Hello, To begin with, thanks to everybody for expressing their thoughts on this matter, and apologies for my not being able to participate in the discussion earlier.
Now, let me provide some background on the record size warning. Historically, tar emitted a warning if the record size differed from the expected value and the verbose level was > 1 (e.g. `tar -tv'). This kind of warning is important for people using tape drives, as the proper record size diagnostics allows them to adjust their drives for optimal performance. This behavior was broken in tar 1.17, and it was not noticed until Ed Leaver reported it in his mail: http://lists.gnu.org/archive/html/bug-tar/2007-09/index.html (this posting also explains how and why the record size report is useful.) When the problem was fixed, Ed drew my attention to the fact that, being emitted in verbose levels > 1, this diagnostics is easily obscured by the copious output tar usually produces in these levels. After a bit of thought I decided to lower verbosity requirement by 1, which behavior made its way into tar 1.21 and subsequent 1.22. Now I can say that this was not the best decision, as the additional diagnostics broke some testcases and scripts. However, I am reluctant to revert to the verbosity>1 requirement, due to the reason explained above. I feel that the --verbose semantics is not clear enough in this context: using the same option to control detailed listing, produced on stdout, and to control additional diagnostic messages on stderr does not seem right to me. I think that an additional option, similar to gcc's -W, would be in place. For example, `tar --warning=record' would enable emitting the record size warning. Opinions? Regards, Sergey
