On 3/13/07, Han Boetes <[EMAIL PROTECTED]> wrote: ...
As the errormessage suggests there is a character in a filename somewhere on my filesystem which updatedb doesn't dig.
The cited error message, "locate database header corrupt, bigram char outside 0, 32-127: 14" *only* indicates that the locate database does not match the format expected by locate. Given that the locate database format involves 'front compression' (the code and published reference can be found in /usr/src/usr.bin/locate/code/locate.code.c), I don't see any particular reason to believe that the problem stems from a filename on your system that contains an unusual character. If I was trying to track this down, I would probably walk 'locate' under gdb to find the problem spot in the database and then study the hexdump of the area against locate.code's source to determine how it might have been generated, with an eye out for obvious stuff like whether the error is on a block boundary or the bad data looks like the output of any programs I know, etc. Or easier, since your said this problem has been around for more than a week, it would appears to be presist past the normal weekly rebuild of the database. If so, simply walking local.code through its passes and checking the output as it goes is probably a faster attack on the issue. Philip Guenther
