Ah thanks so much! This is my first time unretiring something :)
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/proj
I recently requested to unretire rocm-smi, but after it was completed I tried
to rebuild and I get this:
BuildError: package rocm-smi is blocked for tag f39-updates-candidate
See: https://koji.fedoraproject.org/koji/taskinfo?taskID=103328747
So I'm not sure if a) it was not unretired correctly,
+1
Yes this has been mentioned many times on the thread. You can't say the user
has consented but also have it opt-out.
Saying that opt-in data isn't useful because most users won't opt-in is
implying the desire of a dark pattern to encourage more data collection.
___
Agreed 100%. Dark patterning or similar isn't the way to go.
If telemetry is included, it should be opt-in with very clear explanation of
why opt-ing in is important and beneficial.
Opt-out and "by consent" are mutually exclusive in most circumstances.
___
Unfortunately this might just be what happens.
I know that I would personally always opt out on principle, and would vote for
opt-in or dropping the proposal. I am under the impression that most Fedora
users are in the same boat as me.
___
devel mailin
Thanks Luya! I've landed rocm-opencl in rawhide, with epel8/9 and Fedora 36
pending :)
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code of Conduct:
https://docs.fedor
I'm looking to see if anyone wants to review swap with me:
https://bugzilla.redhat.com/show_bug.cgi?id=2090823
Thanks!
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code
Fantastic idea, I've just created a new page. I'll update it as I have time:
https://fedoraproject.org/wiki/SIGs/HC
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code of
A few people contact me directly trying to run things like PyTorch, which
requires large amounts of ROCm to get working (most of which Fedora does not
have yet).
I feel like a SIG, or at least some wiki page would help organize things a bit
for those who want to tackle it but are unaware of the
Thanks Felix,
An open concern that I have been discussing with the ROCm guys is that HIP
(used for things like pytorch) requires rocm-opencl source code to compile.
It seems they want to go with the llvm-project approach in the longterm, having
opencl, hip, and the common static lib "ROCclr" as
I'm up for a review swap if there's no takers.
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduc
For anyone interested I made a review request:
https://bugzilla.redhat.com/show_bug.cgi?id=2090823
I'm not 100% sure what to do with some of the debug related rpmlint errors. Any
help would be much appreciated.
___
devel mailing list -- devel@lists.fedo
As far as I know, RedHat is working with Nvidia to get this upstream and
working with nouveau.
I'm sure there's a bunch of challenges around that though, so I don't expect
much over the next few quarters.___
devel mailing list -- devel@lists.fedoraproj
+1 to using an rpm macro to avoid adding an external script, if spectool can
work with it.
Something like:
%global source0_generate_script ( \
curl ... \
rm -rf ... \
tar ... )
I'm not sure if that syntax is correct.___
devel mailing list -- devel@li
> Nice work!
Thanks :)
> The only x86 32-bit use I can think of would be wine if it supports ROCm.
Yeah I looked into it, and it looks like rocm-runtime fails on 32bit due to
some assumptions for 64bit in the code.
I don't think there's too much value to 32bit, so it's not really worth trying
I made a thread late last year inquirying about interest in ROCm packaging; in
that time I've introduced a few packages amd updated a few existing packages to
the latest version.
Right now, Fedora is just short of making good use of ROCm, as it needs a
frontend like OpenCL or HIP. I have a COPR
It seems like my needs are pretty common: Steam, Wine, misc games (OpenGL, SDL,
maybe vulkan, XWayland).
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code of Conduct:
Quick update, I've made some new package reviews:
ROCm-Device-Libs: https://bugzilla.redhat.com/show_bug.cgi?id=2044664
ROCm-CompilerSupport: https://bugzilla.redhat.com/show_bug.cgi?id=2045955
ROCm-Device-Libs is needed to update "rocm-runtime" and for
ROCm-CompilerSupport.
ROCm-CompilerSupport
I created a new review request, hopefully that encourages some conversation on
the topic :)
https://bugzilla.redhat.com/show_bug.cgi?id=2044664
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fed
In order to update "rocm-runtime" to the latest, it requires a new package
"ROCm-Device-Libs" as a build requirement.
The issue is that the project installs files into /usr/amdgcn, which seems
incorrect to me based on the FHS and Fedora guidelines. Here's the upstream for
reference:
https://gith
Yes, this can't be updated until someone packages ROCM-Device-Libs
unfortunately.
If anyone volunteers, I'm happy to help review.
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
> On Thu, Dec 16, 2021 at 05:07:10PM -0000, Jeremy Newton wrote:
>
> I think that'd be awesome -- and those internal clean-ups are really
> appreciated. Having the infrastructure there is nice, but I'm also curious:
> are there any application-level tools that are i
Yeah I think the technical leads are mostly on board with following FHS as
close as possible, which is an obvious plus for Fedora.
I think the biggest issue is the scale of the problem, and it almost feel likes
they need to work component by component, but [2] will definitely be fixed for
all c
Full disclosure, I am both a Fedora packager and an employee at AMD.
To be clear, the following is not at all endorsed by my employer; my interest
and use of Fedora is purely a personal hobby and I would like to keep it that
way.
There has been a recent effort to step up Debian packaging of ROCm
Indeed, but they should add the provides: bundled(cubeb) in the spec
Anyway, I made a review request for anyone interested:
https://bugzilla.redhat.com/show_bug.cgi?id=1826034
I plan to make a pull request to Firefox later if I can get it to unbundle.
Good to know, thanks!
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https
I package dolphin-emu, which bundled libcubeb in the latest version.
I've built it in rawhide and I'm trying to systematically unbundle things.
Looks like cubeb is apart of the Firefox source tree (./meda/libcubeb/). Should
I email gecko-bugs-nob...@fedoraproject.org or make a bug report?
It look
I feel like the batched update makes a lot of sense, providing the same amount
of QA/testing time is still provided and some rules are set on what can and
cannot be pushed in that update. E.g., since GTK now has a LTS model, I would
assume major release updates would only ever be pushed to rawhi
Thanks! I'll take a look into the port guide.
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
Anyone have any idea what causes these errors? Trying to update a package and
it failed during a local test build in mock (f25):
In file included from /usr/include/c++/6.2.1/stdlib.h:36:0,
from expr.ypp:5:
/usr/include/c++/6.2.1/cstdlib:124:11: error: '::div_t' has not been decla
So does exclusive arch actually block the unsupported arches come f26? The
emails are annoying but I'm more concerned that things will be broken when
branch from rawhide happens.
On Mon, Nov 14, 2016 at 6:50 AM, Pavel Raiskup wrote:
> On Sunday, November 13, 2016 4:28:26 PM CET Jerem
Hi,
I was wondering if any of the RPM guru's know how to fix an issue I'm having.
I keep getting this email:
>orthorobot has broken dependencies in the rawhide tree:
>On ppc64le:
>orthorobot-1.1-4.fc26.noarch requires love
>Please resolve this as soon as possible.
This is because it's noa
At risk of asking a redundant question, I'm assuming Wayland is still a go?
IIRC the contingency deadline was the beta.
I ask because it does not appear to be a part of the changeset, yet this FEDCo
ticket seems to suggest otherwise:
https://fedorahosted.org/fesco/ticket/1615
>>* Do you think that the average user with a clicking sound card or disk
*>>* corruption when suspending would be able to make the link to this new
*>>* package?*
> Even better ... the integrated mouse pointer on my external ThinkPad USB
> keyboard stops working if USB suspend is enabled for this
34 matches
Mail list logo