diff options
author | Stavros Aronis <[email protected]> | 2010-06-18 03:44:25 +0300 |
---|---|---|
committer | Lukas Larsson <[email protected]> | 2011-02-18 12:03:18 +0100 |
commit | 98de31e836a04ccc8f5f9acd90b9ba0803a24ab5 (patch) | |
tree | 3f26237297b0b2d9040de1b97eeb7cd75bce2dfe /lib/dialyzer/test/small_tests_SUITE_data/src/trec.erl | |
parent | 08cec89bb1e781157a75c13e72562258b271b469 (diff) | |
download | otp-98de31e836a04ccc8f5f9acd90b9ba0803a24ab5.tar.gz otp-98de31e836a04ccc8f5f9acd90b9ba0803a24ab5.tar.bz2 otp-98de31e836a04ccc8f5f9acd90b9ba0803a24ab5.zip |
Test suites for Dialyzer
This is a transcription of most of the cvs.srv.it.uu.se:/hipe repository
dialyzer_tests into test suites that use the test server framework.
See README for information on how to use the included scripts for
modifications and updates.
When testing Dialyzer it's important that several OTP modules
are included in the plt. The suites takes care of that too.
Diffstat (limited to 'lib/dialyzer/test/small_tests_SUITE_data/src/trec.erl')
-rw-r--r-- | lib/dialyzer/test/small_tests_SUITE_data/src/trec.erl | 37 |
1 files changed, 37 insertions, 0 deletions
diff --git a/lib/dialyzer/test/small_tests_SUITE_data/src/trec.erl b/lib/dialyzer/test/small_tests_SUITE_data/src/trec.erl new file mode 100644 index 0000000000..b36b0cafba --- /dev/null +++ b/lib/dialyzer/test/small_tests_SUITE_data/src/trec.erl @@ -0,0 +1,37 @@ +%% +%% The current treatment of typed records leaves much to be desired. +%% These are not made up examples; I have cases like that the branch +%% of the HiPE compiler with types in records. I get very confusing +%% warnings which require a lot of effort to find their cause and why +%% a function has no local return. +%% +-module(trec). +-export([test/0, mk_foo_exp/2]). + +-record(foo, {a :: integer(), b :: [atom()]}). + +%% +%% For these functions we currently get the following warnings: +%% 1. Function test/0 has no local return +%% 2. The call trec:mk_foo_loc(42,any()) will fail since it differs +%% in argument position 1 from the success typing arguments: +%% ('undefined',atom()) +%% 3. Function mk_foo_loc/2 has no local return +%% +%% Arguably, the second warning is not what most users have in mind +%% when they wrote the type declarations in the 'foo' record, so no +%% doubt they'll find it confusing. But note that it is also inconsistent! +%% How come there is a success typing for a function that has no local return? +%% +test() -> + mk_foo_loc(42, bar:f()). + +mk_foo_loc(A, B) -> + #foo{a = A, b = [A,B]}. + +%% +%% For this function we currently get "has no local return" but we get +%% no reason; I want us to get a reason. +%% +mk_foo_exp(A, B) when is_integer(A) -> + #foo{a = A, b = [A,B]}. |