Enrico Forestieri wrote: > Been there, done that. The attached would be the patch you are looking > for, but it only fixes the assertion, not the other bug I mention. > This is because fileSearch is designed to replace an extension, not to > add one.
I don't think that was intended. I rather think that the problem is that changeExtension does not really know if a file has no extension. Maybe we should introduce aan addExtension and use that if we know that a filename has no extension? > Of course, it would be possible to modify fileSearch such that > it adds the extension when must_be_readable is false, but I find this > approach rather ugly. I agree. Nevertheless I like your second patch much more, because it makes the intention clear. In a second step we can introduce addExtension() and use that in fileSearch. Georg
