How to inspect the sandbox for cram tests #11167
Blaisorblade
started this conversation in
General
Replies: 1 comment
-
@Alizter I think you've added a setting to keep a sandbox around if a rule fails? Shall we bring that back perhaps. I think we can just go ahead with something like that except instead of keeping the sandbox spot, we can move it into some safe space for such broken sandboxes. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Cram tests run into a temporary sandbox folder that gets deleted — is there a way to prevent deletion, to allow inspecting the sandbox or even modifying it (when debugging the test)?
In my case, my cram tests generates some file
foo.v
, and then compiles it with Coq. I'm getting unexpected errors so I'd like to inspectfoo.v
— the only way I found is to addcat foo.v
to my cram test, use promotion to updaterun.t
, then copy-paste the output out of the.t
. But this seems hacky, and would get worse if I had multiple generated files, or some of them were binary (and yes, something like base64 encoding could make the hack applicable there too).Or I can extract the commands from the
run.t
file and run them outside — again, a hassle if there are many commands.All this gets worse if the problem isn't repeatable — but I wouldn't restrict a solution to failing tests: I might want to inspect the sandbox even when the test succeeds.
EDIT: I already tried
--sandbox=none
— but I can't find outputs, and ifrun.t
needs changes, the diff shows that the test still runs under_build/.sandbox
. Example:Beta Was this translation helpful? Give feedback.
All reactions