Hi, On Tue, Feb 24, 2009 at 12:24 PM, Nick Burch <[email protected]> wrote: > On Tue, 24 Feb 2009, Yegor Kozlov wrote: >> Another potential issue are our test files. Many of them were uploaded in >> Bugzilla and then included in svn. > > I'm tempted to say we ask around at apachecon, and find out what other > projects do wrt test files. Once we have an idea of what everyone else does, > bring those back to the list, and we can pick an outcome
I've recently worked on this same issue with the PDFBox project in the Apache Incubator. PDFBox has a number of test files that have been submitted as examples in the issue tracker with little concern over the copyright or licensing status. See https://issues.apache.org/jira/browse/PDFBOX-391 and http://markmail.org/message/cuyylr6zqs4fwdiz for more background. The situation is a bit better for Apache POI, at least for any test files contributed after POI switched to ALv2, as the section 5 of the license gives a good baseline interpretation for the licensing of contributed files. So if a file has been submitted through Bugzilla and contains no obviously troublesome content, then it's probably fine licensing-wise. However, many users don't really care about licensing that much, and can just submit any documents that they are having problems with. For example PDFBox contains the entire JavaMail specification and an IRS tax guide as test files. To best way to deal with such cases is to check the licensing status already before the file is added to svn. If a user submits a test document that looks like something that they may not have sufficient rights to allow us to redistribute the file as a part of POI, then it's best to ask for clarification or an alternative test file if possible. It would also be a good idea to review the test files we already have in svn. Any files with an unclear licensing status are probably better dropped from official releases. In PDFBox we've decided to create a separate test package that contains all the troublesome test files. This package will not be officially released or even kept in Apache svn. But we will make it informally available to committers and other interested developers for use as a part of the PDFBox test suite until better test cases are created. Perhaps a similar approach would work also for POI in case such troublesome test files are found. >> - programmatically insert the ASF license, for example, in the Comments >> OLE property. Create an utility to audit test files before release. > > My slight worry is that for the files where the bug report relates to iffy > ole2 blocks / properties / etc, that'll break/fix them Yeah, it's probably not a good idea to modify the test files. BR, Jukka Zitting --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
