Andy Bradford wrote: > > I think this will work, but I still wonder if it used to work this way > before (e.g. test results ended up in current working directory, not > elsewhere), why it should not continue to work that way? The reason why > I suggested [pwd] is because that is how it currently works (on trunk). >
The way trunk works is probably not the best model. It requires the user running the test to actively take steps to avoid polluting/mangling the Fossil checkout while testing. > > Also, I look at it from a different perspective. Having a tool write > things in a magical location that is *not* my current working directory > is odd behavior. For example, if I type ``vi somefile.txt'', I know it > will produce that file in my current working directory. If I type > ``fossil --httptrace sync'' it it will produce HTTP request/reply things > in my current working directory, etc... Why would tester.tcl behave > differently? > Lots of tools on both Windows and Unix (including 'vi'?) write temporary files (and folders) to a temporary directory. With the new handling on the testerCleanup branch, these files and folders are written and cleaned up for each test file, in a "fail-safe" manner. -- Joe Mistachkin _______________________________________________ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users