<mnkydeth>
I remember mention of a wayland how to? Maybe a wiki in the making? Gonna try my hand at a wayland + sway + pipewire install. I cut out most of the install CD and got the bare minimum of base and a few things from opt. Any hints or tips anyone can give before I get started?
<mnkydeth>
cool, some searching I found TimB site with some info pertaining to Crux and an oper.io site that also has specific programs used to get basic functionality back.
chrcav has quit [Ping timeout: 272 seconds]
chrcav has joined #crux
crash_ has quit [Remote host closed the connection]
<farkuhar>
mnkydeth: what constitutes "basic functionality" for you? If you need to run GIMP, you should be aware that native Wayland support is only available in the 2.99 (development) version. GIMP 2.10 in the repos will require the xorg-xwayland compatibility layer.
<farkuhar>
similarly, if any of your apps use FLTK for their GUI, you can run them under Wayland after installing xorg-xwayland, or you can maintain your own port of FLTK 1.4 (the development branch) if you want to avoid Xorg as much as possible.
ppetrov^ has joined #crux
crash_ has joined #crux
<nggit>
hello is there any tools to scan the health of urls on ports. some upstreams make urls not immortal.
<farkuhar>
nggit: checking the source urls would be a nifty feature to add to prtverify. Feel free to write a patch and submit it to the CRUX team for review.
<farkuhar>
with the output of prtverify being used for the color-coding of individual ports and repos on the unofficial portdb (which might eventually replace the official portdb), nggit's suggestion would make it easy to see at a glance which ports have broken source urls.
<SiFuh>
Cool idea
<SiFuh>
I usually run 'pkgmk -do' but that downloads the archive
<SiFuh>
I am sure there will be some bugs, but I just whipped it together pretty quickly.
jue has joined #crux
<cruxbot>
[contrib.git/3.7]: asciidoc: new dependencies: python3-build python3-installer
<nggit>
yes, let's say if the owner changes the url that should have been 404 to 200. so it can't be accurate.
<SiFuh>
It's a start, hopefully it inspires someone to write ;-)
<nggit>
great
<SiFuh>
I ran it on 3.6 core and I found this. ./sqlite3/Pkgfile: line 9: printf: 3 39 3: invalid number
<SiFuh>
Heh
<SiFuh>
_version=$(printf "%i%.2i%.2i%.2i" ${version//./ }) :-) Haha that's a bit uneccessary
<nggit>
hmm.. i run debian desktop on crux. hope you try. this is easy if you have docker installed :D - Run Docker Container with Desktop Environment on tty10 (Ctrl-Alt-F10), coexist with the host. https://github.com/nggit/docker-tty
<cruxbot>
[opt.git/3.7]: ffmpeg4: 4.4.2 -> 4.4.3
<nggit>
will try to find other ways of checking port health, if any.
<jue>
SiFuh: why unneccessary?
<jue>
SiFuh: if you look at sqlite's version string you'll see why we did that
<SiFuh>
jue: sqlite's version string is uneccessary
<jue>
oops, you need that to download the sources, right?
<jue>
SiFuh: if you have a better idea how we can have a propper version number in our port, just speak up
<jue>
*proper
<cruxbot>
[xorg.git/3.7]: xorg-libxrender: update to 0.9.11
<cruxbot>
[opt.git/3.7]: glslang: update to 11.12.0
<cruxbot>
[opt.git/3.7]: spirv-tools: update to 2022.3
ahmadraniri[m] has left #crux [#crux]
<cruxbot>
[contrib.git/3.7]: ddgr: 2.0 -> 2.1
<cruxbot>
[opt.git/3.7]: flac: 1.4.1 -> 1.4.2
<farkuhar>
SiFuh: would you prefer to define the sqlite3 version as a 7-digit integer right at the start, rather than running printf and bash substitutions on the decimal-separated string of major/minor version numbers?
<farkuhar>
The present method of defining _version is much cleaner than some transformations I can imagine.
<farkuhar>
Example using bash array: vparts=(${version//./ }); _version=0; i=0; while [ $i -lt ${#vparts[@]} ]; do j=$(( 6-2*i )); dshift=$(echo "10^$j" | bc -l); _version=$(( _version+${vparts[i]}*$dshift )); i=$(( i+1 )); done
<farkuhar>
I'd prefer to reserve the criticism "a bit unnecessary" for stuff like the latter, rather than what's currently done in sqlite3.
<cruxbot>
[contrib.git/3.6]: python3-importlib_metadata: updated to version 5.0.0
<cruxbot>
[contrib.git/3.7]: python3-importlib_metadata: updated to version 5.0.0
<SiFuh>
farkuhar: I am not critisizing the _version in the Pkgfile. I am critizing the sqlite version.
<SiFuh>
The version is encoded so that filenames sort in order of increasing version number when viewed using "ls". For version 3.X.Y the filename encoding is 3XXYY00. For branch version 3.X.Y.Z, the encoding is 3XXYYZZ.
<SiFuh>
So for the sake of having it in order, they'd rather not follow their actual version number. So why bother calling version 3.39.3 when they can just call it version 3390300.
<jue>
SiFuh: No, i don't see any valid reason to change the port
<SiFuh>
No worries
ppetrov^ has quit [Quit: Leaving]
jue has quit [Ping timeout: 276 seconds]
<ocb>
jaeger: have bringed back up my repo.
<ocb>
would be nice to have it back on the list.
zou1 has joined #crux
zou1 has quit [Ping timeout: 244 seconds]
<farkuhar>
SiFuh: I find it amusing that the project even bothers to have a github mirror, considering their preference for fossil: https://sqlite.org/whynotgit.html
<SiFuh>
farkuhar: or the fact they name the versions correctly in github ;-)
brian|lfs has quit [Remote host closed the connection]
brian__ has joined #crux
brian__ is now known as brian|lfs
<farkuhar>
Quoting https://github.com/sqlite/sqlite , "The names of check-ins and other artifacts in a Git mirror are different from the official names for those objects."
<stenur>
removed asciidoc.. an unbelievable amount of python modules wanted to come in .. just to _install_ that thing!
<mnkydeth>
There is a signature mismatch on egl-wayland in contrib.
<farkuhar>
SiFuh: the sqlite tarball you get from github (or even using www.sqlite.org/src/tarball/version-$version/${name%3}.tar.gz) would require the build-time dependency contrib/tcl to generate the correct Makefiles. Since this is a core port, it should not depend on contrib/tcl; hence the use of a "pre-processed" source archive under the 2022/ directory.