gnustep-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Test results from base


From: Fred Kiefer
Subject: Re: Test results from base
Date: Wed, 23 Mar 2011 10:48:02 +0100
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; de; rv:1.9.2.14) Gecko/20110221 SUSE/3.1.8 Thunderbird/3.1.8

On 23.03.2011 08:56, Richard Frith-Macdonald wrote:

On 21 Mar 2011, at 23:06, Fred Kiefer wrote:

My current test results are these:

   5547 Passed tests
     11 Dashed hopes
      8 Skipped sets
      5 Failed builds
      1 Failed test

The annoying bits are of course the failed builds. These are the new tests 
David just added. The problem here is that I don't have a compiler for ObjC++ 
installed. For me it is easy to tell that the failed builds are all caused by 
this. For others this would be a lot easier if these tests where grouped 
together. And it would be even better if these tests would be skipped if that 
compiler isn't available.

What I did for this yesterdays was:
1. added stuff in gnustep-make's configure to work-around a bug in autoconf 
where it was reporting the presence of g++ for CXX even when no working 
compiler existed!
2. fixed a bug in gnustep-tests so that it really does skip .mm files if there 
is no objective-c++ compiler
3. fixed those new tests to be bracketed with #if defined(GSTESTDEV) ... so 
they are only run in 'developer' mode .... these tests make no sense in general 
since they are using functionality which does not exist on the vast majority of 
systems, so we shouldn't be running them routinely.

Thank you, this seems to have fixed the issue.

The skipped tests are because I don't have ICU installed and the failed tests 
is the new test in NSNumberFormatter, which seems to be an actual bug in the 
code:
Failed test:       basic.m:31 ... Handle leading zeroes in fractional part: 1.1
Expected '1.1' and got '1.01'

Yes, that's a real bug ... I had a quick look at the code there and don't 
really understand what it's trying to do ... perhaps Stefan can manage a quick 
fix?

I had a short look at the code as well and to me the order of the statements looks wrong. We first strip off trailing zeros and then add zeros at the front until the requested length is reached. This is of course wrong. Maybe doing it the other way around would help?



reply via email to

[Prev in Thread] Current Thread [Next in Thread]