I finally put my FFI testsuite into CVS.
There are three known areas causing it to fail, so please don't report =
them again (unless you solved an issue :-).
The testsuite will even cause CLISP to dump core! (Issue 3)
1. one validp/foreign-value issue
I'll have to add an unwind-protect calling mark-invalid to =
with-foreign-object, as I already did for with-foreign-string.
2. errors involving c-struct
Somebody modified C-STRUCT and it does not seem internally consistent =
Had I submitted my testsuite earlier this would have been noticed back =
3. Errors & crash involving c-array and c-array-max with character on =
See bug "cast to char array crashes"
Work-around so far: compile foreign.d using -O0 -falign-functions=3D4
Or add some printf() amid convert_from_foreign_array_fill(!)
From: Sam Steingold <sds@gn...> - 2003-04-30 14:21:44
> * In message <9F8582E37B2EE5498E76392AEDDCD3FE03ECF309@...>
> * On the subject of "why the new FFI testsuite fails - glad there is one"
> * Sent on Wed, 30 Apr 2003 09:56:14 +0200
> * Honorable "Hoehle, Joerg-Cyril" <Joerg-Cyril.Hoehle@...> writes:
> I finally put my FFI testsuite into CVS.
apparently, too early.
> The testsuite will even cause CLISP to dump core! (Issue 3)
the restsuite is for regression testing, not for bug indication.
the general rule is that _before_ "cvs ci",
you must make sure that "make check" passes,
at least on your own platform.
please disable all tests that fail (or fix them).
PS. there is one test that have been failing for quite some time.
I put it there because Bruno promised to fix it.
I now disabled it.
Sam Steingold (http://www.podval.org/~sds) running RedHat9 GNU/Linux
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
Life is like a diaper -- short and loaded.