Matt Mastracci
Matt Mastracci
Tweak to publish script: we can actually use `cargo info crate@version` and test the return value now.
Experiment to add low-level optimization to streams based on stream type. Not complete yet. This may be able to improve throughput for bulk restore operations.
Discussed offline. We should implement two environment variables that will redirect the FS access for two parts of the gel-rust driver (and eventually, the other drivers as well). This will...
One missing helper method -- a safe wrapper for `IsProcessInJob`. https://learn.microsoft.com/en-us/windows/win32/api/jobapi/nf-jobapi-isprocessinjob This could be exposed as something like `Job::contains(&self, &impl AsRawHandle) -> Result`.
It would be useful to have an `assign_child` function on `Job` which would simply be this ``` pub fn assign_child(&mut self, child: &std::prcess::Child) -> Result { use std::os::windows::io::AsRawHandle; job.assign_process(child.as_raw_handle() as...
After #8590 landed, the new connection pool appears to be working. The new pool may have triggered the connection leak more than the old one, causing it to show up...
This should be possible by removing the `super::...` call to the ctor method, but will need testing.
For 1.0, the docs need to be cleaned up, de-duplicated and re-reviewed. This includes: - [ ] Repo README - [ ] ctor README - [ ] ctor crate root...
#[dtor] should exist in both atexit and link_section varieties, with the user controlling how this works. ``` #[dtor(atexit)] #[dtor(cxa_atexit)] #[dtor(link_section=...)] ``` By default, we should use the most reliable form...