On Tue, 11 Oct 2022 15:17:13 GMT, Weijun Wang <wei...@openjdk.org> wrote:
>> Bill Huang has updated the pull request incrementally with one additional >> commit since the last revision: >> >> Implemented review comments. > > I have a question, why must this test be manual? Can't we compare the > localized texts? Hi @wangweij, I have a discussion with @naotoj about automating this test. Here is a summary. 1. We can't hard code localized texts into the test and compare to the localized message form the tool. The main reason is mentioned above by @naotoj. Another reason is that localization sometimes may come in slightly different, correctness of translation is not the goal of this test. 2. We can't do a word-to-word comparison. But even so, we can't rely on the same resource bundle used by the tools for localization validation. It provokes a risk that test may fail to detect any localization error in the tool as they use the same mechanism. 3. It makes a bit more sense for this test being a manual test because it matches the look-and-feel criteria. With someone looking into the result, at least it can raise a flag if something doesn’t look right. And only engineers can do this. ------------- PR: https://git.openjdk.org/jdk/pull/10635