jrgd

joined 10 months ago
[–] jrgd@lemm.ee 10 points 13 hours ago

I believe it's mostly drawing tablet support in Qt and in turn porting to Qt6 that's holding native Wayland builds back.

[–] jrgd@lemm.ee 6 points 6 days ago* (last edited 6 days ago)

A good amount of distros actively have this functionality. To avoid breaking system packages, you can install the distro package for the given module or as the error recommends: use a venv for the given project.

As to why many guides don't include it, I suspect as typical for many Linux-centric articles: they weren't been written by knowledgeable individuals or just in general are writing with knowledge that is often 5+ years out of date.

[–] jrgd@lemm.ee 10 points 1 week ago (2 children)

My top picks currently for distros that support KDE are the following:

For your use case (Nvidia, Wayland preferential), the better choices among these will likely be the rolling releases (OpenSUSE Tumbleweed, ArchLinux) or 6 month point releases (Fedora KDE). Debian and OpenSUSE Leap are solid choices for LTS, but given the state of Nvidia and Wayland, it's best to use the latest releases of KDE and the proprietary Nvidia drivers. If you switch GPUs to AMD or Intel in the future, you should have no issues using any of the distros listed.

To put points against some of the distros your contending list:

Many of the direct Ubuntu-based distros tend to have a certain level of lesser quality in packages (such as many releases never end up pushing bugfix patches that get patched in many other distros including Debian). Additionally, there is no guarantee that Ubuntu-derivative distros that don't directly source from Ubuntu software repos may have breakages when using PPA repos or developer-distributed .deb packages.

I'm sure you're aware of this bit as well, but the mainline Canonical-maintained distros (Ubuntu, Xubuntu, Kubuntu, etc.) rely heavily on Snap: a containerized application platform similar to flatpak, but with no freedom of choice of package sourcing. Every Snap package will be pulled from Canonical's proprietary publishing platform. A lot of derivative distros (Linux Mint, Pop! OS, etc.) end up stripping out Snap from default installations and removing package redirects, recommends for Snap.

For Arch derivatives (Endeavour, Manjaro, etc.), don't expect to be able to use AUR packages without issues unless your derivative directly sources from the ArchLinux repos. Many AUR packages explicitly expect the latest packages, which some derivatives defer updates to, causing breakages.

In particular, Manjaro has a track record of poor maintenance and questionable choices (recommending users to roll back system clocks after forgetting to renew TLS certs, shipping outright broken versions of Asahi Linux in order to tout support for Apple hardware, DDOS'ing the AUR, etc.)

Debian Sid (the unstable (rolling) variant Debian) is an option, but it's really not recommended for end-use, and mostly only for testing.

To put points against some of the distros on my recommendation list:

Fedora explicitly only ships with FOSS software. This does mean that initial NVidia driver setup is more involved compared to most distros. The process shortlist is initial boot with nomodeset, install rpmfusion repos, and then install the NVidia drivers from RPMFusion-nonfree. Once that is done, the proprietary drivers should be installed and all configurations necessary should already be made. Simply rebooting should allow using the system accordingly.

Installing ArchLinux specifically expects some knowledge of the inner workings of a Linux system. Modern Arch live images do come with Archinstall: a utility that assists in getting an installation from configuration options. In general, an Arch install is a more involved process. ArchLinux also expects that you read from the news page before pushing updates to your system. While this kind of practice can also be true for many other rolling systems/point releases between feature upgrades, it is fairly imperative that due diligence and backups are taken on Arch systems when updating.

[–] jrgd@lemm.ee 4 points 2 weeks ago

The best three brands with natively-supported hardware:

  • HTC's Vive series headsets
  • Bigscreen's Beyond
  • Valve's Index

Pretty much everything else requires a lot more tinkering than just launching SteamVR/OpenVR applications.

Some helpful links for diagnosing compatibility:

[–] jrgd@lemm.ee 16 points 2 weeks ago

For the most part, you won't be able to escape Unix-like paradigms when using Unix-like systems. Notably, users have to exist in some form. You don't necessarily need to give them passwords for the frontend signage, but they need to exist. The shortlist of setting up cage would be:

It's not quite a few clicks, but this can in contrast also be fully automated trivially if it's something you need to setup more than once.

