Re: [PATCH 00/42] repo setup test cases and fixes
- From
Jonathan Nieder <jrnieder@gmail.com>
- Date
- Oct 29, 2010, 17:09 UTC
- Message-ID
- <20101029170906.GA29249@burratino>
- In-Reply-To
- <AANLkTikJJnTXfWXKe5bm_Qyjmgna9g3vChdkPzoiAb4i@mail.gmail.com>
Sverre Rabbelier wrote:
> 2010/10/28 Nguyễn Thái Ngọc Duy <pclouds@gmail.com>:
Show 5 quoted lines
>> This series includes setup coverage tests (180/376 fail). Then the fixes, >> which bring down to zero failed test in the end. 02/42 describes the >> rules. New rules are: > > Shouldn't the fixes come first (for bisectability?).
Good point. Ideally the tests would come first if the fixes are contraversial, or be squashed with the fixes if neither is contraversial, or after if the tests are contraversial but the fixes are not.
For bisectability and clarity, the tests should use test_expect_failure when they fail (like this patch series does) and change that to test_expect_success in the patch that fixes them.