On Wed, 8 Jun 2022 16:59:52 GMT, Alexey Ivanov <aiva...@openjdk.org> wrote:
>> Thanks for catching that. Also, the reasoning for the test originally using >> a temporary file to create HTML was to have an HTML file that could be >> opened itself and compared towards. I see that it's not necessary for this >> test, but was useful at the time. I thought it could still be useful, but >> this test with the HTML stored in a string is probably sufficient. > >> the reasoning for the test originally using a temporary file to create HTML >> was to have an HTML file that could be opened itself and compared towards. I >> see that it's not necessary for this test, but was useful at the time. I >> thought it could still be useful, but this test with the HTML stored in a >> string is probably sufficient. > > It could still be useful. An option to the test, say `-saveHTML`, could be > used to get the HTML file to open in browser. > > Yet I think it's easier to run the test on its own when you're developing it, > no need to define the system property. At the same, I noticed a weird behaviour with the file. When I ran the test, class file was located in another directory but the HTML file was in the current directory. Calling `setPage` produced an exception. And `PassFailJFrame` started to behave erratically: it didn't close after clicking Pass or Fail. I didn't look into the reasons why. But it looks as bug in our manual testing framework… ------------- PR: https://git.openjdk.java.net/jdk/pull/7446
