| 13 Aug 2025 |
 Alyssa Ross | With strictDeps I believe checkInputs don't end up in PATH, only nativeCheckInputs | 18:08:34 |
 emily | I think we have a HOST_PATH thing or something. | 18:08:50 |
| emily | for patchShebangs --host I guess. | 18:09:05 |
| emily | but yeah cross testing is a mess. even with canExecute. | 18:09:11 |
| emily | tbh I think it's fine to use whatever variable is convenient and leave it to the future generation that decides it wants to figure out how cross tests should work. | 18:09:24 |
| emily | (to which the answer will be "separate installCheckPhase derivation" if we raised them well) | 18:09:36 |
 Artem | to be clear: Z3 is called by the test suite when the test suite is run | 18:12:10 |
| Alyssa Ross | that's a nativeCheckInputs case then, definitely | 18:13:32 |
| Artem | simply replacing testSystemDepends with nativeCheckInputs here doesn't work. I'm happy to do anything before I open a PR if someone tells me what :-) otherwise I proceed with this | 18:14:25 |
| Artem | * simply replacing testSystemDepends with nativeCheckInputs here doesn't work. I'm happy to do anything before I open a PR if someone tells me what :-) otherwise I proceed with testSystemDepends | 18:14:32 |
| Alyssa Ross | It sounds like there needs to be another builder option, testSystemNativeDepends or some less terrible name, that maps to nativeCheckInputs | 18:16:55 |
| emily | (I don't see how. if you're cross-compiling a system that uses this thing, then it's going to invoke the hostPlatform Z3 when you actually use it) | 18:18:32 |
| Alyssa Ross | But it's the test suite that runs the binary | 18:18:49 |
| emily | (testing it against the build platform Z3 is strictly less realistic for testing whether the thing you are building will actually work) | 18:18:50 |
| Alyssa Ross | Not the program itself | 18:18:54 |
| Alyssa Ross | Is what we were told | 18:18:56 |
| emily | yeah, and the test suite runs on the host platform | 18:18:59 |
| emily | via the magic of canExecute | 18:19:08 |
| emily | it's testing a host platform thing | 18:19:14 |
| emily | not a build platform thing | 18:19:17 |
| Alyssa Ross | Then what would nativeCheckInputs ever be for? | 18:19:29 |
| emily | make test builds a buildPlatform test runner utility that has a binary dependency or something. | 18:19:59 |
| emily | maybe that dependency is GCC that targets hostPlatform so it can dynamically compile some test code before running it or something. | 18:20:18 |
| emily | like I think the same rule of thumb of "tools used by the build system" applies. | 18:21:06 |
| Alyssa Ross | I guess I'd consider test framework part of the build system | 18:21:58 |
| Alyssa Ross | But really it depends for what purpose it's invoking this tool | 18:22:22 |
| Alyssa Ross | Is it to generate test data? If so, build platform | 18:22:29 |
| emily | it's a Haskell library that provides an interface to an SMT solver | 18:22:35 |
| emily | you write Haskell code and it invokes Z3 at runtime to solve SMT things | 18:22:41 |
| emily | if Z3 was a library, it would be a buildInput | 18:22:48 |
