tag:blogger.com,1999:blog-38606689.post2798330986836934240..comments2023-03-25T06:12:33.283-07:00Comments on Xoatlicue: A Systematic Look at Testing, PerhapsTau Centralhttp://www.blogger.com/profile/04639944664204798442noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-38606689.post-20143508393339890542007-12-05T14:24:00.000-08:002007-12-05T14:24:00.000-08:00On the trunk: except for the test added after this...On the trunk: except for the test added after this regression, I can only see a single test case which uses drawImage, and it doesn't even use the drawImage(HTMLImageElement, ...) form which was broken. And that test wasn't written by Mozilla developers.<BR/><BR/>(I've posted some more comprehensive tests <A HREF="https://bugzilla.mozilla.org/show_bug.cgi?id=407049" REL="nofollow">here</A> in the hope that it will help prevent future canvas bugs, though that doesn't solve the general problem for the rest of the browser.)Philip Taylorhttps://www.blogger.com/profile/13121419389252492670noreply@blogger.comtag:blogger.com,1999:blog-38606689.post-58447327478398288982007-12-05T11:43:00.000-08:002007-12-05T11:43:00.000-08:00Well, specifically in this instance, there are aut...Well, specifically in this instance, there are automated tests for the failure that caused 2.0.0.11 to be created. The problem is that these reference tests, which were created by developers working on Canvas (I believe), only run on the trunk, which is Firefox 3. The failure was in a single method that Canvas uses.<BR/><BR/>Most of the ongoing work to improve unit tests and automated coverage has been future focused, so it has happened in the Firefox 3 space. That's where the vast majority of development work is occurring. That's not a justification but simply a statement of fact.<BR/><BR/>We're looking at ways to address this issue but don't have a short-term solution.Alhttps://www.blogger.com/profile/04561666880371974930noreply@blogger.com