[–] jrgd@lemm.ee 39 points 2 weeks ago (3 children)

In what way does Windows fulfill a 'kiosk' display mode better than Linux for you? Are you looking for permanent installations or just temporary lockdown to a single application. One of the more modern and straightforward methods currently is using cage.

Cage lets you spawn a Wayland compositor from command-line (or via system service, obviously) that launches either a singular or multiple exclusively-fullscreen applications.

 

A bit of basic information beforehand that should be relevant:

  • OS: Fedora 40 (x86_64)
  • Desktop: KDE Plasma 6.1.3 (5.27.x -> current) (Wayland)
  • Motherboard: ASUS PRIME X470-PRO
  • Primary keyboard: Keychron S1 (tested on stock firmware, Windows layers)

The issue in some more detail: alt-tabbing windows in KDE sometimes leaves behind an alt keypress in the window that was alt-tabbed from that won't go away until alt is registered as pressed again by the window.

This has certainly been an interesting issue that has been a problem for at least a year at this point. I've finally gone in to do basic troubleshooting regarding the issue. I have pretty much ruled out hardware at this point. Originally, I had my primary keyboard plugged into my monitor's USB hub. In testing, I tried another keyboard, migrated the connect to my motherboard rather the monitor's USB hub, and the alternate keyboard plugged into the motherboard. All tests eventually result in the same issue happening over time.

One thing I see as a potential pattern but cannot fully confirm is that it appears these remnants of alt keypresses only show up in XWayland windows, and not native Wayland applications. Applications I have seen it occur in include Minecraft Java Edition, various Proton games, Electron applications that don't enable native Wayland support (e.g. Discord), Steam. This issue also occurs less if I allow XWayland apps to always see modifier keypresses. Though I assume this is more from the tendency to alt-tab back into affected windows, registering another alt keypress.

At this point, I am pretty confident this is software-related, but I cannot find any existing bug reports within the past three years (especially noting any relevant to KDE) that describe this problem. I am unsure of how one would go to debug this category of issue or even find the root cause component that is causing the issue. I am curious if anyone here has also had this irritating quirk occur and/or if you know any workarounds/fixes for this given problem.

[–] jrgd@lemm.ee 6 points 2 weeks ago

An aside to the technical question of how to migrate profiles to older versions:

DO NOT DOWNGRADE FIREFOX BELOW 131.0.2 OR ESR 128.3.1, 115.16.1

I feel that given this recent vulnerability, it is important to make this notice.

Otherwise:

For migrating profiles between the same major version, Mozilla provides a guide for full profile migration. This also works with forwards compatibility. I generally wouldn't try to go backwards however as many new major versions change the data format and contents of your profiles, which older versions have no idea how to interpret.

For downgrading, it's best to export bookmarks, go through your important addons and backup the settings for each one that needs configuration, and take note of anything you're previously modified in about:config to your preference. Perhaps take screenshots of your tab bar and overflow menu as well so you can recustomize them to your liking easily on the downgraded version.

[–] jrgd@lemm.ee 1 points 3 weeks ago (1 children)

What hardware, audio interface, and sound server is in use for your 5.1 Surround setup?

[–] jrgd@lemm.ee 3 points 3 weeks ago

I am under the presumption that the current state of the Intel Arc Alchemist GPUs will likely remain about the same under Mesa even if support is dropped today by Intel. Am I mistaken in the amount of continued driver effort Intel has been putting in for the Mesa GPU drivers?

Obviously if this is true, one should probably remain wary of upcoming Battlemage GPUs.

[–] jrgd@lemm.ee 26 points 3 weeks ago (6 children)

A key list of compatible/incompatible components to look for:

  • GPU
  • Network Interfaces (Ethernet and Wi-Fi)
  • Audio Interfaces (not that much of an issue anymore)
  • Disks
  • Motherboards
  • CPU (excluding x86 ecosystem)
  • Peripherals

The explanations for this are pretty long, but are meant to be fairly exhaustive in order to catch most if any pitfalls one could possibly encounter.

GPU:

A big one is the choice between AMD, Intel, and NVidia. I am going to leave out Intel for compute as I know little about the state it is in. For desktop and gaming usage, go with AMD or Intel. NVidia is better than it used to be, but still lags behind in proper Wayland support and the lack of in-tree kernel drivers still makes it more cumbersome to install and update on many distros whereas using an AMD or Intel GPU is fairly effortless.

