edwards wrote: > On 2017-09-20, Ian Zimmerman wrote: > >> When I add multiple net.* services to a single runlevel (basic example: >> both a net.en* and a net.wl* service in default runlevel), it has a >> surprising and undesirable effect: when I bring one of them down by >> stopping the service, dnsmasq also gets stopped. >> >> It is as if openrc thinks dnsmasq "depends" in the openrc sense on >> _both_ the net.* services, instead of just one. Is there a way to hack >> the dependency machinery to teach openrc the right way? > > From my /etc/rc.conf: > > # Do we allow any started service in the runlevel to satisfy the dependency > # or do we want all of them regardless of state? For example, if net.eth0 > # and net.eth1 are in the default runlevel then with rc_depend_strict="NO" > # both will be started, but services that depend on 'net' will work if either > # one comes up. With rc_depend_strict="YES" we would require them both to > # come up. > rc_depend_strict="NO" > This might also be considered a bug in dnsmasq's init script: https://bugs.gentoo.org/show_bug.cgi?id=need-net Most things that "need net" don't -- they just need any old interface to be up. In particular, they might work just fine with only the loopback interface up, and so what they really want is "net.lo". However (and a little ironically), "need net" is not satisfied by the loopback interface, so "need net" almost always waits longer than it has to because most daemons have default configuration that use either 0.0.0.0 or 127.0.0.1. I wouldn't swear to it, but I bet as a forwarding DNS server, dnsmasq can start just fine with only the loopback interface up. (You still won't be able to make DNS queries until the WAN comes up, but that's no reason to delay the start-up of dnsmasq with a "need" dependency). You might want to ask around, but I think the best solution for those services is to drop "need net" and add rc_need="net.lo" to their conf.d files. That way, you can place a comment right next to it that says # if you listen on a particular interface, specify it here rc_need="net.lo" allowing you to preempt the situation where a user specifies one interface in the config file but "need net" is satisfied by the wrong one.
