From: Uri Guttman >>>>>> "BM" == Bob McConnell <r...@cbord.com> writes: > > BM> From: Uri Guttman > > >> that implicates the tar.exe in your box as the guilty party. put > >> the blame on camelbox for not handling some variation in tar > >> headers that all other tar utils and modules can handle. report > >> this as a big bug > BM> to > >> them. > > BM> Sorry Uri, but I don't agree with any of that. The way I see it: there > BM> is a bug in the program building those tar files that is erroneously > BM> splitting the path between the prefix and name fields in the headers. > BM> The Archive::Tar module handles the resulting corrupted files, but it is > BM> still abuse of the file structure. If they were built correctly, the GNU > BM> tar (and WinZIP) would still be able to extract them. > > given the fact that my gnu tar on linux handled the tarball just fine > means the tarball is fine. that is a fact. perl's tar module handles it > fine. your tar from camelbox fails with it. those point to camelbox as > guilty, not anyone else. the tar header info may have changed in harness > releases but it obviously is supported by standard tar extractors so it > is not a broken tarball. camelbox may have coded their tar only to some > older or more narrow version of the tar header spec and that is what is > failing. > > BM> A Camelbox developer provided the workaround via their mailing > BM> list, so they are aware of it. But I haven't looked at this year's > BM> Summer of Code projects, so I don't know if they will be working > BM> on it or not. > > they know of the workaround since they know their tar is failing. no one > else is complaining to the harness author but you. they all must be > using working tar extractors (either a program or the perl module). so > they did nothing wrong and the harness tarball is good. i just use > sherlock holmes' rule here are eliminate all suspects and the one that > is left must be guilty. :)
You do know that Sherlock Holmes was a fictional character, don't you? If any of my diagnostics were that superficial, I would never have lasted 30 years in this profession, nor any other. A much better guideline would be one Bob Pease (an analog EE) used to say in Electronic Design and EDN magazines; "If you notice anything funny, record the amount of funny." Can you point me at any specification that states a tar archive may split the path between the two header fields? I haven't found one and until I see that, I will continue to believe the archives are corrupted, no matter how many utilities have implemented a workaround. Bob McConnell -- To unsubscribe, e-mail: beginners-unsubscr...@perl.org For additional commands, e-mail: beginners-h...@perl.org http://learn.perl.org/