For compute, NVidia is still the optimal choice for Blender, Resolve, and LLM. Though that isn't to say that modern AMD cards don't work with these tasks. For Blender and Davinci Resolve, you can get them to use RDNA+ AMD cards through ROCm + HIP, without requiring the proprietary AMD drivers. For resolve especially, there is some serious setup involved, but is made easier through this flatpak for resolve and this flatpak for rocm runtime. ML tasks depend on the software used. For instance, Pytorch has alternate versions that can make use of ROCm instead of CUDA. Tools depending on Pytorch will often have you change the Pytorch source or you may have to manually patch in the ROCm Pytorch for the tool to work correctly on an AMD card.

Additionally, I don't have performance benchmarks, but I would have to guess all of these tasks aren't as performant if compared to closely equivalent NVidia hardware currently.

Network Interfaces:

One section of hardware I don't see brought up much is NICs (including the ones on the motherboard). Not all NICs play as nicely as others. Typically I will recommend getting Ethernet and Wireless network interfaces from Intel and Qualcomm over others like Realtek, Broadcom, Ralink/Mediatek. Many Realtek and Mediatek NICs are hit-or-miss and a majority of Broadcom NICs I have seen are just garbage. I have not tested AMD+Mediatek's collaboration Wi-Fi cards so I can't say how well they work.

Bluetooth also generally sits into this category as well. Bluetooth provided by a reputable PCIe/M.2 wireless card is often much more reliable than most of the Realtek, Broadcom, Mediatek USB dongles.

Audio Interfaces:

This one isn't as much of a problem as it used to be. For a lot of cards that worked but had many quirks using PulseAudio (a wide variety of Realtek on-board chipsets mainly), they tend to work just fine with Pipewire. For external audio interfaces: if it is compliant to spec, it likely works just fine. Avoid those that require proprietary drivers to function.

Disks:

Hard drives and SSDs are mostly fine. I would personally avoid general cheap-quality SSDs and those manufactured by Samsung. A lot of various SATA drives have various issues, though I haven't seen many new products from reputable companies actually releasing with broken behavior as documented by the kernel. If you wish to take a detailed look of devices the kernel has restricted broken functionality on, here is the list.

Additionally, drives may be one component beside the motherboard where you might actually see firmware updates for the product. Many vendors only release EXE files for Windows to update device firmware, but many nicer vendors actually publish to the LVFS. You can search if a vendor/device is supplied firmware here.

Motherboards:

In particular, motherboards are included mainly because they have audio chipsets and network interfaces soldered and/or socketed to them. Like disks, motherboards may or may not have firmware updates available in LVFS. However, most motherboard manufacturers allow for updating the BIOS via USB stick. Some laptops I have seen only publish EXE files to do so. For most desktop boards however, one should be able to always update the motherboard BIOS fine from a Linux PC.

Some motherboards have quirky Secure Boot behavior that denies them being able to work on a Linux machine. Additionally some boards (mostly on laptops again) have either broken or adjustable power state modes. Those with adjustable allow for switching between Windows and standard-compliant modes.

Besides getting a Framework laptop 'Chromebook edition', I don't think there is much you will find for modern boards supporting coreboot or libreboot.

CPUs:

For your use case, this doesn't really matter. Pretty much every modern x86 CPU will work fine on Linux. One only has to hunt for device support if you are running on ARM or RiscV. Not every kernel supports every ARM or RiscV CPU or SoC.

Peripherals:

Obviously one of the biggest factors for many new users switching to Linux is their existing peripherals that require proprietary software on Windows missing functionality or not working on Linux. Some peripherals have been reverse engineered to work on Linux (see Piper, ckb-next, OpenRazer, StreamController, OpenRGB).

Some peripherals like printers may just not work on Linux or may even work better than they ever did on Windows. For problematic printers, there is a helpful megalist on ArchWiki.

For any other peripherals, it's best to just do a quick search to see if anyone else has used it and if problems have occurred.

 

Greetings all!

I have been working on getting a new network setup. The current test host (A server running OpenSUSE Leap 15.6 w/ Wicked) is able to get routes and obtain an address via DHCP from the router of the network (running OPNSense 24.7.6), but is unable to resolve routes and obtain an address via the local DHCPv6 server. Admittedly, I am not great with IPv6 doubled with the ISP for this network granting a statically-defined /128 address for the router and manually-delegated /64 address blocks.

