On 2021/08/12 12:11, Carlos E. R. wrote:
On 12/08/2021 20.58, L A Walsh wrote:
On 2021/08/12 11:54, Stefan Seyfried wrote:
On 12.08.21 20:31, L A Walsh wrote:
I noticed a bunch of 32bit packages that often get installed when there is an identically named package but without the -32bit- in the name.
What's the difference between the ones that say -32bit- vs. not, and why is a 64bit installation preferring the -32bit- variant of the package. The -32bit packages are 32bit, the others are 64bit.
That was easy ;-)
You need the -32bit packages if you want to run 32bit binaries, e.g. readily compiled stuff downloaded from the internet.
Um...I thought that's what the i586 packages were for.
i586 are for a full 32 bit system and processor.
In general, I try not to run 32bit so I don't need duplicate libraries.
Wine. Stefan mentioned it in passing, it is the most common application in a 64 bit system that demands 32 bit libraries because some Windows applications are 32 bit.
If you have wine installed, they come from that one.
Don't have wine installed. I run 32-bit applications on my windows machine. The linux machine is a server to provide services like filesystem, backups, mail, domain (single signon), DNS, web-access (squid). I can even put large win programs (like games) on the server and run them on windows (usually). So in the past, I haven't needed 32bit packages on the server -- and was wondering why I can replace them with the 64-bit packages, manually, and not run into dependency probs. I have zypper setup to not install recommended packages -- just deps. So still don't see why 32-bit packages have been getting installed over their 64-bit alternatives -- it didn't used to happen, so was wondering what might have changed, that's all...