Feoramund

`linux.ppoll` always returns `E_INVAL` due to invalid `sigsetsize`

5

## Context It seems the Linux kernel is expecting a specific value for `sigsetsize` (which glibc handles), despite what the man pages say about this syscall. In Odin, we pass...

The tokenizer and the parser were reporting it in different positions. This way, they'll report in the same spot. I removed the `TODO(bill)` comment under the assumption that since it...

encoding/json: unmarshal uses the first struct in a union of structs

5

## Context `json.unmarshal` seems to be ignoring alternative structs in a union, because if it doesn't find the fields it's looking for on the first struct variant it tries, it...

I am not 100% sure if this is a bug or a system-dependent quirk, so I'm outlining my thoughts ahead. This _does_ fix my issue. I was having an issue...

## Context Odin: dev-2024-04:a61ae7c86 OS: Arch Linux, Linux 6.8.5-arch1-1 CPU: 12th Gen Intel(R) Core(TM) i7-12700K RAM: 31916 MiB Backend: LLVM 14.0.6 The following scenario causes a segmentation fault: ### main.odin...

text/table uses bytes instead of runes for length

4

## Context The `text/table` library is using `len` where `utf8.rune_count` would probably be more appropriate. See the following program and its output. ```odin package main import "core:os" import "core:text/table" main...

Provide error bool for FreeBSD syscall return value

7

I'm working on porting `core:net` to FreeBSD, and I ran into this predicament where `syscall` was returning only positive numbers, which isn't too helpful for detecting errors. I know very...

Triggered this with `syntax_error_va` a moment ago and noticed that this was a pattern I missed when I fixed the race condition in `error_va`. I fixed all the ones I...

## Context I discovered this bug after making a typo with an `append` parameter. It's a regression caused by the fix I introduced in commit ebfbe4d26031bde4e91ba489102d60fe6959173f. I narrowed the reproduction...

Support `testing.set_fail_timeout` on UNIX-likes

5

This will require #3468 in order to terminate tests that are stuck in a loop or otherwise preoccupied, but it does register the test as a failure, if it ever...