James Carlson wrote: > > I was commenting on Ali's seeming misunderstanding of my original > comment; he brought up adding -w enhancements, when I was merely > saying that renaming files (if that's how the changes are wrought) > shouldn't be a serious concern. >
You're right --- it never occurred to me that renaming the mapfiles would be a problem, so I assumed you were talking about the post-build tests for nightly that we had discussed earlier. You're also right that my exception list turns out to be comically small. I spent some time this afternoon finalizing the list, and it turns out to be about 5 lines. If I were to include the mapfiles in ON that don't specify versioning to this, I might end up with 10. However, I want those files to have the comment anyway. So really, it's nothing. With some minor effort, I could make the mapfilechk exception list go away entirely. However, I do like it better this way, and as you and Mark seemed to be in agreement that a general exclusion mechanism for the 'hg' tools has some value, I'm going to go ahead with it. --- Mark J. Nelson writes: You're now talking about checking the shared objects, instead of the > mapfiles. Seems reasonable, but I still would like mapfile changes to > be more directly tested, even when wsdiff is not invoked. I spoke with Mark about this earlier this afternoon. His feeling is that gatelings are more likely to run check_rtime in their test builds than wsdiff. Putting it in wsdiff would probably result in the gatekeeper being the first one to discover a problem, *after* integration, which is better than not finding it at all, but not by much. So, I'll look into doing something that is triggered with the other ELF checks (-r nightly flag). I'll also look at wsdiff, because it seems worthwhile. - Ali
