/x*/ regexp should match an empty substring about v7

[vrz]

fixed

[mkmik]

Is there any test case that exercises this?

[cpq]

w.r.t. tests: the plan is to make VM to execute standard ecma262 tests.
Regexp tests will be covered by those.

@mmikulicic do you think we should add our custom hand-written tests as well?

[mkmik]

I agree we should avoid writing useless tests, but still it would be nice to associate a bug fix with a passing test that used to fail. It should be either possible to point out the ecma262 test chapter+number or write a short test case that covers the bug fix.

[cpq]

@vrz could you send a PR with the test case please - just add it to tests/unit_test.c?

[mkmik]

FYI, #30 reenables some of the pre-existing tests also fixed by this issue, which were disabled because we were having too many broken tests and we were becoming test blind.

[vrz]

PR with unit test is sent
FYI: I still have unit_test not working at lines:
121, 144, 148 -- ASSERT(check_num(v7, v, NAN));
Somthing not good with NAN definition.

[mkmik]

@vrz you mean that test is not working on current master ?
All test pass on mac and linux, see our CI server: https://circleci.com/gh/cesanta/v7
Does it mean that these tests fail on windows only?

BTW, you can easily run a build and test on a linux+gcc platform by installing docker for windows (https://docs.docker.com/installation/windows/) and then running "make docker".

[vrz]

Yes, tried on windows+gcc.

Fail on line 121: [check_num(v7, v, NAN)]

[mkmik]

Reproduced on vs2013, but line 121 works fine

  Fail on line 144: [check_num(v7, v, NAN)]
  Fail on line 148: [check_num(v7, v, NAN)]

[mkmik]

Caused by redefinitions of NAN done for earlier windows versions. Fix pending in #34.

Closing this issue since it has been fixed in 6888ceb and 7a4e832.

FYI: https://help.github.com/articles/closing-issues-via-commit-messages/