TobiSGD, I haven't had any error messages before.

Which ones are you facing ?
are you building with pam support, or without etc.

I am building without PAM. The errors I get are about missing libraries, issued by libtool. Example:Building with -j1 however works, now I only have to figure out how to actually use systemd.

I think it is a missing dependency in the Makefile.

Especially since it works with -j1.

Are you building with or without python support?

I use the Slackbuild as is, except of course switching out -j7 with -j1.

Systemd should build with -j2 or higher MAKEOPTS flags. If it's not then I'd do a query of any dependencies in the system and double or triple check everything before continuing. Check your compiler as well.

One minimal build I've seen (LFS-systemd experimental) disables gudev and python support from the main build.

As an option try testing flags...

--disable-gudev

And...

--without-python

These flags are for entirely optional features according to the documentation. Maybe it's something else though if this isn't it.

Has anyone checked the configure script to see if a flag for

--disable-shared

Exists?

Tobi, I was able to reproduce the same error with an AMD CPU only.

libsystemd-shared.la is being build by systemd itself, so either amd cpu's are doing some different things in parallel, or there is something different wrong.

I have currently no Intel systems available, the building was done on an Athlon QL-66 (dual-core) and a Phenom II X6 (six cores), both had that problem. Might be reasonable to add the information to your Wiki that in case of errors on AMD CPUs compiling with -j1 is recommended.

It has nothing to do with the processor type.

It only has to do with a missing dependency in the Makefile.

1 members found this post helpful.

Which?

I assume you are meaning which Makefile... Can't help there.

Systemd has tentacles in so many places. But I would guess it is missing from the core systemd Makefile.

Like I said in this thread, I run that SystemD build in two of my machines. Nothing spectacular, two desktop systems, on somewhat old hardware (think about P4 w. socket 478, with 1G RAM & Co.)...

Also, both runs fine that "disgusting bloated and resurce hog" called KDE(4).

I like to say that everything works fine, from all that time, feeling it as stable solution.

Congratulation, bartgymnast! Looks like you successfully ported that (so hatred) SystemD into Slackware!

Then, we have SystemD. Beyond of that, is just a question of taste (or religious beliefs) on using it or not.

Agree. It looks like it is compiled. Now to actually make it work takes a good bit of changes to the startup scripts, and in some cases, changes to the services themselves.
Not exactly. It is a question of reliability. Network analysis for dependencies is an NP hard problem. It can be done for small environments such as desktops. But then, the time it takes is better put to use in actually starting them than in computing which comes first. Systemd attempts to do that with a "preprocessed network state"... Unfortunately, it can't be reliable as some things depend on outside operation before they can be started. Also systemd has trouble with timing issues - look at NetworkManager. Systemd has to depend on NetworkManager to tell it when things are ready, unfortunately, NetworkManager also depends on external things (like DHCP, DNS...) to be ready, and some of them aren't. By default systemd assumes the network is ready when NetworkManager is started... or (if directed) when NetworkManager tells systemd the network is ready. Then systemd starts things that depend on the network... like apache, which has been known to fail because the network isn't quite ready (dhcp took a little too long), or DNS isn't quite ready (after all it takes time to load tables...). Sometimes it works - if other services cause these to delay a bit while they get started. But it is now a VERY hard problem to solve when it doesn't work. And then there are services that SOMETIMES have additional dependencies - like apache needing a database (is it local... or remote? or both?)

It used to be trivial to handle...

1 members found this post helpful.

I've just compiled the systemd package on a freshly installed Virtualbox Slackware64 14.1 with an AMD Phenom II X4, with -j7, it did error on the readme while copying stuff over to the package. Tried to build it again with the "install readme" line commented out and it build successfully.

Getting the network up and running has been a long running argument for server usage of systemd and parallel loading of daemons compared to traditional linear loading methods used by sysvinit.

Here's a good question, could systemd be scripted to load network related services without using parallel loading?