previous: >>107939911#define __NR_writev 20https://man7.org/linux/man-pages/man2/writev.2.htmloh man, this is basically the same as the last one.... not much to say today. i am really not looking forward to the setuid and xattr doldrums. i honestly am a bit tempted to squish them all together, and then maybe we can fit in some newer syscalls that weren't included in musl libc. maybe even start doing something similar with all the read/write and set/get pairs... but that's for me to decide another day.talk some more about vectored io, i guess. i am very eepy, so that's all you'll get from me today in terms of discussion starters. thank u for subscribing to my blogrelevant resources: man manman syscallshttps://man7.org/linux/man-pages/https://linux.die.net/man/https://elixir.bootlin.com/linux/https://elixir.bootlin.com/musl/https://elixir.bootlin.com/glibc/
#define __NR_writev 20
man man
man syscalls
>>107948632this is actually faster than plain write when writing to the ttyno idea why
>>107949050how extensive was your testing?
>oh man, this is basically the same as the last one....unlike readv, which is worthless, writev has an actual use case though
>>107949498go on?
>>107949303no testing, I realized it once but never cared enough to look into it
>>107949623Short reads are possible so readv is useless.
>>107948632Thanks chatgpt
>>107949840are short writes impossible?
>>107950578possible but much rarer
>>107952365why do you feel that way?
>>107949840Just retry the read surely? Adjusting the slices is annoying but not hard
>>107953873>Adjusting the slicesLinux guarantees that you won't need that
>coding thread on a nocoder board>filled with nocoder postslomoa
>>107953900Can't it stop halfway through a slice as with non-vectored I/O?
