Write a MemoryInfoListStream on linux.#70
Conversation
|
This relies on changes to |
This makes it easier for consumers to read the minidumps, rather than needing to parse linux-specific information. Closes rust-minidump#8.
The sanitized stacks test was previously not running the `Context::With` case. Now that the `With`/`Without` tests are programmatically toggled (to prevent such errors as that!), it is clear that there is something weird with the context's stack pointer, which seems to rarely (never?) lie in mapped memory. For now, this test is disabled.
f25207f to
7ef6250
Compare
|
Since a new release of |
The tests and helper functions already account for these platforms.
|
Thanks, I'll take a look at this tomorrow! |
Jake-Shadle
left a comment
There was a problem hiding this comment.
Oof, this is a big change, but the added/improved testing definitely helps!
|
@Jake-Shadle Yeah, I'm not sure if you saw but the changes to the tests exposed a bug in the tests which was hiding a failure (7ef6250). For now, I had no solution to the failure. |
gabrielesvelto
left a comment
There was a problem hiding this comment.
This is fine save for one small comment I left. The refactoring resulting from using the procfs crate is really nice and I also like how you macro-ized the tests.
|
CI failure seems to be from network errors. |
|
I'm re-triggering the jobs, I'll merge the changes as soon as they all turn green |
I'm reworking that stuff in #78. Right now situations in which the stack pointer falls out of a mapped range just don't work. |
This makes it easier for consumers to read the minidumps, rather than needing to parse linux-specific information.
Closes #8.