The OPNSense configuration has a /64 address block assigned as its address space for the LAN interface. The configuration has the ISC DHCPv6 server allocating address range 2602:xxxx:xxxx:xxxx::8888:0 - 2602:xxxx:xxxx:xxxx::8888:ffff. The radvd server is set to managed, set with an automatic source address, set to advertise the default gateway, set to use the dhcpv6 dns configuration, and set with no additional routes advertised.

As noted, the OpenSUSE machine is unable to get any routes beyond link-local via ipv6 nor is it able to automatically be assigned an ipv6 address from the DHCPv6 server. I have done some diagnostics, but have been unable to determine any conclusive issue.

Starting ip route and address checks:

ip -6 addr

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 state UP qlen 1000
    inet6 fe80::xxxx:xxxx:xxxx:a4ee/64 scope link proto kernel_ll [OpenSUSE Leap 15.6 Server link-local address]
       valid_lft forever preferred_lft forever

ip -6 route

fe80::/64 dev eth0 proto kernel metric 256 pref medium

The eth0 interface noted is using a standard configuration as provided by Wicked (BOOTPROTO=dhcp, STARTMODE=auto, ZONE=public). Testing dhcpv6 address acquisition by hand results in nothing:

wicked test dhcp6 -m auto eth0

wicked: eth0: Request to acquire DHCPv6 lease with UUID <$uuid-a> in mode auto

However, testing in forced managed mode does get results from the DHCPv6 server:

wicked test dhcp6 -m managed eth0

wicked: eth0: Request to acquire DHCPv6 lease with UUID <$uuid-b> in mode managed
INTERFACE='eth0'
TYPE='dhcp'
FAMILY='ipv6'
UUID='<$uuid-b>'
IPADDR='2602:xxxx:xxxx:xxxx::8888:807/128' [theoretical bound address on LAN]
PREFIXLEN='128'
DNSSERVERS='2602:xxxx:xxxx:xxxx::1' [LAN address of router]
DNSSEARCH='<$domain>'
ACQUIRED='1729020515'
CLIENTID='<$clientid>'
SERVERID='<$serverid>'
SERVERADDR='fe80::xxxx:xxxx:xxxx:a4ee' [OpenSUSE Leap 15.6 Server link-local address]

So unless I am mistaken at this point, this likely means that something is going wrong with the Router Advertisements for the system to not automatically try get assigned an ipv6 address. Checking a router advertisement broadcast to the OpenSUSE server, I am not seeing anything out of the ordinary:

radvdump

#
# radvd configuration generated by radvdump 2.17
# based on Router Advertisement from fe80::xxxx:xxxx:xxxx:4eb4 [router link-local on LAN]
# received by interface eth0
#

interface eth0
{
        AdvSendAdvert on;
        # Note: {Min,Max}RtrAdvInterval cannot be obtained with radvdump
        AdvManagedFlag on;
        AdvOtherConfigFlag on;
        AdvReachableTime 0;
        AdvRetransTimer 0;
        AdvCurHopLimit 64;
        AdvDefaultLifetime 1800;
        AdvHomeAgentFlag off;
        AdvDefaultPreference medium;
        AdvLinkMTU 1500;
        AdvSourceLLAddress on;

        prefix 2602:xxxx:xxxx:xxxx::/64 [public /64 address block manually delegated as LAN]
        {
                AdvValidLifetime 86400;
                AdvPreferredLifetime 14400;
                AdvOnLink on;
                AdvAutonomous off;
                AdvRouterAddr off;
        }; # End of prefix definition


        RDNSS 2602:xxxx:xxxx:xxxx::1 [LAN address of router]
        {
                AdvRDNSSLifetime 600;
        }; # End of RDNSS definition


        DNSSL <$domain>
        {
                AdvDNSSLLifetime 600;
        }; # End of DNSSL definition

}; # End of interface definition

sysctl -a | grep eth0.accept_ra

net.ipv6.conf.eth0.accept_ra = 1
net.ipv6.conf.eth0.accept_ra_defrtr = 1
net.ipv6.conf.eth0.accept_ra_from_local = 0
net.ipv6.conf.eth0.accept_ra_min_hop_limit = 1
net.ipv6.conf.eth0.accept_ra_mtu = 1
net.ipv6.conf.eth0.accept_ra_pinfo = 1
net.ipv6.conf.eth0.accept_ra_rt_info_max_plen = 0
net.ipv6.conf.eth0.accept_ra_rt_info_min_plen = 0
net.ipv6.conf.eth0.accept_ra_rtr_pref = 1

Am I missing something with why Wicked doesn't actually get a proper route to the LAN nor an address via IPv6?

