Re: Drafting a Guix blog post on the FHS container
Jim Newsome writes: > Sorry for (presumably) breaking threading; I came across this online > and don't see a way to set my in-reply-to-email header properly. > > Anyways just thought I'd mention that I recently learned about this > feature, and was able to use it to get a downloaded [Tor Browser > Bundle] running with: > > > ``` > guix shell \ > --container \ > --network \ > --emulate-fhs \ > --preserve='^DISPLAY$' > --share=/run/user/$(id -u)/gdm \ > openssl@1 \ > libevent \ > pciutils \ > dbus-glib \ > bash \ > libgccjit \ > libcxx \ > gtk+ \ > coreutils \ > grep \ > sed \ > file \ > alsa-lib \ > -- \ > ./start-tor-browser.desktop -v > ``` > > `--preserve='^DISPLAY$'` and `--share=/run/user/$(id -u)/gdm` are to > get access to the display. I'm not sure the second parameter is > universally correct; I reverse-engineered it via roughly `ps aux | > grep -- -auth`. > > The `-v` parameter to the browser script keeps it from trying to > background itself, which otherwise causes the container and browser to > terminate. > > It'd ultimately be nice to package the Tor Browser Bundle properly for > guix, but it's nice to be able to use it this way in the meantime. > > -Jim > > [Tor Browser Bundle]: https://www.torproject.org/download/ Any idea how to use this for running appimages? Or anything that requires FUSE in general?
Re: u-boot-am335x-boneblack -> u-boot-am335x-evm-boneblack
On 2022-12-22, Maxim Cournoyer wrote:
> Vagrant Cascadian writes:
>
>> Wondering what necessitated this change from the old variable name to a
>> new name...
>>
>> commit c04528d2a2597d79278833f3607c806278253446
>> Author: Maxim Cournoyer
>> Date: Tue Dec 20 21:25:27 2022 -0500
>>
>> gnu: u-boot-am335x-evm-boneblack: Fix variable name.
>>
>> * gnu/packages/bootloaders.scm (u-boot-am335x-boneblack): Rename to...
>> (u-boot-am335x-evm-boneblack), to match the package name.
>> * gnu/bootloader/u-boot.scm (u-boot-beaglebone-black-bootloader): Adjust
>> accordingly.
>> ...
>> diff --git a/gnu/packages/bootloaders.scm b/gnu/packages/bootloaders.scm
>> index bd9f7bb577..c8b8adbc93 100644
>> --- a/gnu/packages/bootloaders.scm
>> +++ b/gnu/packages/bootloaders.scm
>> @@ -889,7 +889,7 @@ (define*-public (make-u-boot-package board triplet
>> (define-public u-boot-malta
>>(make-u-boot-package "malta" "mips64el-linux-gnuabi64"))
>>
>> -(define-public u-boot-am335x-boneblack
>> +(define-public u-boot-am335x-evm-boneblack
>>(make-u-boot-package
>> "am335x_evm" "arm-linux-gnueabihf"
>> ;; Patch out other device trees to build an image small enough to fit
>>
>> The u-boot-am335x-boneblack was named to match the original target that
>> was removed from upstream, adapting the upstream am335x-evm to fit into
>> a smaller gap in the partition tables... (e.g. 2MB partition offset
>> instead of 4MB offset required by the default am335x-evm board
>> configuration).
>
> The problem was that the *name* of the package was
> "u-boot-am335x-evm-boneblack", as computed by the MAKE-U-BOOT-PACKAGE
> procedure, which includes the board argument in its name (it's been like
> this since its inception in 862e38d5518, 2017).
>
> If the previous variable name should have been its name, the name field
> would have needed to be overridden to it (or perhaps we could introduce
> a #:name argument that would take precedence over any cleverness).
>
> I noticed of the problem when trying to build the package; "guix build
> u-boot-am335x-boneblack" would tell me it didn't exist.
>
> I considered making a deprecated alias but decided against, because in
> the past we didn't when moving/renaming packages *variables*.
