On Wed, Dec 07, 2005 at 08:36:06PM -0800, Ingy dot Net wrote: > On 07/12/05 10:43 -0500, Patrick LeBoutillier wrote: > > Hi all, > > > > Development version 0.50_91 of Inline:Java has been released. Here is > > a summary of what has changed. > > > > Most importantly I had to drop support for perl 5.6 because there were > > some limitations I just couldn't get around anymore (mostly because > > Unicode support). > > Interesting. I support this decision as 5.6.x was never really very > good. I've noticed from bug reports on other projects that 5.8.0 5.8.1 > and 5.8.2 were buggy. Especially 5.8.0.
All versions of perl are buggy to a greater or lesser degree. Hopefully, the further down the maintenance track you get, the fewer bugs there are. > Almost makes me want to just require 5.8.3 and above for all my modules. > > Do folks think that maybe the new version of Inline could just support > 5.8.3 and above? In the docs we could just point users of older perls to > Inline-0.44. I went through this process with Devel::Cover. There is a hard minimum version requirement of 5.6.1, but I know of bugs in perl which cause problems for Devel::Cover up to 5.8.1, so I have set a recommended minimum version of 5.8.2. But I haven't enforced this. Sometimes upgrading just isn't an option. When this is the case there is already pain. I decided that as a module author I would try to share that pain, and support these lesser versions as best as I can. Without exception, Devel::Cover users who expressed an opinion would have been happy for me to bump the minimum version to 5.8, but I eventually decided not to. However, that was just my decision. I understand and support any author who says it is just too much work to support old versions. Are there specific bugs that concern you prior to 5.8.3, or does that version onwards just seem good enough. The primary impact for me is in testing, where I get different results from different versions, making the testing process more complicated and more time consuming. And when I get a bug report from one of these old versions, the first question is whether the problem exists in a more recent perl. -- Paul Johnson - [EMAIL PROTECTED] http://www.pjcj.net