To recap: IPv4 works, this is the only device connected to the network thus far, IPv6 configuration appears (to me at least) correct for the router advertisements and DHCPv6 config.

EDIT:

Found the source of the problem. The OPNSense configuration is in fact correct for what I want to do. The issue is on the OpenSUSE machine. I forgot about a funny little Linux kernel networking quirk regarding ipv6 forwarding. In OpenSUSE, enabling forwarding for IPv6 from the installer keeps net.ipv6.conf.*.accept_ra set to 1. However, setting net.ipv6.conf.*.forwarding to 1 will disable accepting routes from RA, and in my case of expecting automatic IPv6 configuration from DHCPv6 without forcing managed mode on the Linux server.

Unless I feel like bypassing some functionality provided by the router, one needs to set net.ipv6.conf.*.accept_ra to 2 for all affected network interfaces. This enforces accepting routes with forwarding enabled. This in turn for my case also allows for DHCPv6 resolution to function without forcing or bypassing it from the OpenSUSE machine. I can only assume the reason this isn't just default if applied from the installer is that fully-manual static IP addressing is expected rather than wanting to use DHCP reservations for assigning addresses.

So in short:

All is good with the OPNSense configuration. I needed to change the sysctl flag net.ipv6.conf.eth0.accept_ra = 1 to net.ipv6.conf.eth0.accept_ra = 2, in order to forcefully accept RA routes and normal DHCPv6 address assignment on my ethernet interface. This is necessary because I need forwarding over IPv6 for the affected machine.

[–] jrgd@lemm.ee 2 points 3 weeks ago

Reading through the link chain, it seems the Western Digital drives being shipped in those laptops really should have never made it into consumers' hands.

The kernel argument nvme_core.default_ps_max_latency_us=5500 is being used to restrict the power state latency in order to keep the drive out of its lowest power state (because of course yet another cheaply-made device has terrible power state management).

While most distros generally expect NVMe drive to not completely cease functioning while at idle (as should be expected really), AntiX is likely keeping the drive above its minimal power state. Whether this is intentional, unintentional, or from a lack of general power state management provided by the distro isn't something I know. It would require some digging in the source tree for the distro most likely to find if there are any deliberate restrictions to power saving, especially regarding NVMe.

[–] jrgd@lemm.ee 6 points 3 weeks ago* (last edited 3 weeks ago)

A couple things to check using a quick bash script:

#!/usr/bin/env bash

cd /sys/class/power_supply/BAT*/
echo "Charge cycles: $(cat cycle_count)"
printf '%s\0' 'Health: ' &
bc <<< "scale=3; ($(cat charge_full) / $(cat charge_full_design)) * 100"

That should print out the wear cycles the battery has endured and its reported capacity over design capacity. If your battery has less than 1000 cycles and the health reported from the battery is less than 80%, it might be best to contact Framework for warranty replacement as the battery is likely defective.

 

Greetings,

For several years, I have used the wonderful Cantata as a frontend to MPD. Sadly, the frontend stopped receiving updates in 2022 and has started to some problems with age. While I continue to use Cantata for as long as I can, I have been looking around at other music players. However, I haven't seen anything that aims to implement some of the nice things from Cantata.

In short, a few things I have been looking for in a player:

  • suitable for playing single songs, albums, full artists, custom mixes, or playlists (no hyperfocus)
  • can either set a custom artist sort tag (albumartist, composer, etc.) or properly handle semicolons (or some other separator char) in tags
  • semicolon tag split in general would be nice for genre handling
  • powerful active queue handling (move; shuffle and sort by song, album, artist; remove duplicates; consume on play; etc)
  • online lyrics search from multiple providers

Additionally, some nice-to-haves that Cantata handles:

  • CD ripping
  • export library to portable device (with compatibility)

Anyone have a favorite that can handle at least the shortlist of functionality I come to expect? I don't expect specifically a frontend for MPD, but I would prefer a player that doesn't struggle to handle a library with 10^4^ magnitude library size.