Odd. I was certainy able to build u-boot-am335x-boneblack from commit
6b99afeef89233b71d113a63cf04a6b4b49a4679 when it was introduced in 2019,
though it has been quite some time since I tested it...
I will take a guess that it was commit
c2c1dfdf5760873f1db86d14873f725a105f7feb which removed the "name" bit:
(define-public u-boot-am335x-boneblack
- (let ((base (make-u-boot-package
- "am335x_evm" "arm-linux-gnueabihf"
- ;; Patch out other device trees to build an image small enough
- ;; to fit within typical partitioning schemes where the first
- ;; partition begins at sector 2048.
- #:configs '("CONFIG_OF_LIST=\"am335x-evm am335x-boneblack\""
-(package
- (inherit base)
- (name "u-boot-am335x-boneblack")
live well,
vagrant
signature.asc
Description: PGP signature
Re: u-boot-am335x-boneblack -> u-boot-am335x-evm-boneblack
Hi Vagrant, Vagrant Cascadian writes: > Wondering what necessitated this change from the old variable name to a > new name... > > commit c04528d2a2597d79278833f3607c806278253446 > Author: Maxim Cournoyer > Date: Tue Dec 20 21:25:27 2022 -0500 > > gnu: u-boot-am335x-evm-boneblack: Fix variable name. > > * gnu/packages/bootloaders.scm (u-boot-am335x-boneblack): Rename to... > (u-boot-am335x-evm-boneblack), to match the package name. > * gnu/bootloader/u-boot.scm (u-boot-beaglebone-black-bootloader): Adjust > accordingly. > ... > diff --git a/gnu/packages/bootloaders.scm b/gnu/packages/bootloaders.scm > index bd9f7bb577..c8b8adbc93 100644 > --- a/gnu/packages/bootloaders.scm > +++ b/gnu/packages/bootloaders.scm > @@ -889,7 +889,7 @@ (define*-public (make-u-boot-package board triplet > (define-public u-boot-malta >(make-u-boot-package "malta" "mips64el-linux-gnuabi64")) > > -(define-public u-boot-am335x-boneblack > +(define-public u-boot-am335x-evm-boneblack >(make-u-boot-package > "am335x_evm" "arm-linux-gnueabihf" > ;; Patch out other device trees to build an image small enough to fit > > The u-boot-am335x-boneblack was named to match the original target that > was removed from upstream, adapting the upstream am335x-evm to fit into > a smaller gap in the partition tables... (e.g. 2MB partition offset > instead of 4MB offset required by the default am335x-evm board > configuration). The problem was that the *name* of the package was "u-boot-am335x-evm-boneblack", as computed by the MAKE-U-BOOT-PACKAGE procedure, which includes the board argument in its name (it's been like this since its inception in 862e38d5518, 2017). If the previous variable name should have been its name, the name field would have needed to be overridden to it (or perhaps we could introduce a #:name argument that would take precedence over any cleverness). I noticed of the problem when trying to build the package; "guix build u-boot-am335x-boneblack" would tell me it didn't exist. I considered making a deprecated alias but decided against, because in the past we didn't when moving/renaming packages *variables*. -- Thanks, Maxim
Re: security tracking NLNet grant
Csepp writes: > Anyone knows who is/was working on this and what happened to the > project? > The site only links to general Guix pages. > > https://nlnet.nl/project/GUIX-securitytracking/ The relevant mailing list thread is here https://lists.gnu.org/archive/html/guix-devel/2021-04/msg00036.html For various reasons, it didn't go anywhere. signature.asc Description: PGP signature
Re: u-boot-am335x-boneblack -> u-boot-am335x-evm-boneblack
Vagrant Cascadian writes: > With all that said... having 512MB of ram, I wonder how well a > beaglebone black would do running guix at all... I used to use Guix on Debian (not Guix System) on my BeagleboneBlack. With swap space, it worked well, even though `guix pull` took multiple days to complete. But I haven’t used it since 2020 and lost my UART (gotta be somewhere…) and the BBB’s screen remains black now for some reason. Regards, Florian
