clean compile with compiler flags /WX /W4

[frederich]

Dear Tobias,

I use /WX and /W4 in my projects. This commit tidies up some minor compile issues. The boost::ignore_unused might be replaced with a better VERIFYEQUAL mock-up. In general, I think we should add proper default implementations for VERIFYEQUAL, _ASSERTPRINT instead of empty mock-ups.

Best Jens

[frederich]

Tobias, any comment?

[treh]

Jens, Could you resolve the conflicts arising from the latest update?

[frederich]

Sure

[frederich]

Tobias, please take a look.

I'd prefer the PR #19 over #17, because there are less foul boost::ignore_unused calls.

[frederich]

Ping, all conflicts resolved.

[treh]

Woudn't it be better to just exclude the unit tests (*.t.cpp) from the release build?

[schoedl]

We probably want to test Release, too.

-- Dr. Arno Sch�dl | schoedl@think-cell.com Technical Director

We are looking for C++ Developers: https://www.think-cell.com/career

---

think-cell Software GmbH https://www.think-cell.comhttps://www.think-cell.com/ Chausseestr. 8/E phone / fax +49 30 666473-10 / -19 10115 Berlin, Germany US phone / fax +1 800 891 8091 / +1 212 504 3039 Amtsgericht Berlin-Charlottenburg, HRB 180042 | European Union VAT Id DE308385481 Directors: Dr. Markus Hannebauer, Dr. Arno Sch�dl On 2 Jan 2017, at 16:32, Tobias Reh notifications@github.com<mailto:notifications@github.com> wrote:

Woudn't it be better to just exclude the unit tests (*.t.cpp) from the release build?

� You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/think-cell/range/pull/17#issuecomment-269979684, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AD_nCwXm-qANpHneUoMMwQguFkN9QO94ks5rOQpIgaJpZM4K9w4t.

[frederich]

PR accepted or shall I rework VERIFYEQUAL to avoid boot::ignore_unused?

[treh]

Release build should either have a meaningful _ASSERT or skip the unit tests. Only silencing the warnings (by boost::unused_arg or by static_cast does not matter) is not useful.

[frederich]

Personally, I prefer to keep asserts in Release, thus meaningful _ASSERTs. I'll send a proposal.

[frederich]

BTW, union_range_test fails in Release mode

Running Unit test 'union_range_tests'...
Fail: Ranges differ:
Expected: [2, 16975008, 3, 4, 4, 32, 47, ]
Is      : [1, 1, 1, 2, 3, 3, 4, 4, 4, 5, ]
Running Unit test 'union_range_generator' ...
Running Unit test 'sub_range_with_tranform' ...
Running Unit test 'Unique_Ranges' ...
Fail: Ranges differ:
Expected: [1, ]
Is      : [-20094001, 15727176, 19268344, 0, 15727124, 19193958, 19153555, 19153555, 13225984, 19013152, -20094941, 15727192, 19153311, 19272000, 19272508, -20094865, 19153555, 19153555, 13225984, 0, 19153555, 2376339, 15727140, 0, 15727268, 19170256, -16415617, 0, 15727212, 2008638148, 13225984, 2008638112, -2010548387, 15727284, 2010451929, 13225984, -2012604734, 0, 0, 13225984, 0, 0, 0, 0, -2012604734, 15727224, 0, 15727292, 2010539952, -12883098, ]

[treh]

What's your test setup? Note that VS 2015 Update 1,2,3 are buggy and have code generation errors. Please use either VS 2015 without Update or VS 2017 RC.

[frederich]

Ah, good to know, thanks! My current setup is VS 2015 Update 3. I'll upgrade soon.