"David Dyer-Bennet" <d...@dd-b.net> wrote: > > Inflammatory, maybe. Incorrect, I don't think so. Joerg has made a > > fairly convincing, and consistent, point of showing where the > > reliability problems with GNU tar lie - in bugs that were filed 16 > > years ago. You even acknowledged them in your response. Out of context > > editing of his posts do not help your argument. > > I haven't looked at the bug he cites. What I do know is that this > discussion is the first time I've heard of it, and I've been using various > tar implementations pretty frequently since, oh, let's say 1986
This is an intersting claim..... In 1986 it was pretty impossible for nearly all people to play with "different tar implementations". Star exists for a while but it has not yet ben published to a greater audience, star was only available to people who had relations to H. Berhold AG or did buy a Sun from them. GNU tar was not existent. PD tar was shown by John Gilmore at USENIX in February??? 1987 and as SUG-tar in December 1987 at the Sun user group meeting. I intended to puiblish star at the SUG-meeting in December 1987 but for some reasons did not. A noticable amount of people was not able to play with different tar implementations before PD-tar aka. SUG-tar was published as GNU tar in 1989. At that time GNU tar was a cleanly written piece of software.... Various maintainer changes caused by Stallman happened before 1993 and in this time the source became harder and harder to read. The fact that GNU tar does not implement a clean and simple basic extract loop but lengthy unstructured code that is used for various tasks makes it unreliable. I made a related bug report in 1993 but this was ignored completely for 10 years! Later, people did offer "fixes" from time to time and it seemed that thex have been helpful but soon other contraints have been found that triggered the same problem. GNU tar, in case that minor problems happen during extract, has a high risk of not extracting the whole archive. Last seen last week with GNU tar-1.19. > (previously I played on Unix a bit, but never worked on it). So his > apocalyptic claims ring strangely hollow. I'm sure the bug is real, and > should be fixed; but I can't find myself considering a tool "broken" > because there's a 16-year-old bug that I've never encountered or even > heard of previously. GNU tar is a real problematic piece of software..... If you are using a different ls implementation that makes problems, you see this when you call it. If you use a tar implementation that does not create standard compliant archives, you see the problem when it is too late. Did you ever read the text in ftp://ftp.berlios.de/pub/star/README.otherbugs It is from 2001, so some of the problems have been fixed meanwhile but it gives a good overview to help to understand what kind of problems you may see. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de (uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
