Hi Jeremias, Jeremias Maerki a écrit : > Given so many regressions lately in the table layout code, wouldn't it
?! It’s debatable I guess, but I’m not sure something like 5 should be considered as that many, especially compared to the features added in the same time. > be worthwhile to construct a few more test cases? All following the > pattern: bug, create test case to reproduce, fix until test case passes. Yes, and that’s usually what I’m doing. I did consider writing a testcase for this one, but given the nature of the bug I didn’t feel like it. It’s better caught by a unit test rather than by a layout engine test, but we don’t have a unit test framework for the layout code. The bug appeared after a change in the logic of creating CellParts, now that it’s fixed it’s not likely to reproduce, even after further refactorings. And the method might well eventually disappear anyway. Vincent -- Vincent Hennebert Anyware Technologies http://people.apache.org/~vhennebert http://www.anyware-tech.com Apache FOP Committer FOP Development/Consulting