 

cross-posted from: https://lemm.ee/post/38676431

A while back I ended up getting tired of making hacks to get custom binaries to launch in Steam for Windows titles. Primarily for modding, I would find a way to simply launch custom EXE files through Steam to ensure the modding tools and the game were contained neatly in the same prefix. My first ventures with this were Skyrim and Fallout: New Vegas. With these titles, I overrode the gamebryo/creation engine launcher EXE with Mod Organizer 2 (renamed to be the launcher). While this worked, the solution doesn't work for other games without a secondary launcher that is targeted through Steam.

I eventually came to the conclusion that one can override launch targets entirely in Steam, and that tools like SteamTinkerLaunch could take advantage of this. However, STL certainly does a lot and honestly, that is way more than I really desired just to launch games with a custom EXE. Thus I made a shell script that essentially allows for the user to write in their own custom target and have it launch right through Steam.

The usage for this is simple. Just copy the 'shim' file into the game directory, override the Steam launch arguments to include "./shim %command%", and all is good. Furthermore, environment variables (such as DRI_PRIME=1), additional launch wrappers (gamemoderun), and game launch arguments (-novid for Source Engine titles) all work. If one needed a combination of all of this, it would look something like "DRI_PRIME=1 gamemoderun ./shim %command% -novid".

The way target editing currently works is on first launch the shim file grabs the default game target and writes it as the contents of 'target', another file in the game directory. From there, one can simply edit the target location in the file and shim will launch the custom executable.

So far, I have used this to get things like RaftModLoader and BeamMP working (mod loader for Raft and multiplayer for BeamNG.Drive respectively). I see no issue with this being able to also work for Bethesda titles and others that need custom executables. As I understand as well, the actual game install directories on a Steam Deck with SteamOS are mutable, and with a bit of tinkering through desktop mode should help get a seamless experience for launching modded Steam games for Deck users as well.

I hope someone finds as much use and utility that I have for getting a lot of modding tools for Windows games working without needing to mangle the prefix using protontricks in some cases or install the absolute multi-tool that is SteamTinkerLaunch.

 

A while back I ended up getting tired of making hacks to get custom binaries to launch in Steam for Windows titles. Primarily for modding, I would find a way to simply launch custom EXE files through Steam to ensure the modding tools and the game were contained neatly in the same prefix. My first ventures with this were Skyrim and Fallout: New Vegas. With these titles, I overrode the gamebryo/creation engine launcher EXE with Mod Organizer 2 (renamed to be the launcher). While this worked, the solution doesn't work for other games without a secondary launcher that is targeted through Steam.

I eventually came to the conclusion that one can override launch targets entirely in Steam, and that tools like SteamTinkerLaunch could take advantage of this. However, STL certainly does a lot and honestly, that is way more than I really desired just to launch games with a custom EXE. Thus I made a shell script that essentially allows for the user to write in their own custom target and have it launch right through Steam.

The usage for this is simple. Just copy the 'shim' file into the game directory, override the Steam launch arguments to include "./shim %command%", and all is good. Furthermore, environment variables (such as DRI_PRIME=1), additional launch wrappers (gamemoderun), and game launch arguments (-novid for Source Engine titles) all work. If one needed a combination of all of this, it would look something like "DRI_PRIME=1 gamemoderun ./shim %command% -novid".

The way target editing currently works is on first launch the shim file grabs the default game target and writes it as the contents of 'target', another file in the game directory. From there, one can simply edit the target location in the file and shim will launch the custom executable.

So far, I have used this to get things like RaftModLoader and BeamMP working (mod loader for Raft and multiplayer for BeamNG.Drive respectively). I see no issue with this being able to also work for Bethesda titles and others that need custom executables. As I understand as well, the actual game install directories on a Steam Deck with SteamOS are mutable, and with a bit of tinkering through desktop mode should help get a seamless experience for launching modded Steam games for Deck users as well.

I hope someone finds as much use and utility that I have for getting a lot of modding tools for Windows games working without needing to mangle the prefix using protontricks in some cases or install the absolute multi-tool that is SteamTinkerLaunch.

view more: next ›