New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fails one test on i386 #2
Comments
That test is likely a harmless failure. It does mean the from_srgb8 function won't return bit-identical results with a naïve implementation, but there's likely no way we can guarantee that anyway because of how passing/returning floats via the x87 stack works. |
Thanks for the confirmation - I'll simply skip that test, then. By the way, in case you care: I am packaging this crate officially for Debian - Develpers' dashboard and package list |
Neat, feel free to let me know if there are any issues. |
I sure will |
Note: This is pretty long-winded and really is just background for a few questions I have -- skip to the end if this is too much right now (no worries). I'm interested in figuring out the cause of the failure (but don't have easy 32bit x86 access, and want to replicate it exactly when I do). Anyway, it has nothing to do with this project, the test is still ignorable if it fails. It's just possibly relevant for rust-lang/rust. So, while I stand by my earlier assessment, it's actually kinda surprising that this fails only on the To be clear, the test is not really robust, and is mostly written to catch a case where I break something locally and don't recompute the table. The most likely reason it's wrong is that the test side calls (after a few layers) the system The other option is that The log says Anyway, assuming it's the normal Anyway, that's all background to a few questions I'd like to ask you, if you know. No worries if you don't. The third question is the one I care the most about:
I think that's it. Honestly, it's very likely that the Footnotes
|
Debian architecture
At the time this was reported, only CI tests for i386 failed - all other architectures doing CI testing succeeded.
Yes, I am pretty sure that on Debian without overriding defaults (which I don't for this packaging) glibc is used.
As per my initial note above, I would indeed expect SSE2 to not be available on the Debian Perhaps you can spot some concrete evidence from the logs of the Debian machinery.
Sorry, I don't know such details. |
(Sorry for the delay in responding, been swamped at
Thank you, this answers my questions. And disabling that test across the board is totally reasonable -- probably the right call (the test is sloppily written, and intended to catch my own mistakes rather than anybody elses). Thanks for getting back to me, it's cool to see that Debian runs package tests. |
Fat fingered the close button, my bad. (This isn't closed until I fix the test) |
Testsuite fails on i386:
The text was updated successfully, but these errors were encountered: