johnstonskj
2019-2-11 16:54:06

@johnstonskj has joined the channel


bkovitz
2019-2-11 18:49:25

In Typed Racket, how can a macro find out the type of an identifier?


soegaard2
2019-2-11 19:03:11

Not a beginner question :wink:


samth
2019-2-11 19:06:59

@bkovitz you can’t do that


greg
2019-2-11 19:33:08

@robert.postill One common way to run tests is to put them in test submodules. By default raco test finds and runs test submodules. However you can put other, optional/slow/whatever tests in submodules named whatever you want, and tell raco test to run those submodules using the -s command-line flag.


greg
2019-2-11 19:33:52

Although I don’t know if this is a great example, I did this with my markdown project, and for the reason you said (to avoid running some tests on CI): https://github.com/greghendershott/markdown/blob/master/Makefile#L37-L46


greg
2019-2-11 19:35:37

I don’t bother with Racket test suites, so idk if this applies to what you’re doing.


greg
2019-2-11 19:37:47

Absent test suites, there’s really not too much special about tests vs. non-tests, in Racket. There is just code you run or do not run at different times. You can put tests in their own files, and run them or not. What’s neat about test submodules is you can mix things in the same file, and it only gets loaded at specific times: “compile time”, “run time”, “test time”, “doc time”, etc.


greg
2019-2-11 19:40:15

Of course people have opinions about testing. :slightly_smiling_face: I’m normally in the rackunit camp, which is “no news is good news”. I only want to hear about failures, and act on them. Sort of like “management by exception”. However, another valid idea is to want affirmative feedback on how many tests ran, and passed, too. I don’t do that with Racket so don’t have much advice to offer about that. Someone else probably can, if that’s more your preference.


bkovitz
2019-2-11 21:30:31

@samth OK, that answers that question.


samth
2019-2-11 21:48:31

@greg @robert.postill note that raco test by default says how many tests passed


philip.mcgrath
2019-2-11 22:01:52

Are there any known quirks to how ffi-lib opens DLLs on Windows? I can build libgit2 and run its test suite successfully, but when I try to open it with ffi-lib, it fails with the error %1 is not a valid Win32 application.; errid=193.


samth
2019-2-11 22:03:07

@philip.mcgrath I doubt that’s a known quirk


philip.mcgrath
2019-2-11 22:52:12

@samth I expect it’s something I’m not doing properly, but I don’t know what: I don’t have a Windows dev box locally (I’m building this on AppVeyor), and I haven’t really used Windows in over a decade. (Also, I meant to post this on #general … oops.)