On Fri, Jun 27, 2008 at 7:16 PM, Adam Petaccia adam@tpetaccia.com wrote:
Subject: [Gdiplus 6/6] We're not checking to see if these functions succeed or not, they all just return Ok.
The point of regression tests is to prove that this is always the case...
On Fri, 2008-06-27 at 19:19 -0500, James Hawkins wrote:
On Fri, Jun 27, 2008 at 7:16 PM, Adam Petaccia adam@tpetaccia.com wrote:
Subject: [Gdiplus 6/6] We're not checking to see if these functions succeed or not, they all just return Ok.
The point of regression tests is to prove that this is always the case...
The GetGeneric* functions forward to CreateFontFamilyFrom name; I wrote the regression tests to make sure that we retrieve the correct font; Also, if one fails, it will be caught when the FontName string doesn't compare correctly.
On Sat, Jun 28, 2008 at 9:50 AM, Adam Petaccia adam@tpetaccia.com wrote:
On Fri, 2008-06-27 at 19:19 -0500, James Hawkins wrote:
On Fri, Jun 27, 2008 at 7:16 PM, Adam Petaccia adam@tpetaccia.com wrote:
Subject: [Gdiplus 6/6] We're not checking to see if these functions succeed or not, they all just return Ok.
The point of regression tests is to prove that this is always the case...
The GetGeneric* functions forward to CreateFontFamilyFrom name; I wrote the regression tests to make sure that we retrieve the correct font; Also, if one fails, it will be caught when the FontName string doesn't compare correctly.
Conformance tests are:
* independent of implementation. We don't know how native implements this function (and we don't really care), and there's no guarantee that our implementation of CreateFontFamilyName won't change.
* independent of other tests (unit). You're watering down the 'FontName string comparison' test by making it also rely on the results of the previous GetGeneric* functions as well. Each test tests one thing.
On Sat, 2008-06-28 at 10:25 -0500, James Hawkins wrote:
On Sat, Jun 28, 2008 at 9:50 AM, Adam Petaccia adam@tpetaccia.com wrote:
On Fri, 2008-06-27 at 19:19 -0500, James Hawkins wrote:
On Fri, Jun 27, 2008 at 7:16 PM, Adam Petaccia adam@tpetaccia.com wrote:
Subject: [Gdiplus 6/6] We're not checking to see if these functions succeed or not, they all just return Ok.
The point of regression tests is to prove that this is always the case...
The GetGeneric* functions forward to CreateFontFamilyFrom name; I wrote the regression tests to make sure that we retrieve the correct font; Also, if one fails, it will be caught when the FontName string doesn't compare correctly.
Conformance tests are:
- independent of implementation. We don't know how native implements
this function (and we don't really care), and there's no guarantee that our implementation of CreateFontFamilyName won't change.
Good point.
- independent of other tests (unit). You're watering down the
'FontName string comparison' test by making it also rely on the results of the previous GetGeneric* functions as well. Each test tests one thing.
Ok. I'll scrap this patch.
-
Adam Petaccia -
James Hawkins