On Tue, May 21, 2002 at 11:29:24PM +0200, Lars Weber wrote: > All this talk about reasons for using `/hurd' got me wondering: Do there > exist potential problems when a translator that translates a certain > directory is itself located somewhere inside that directory?
No, because if you exec a file, you "open" it (what the Hurd equivalent of the open it), and only pass the descriptor to the exec server (actually the Hurd equivalent of a descriptor). So, you keep an active reference to the file, and always talk to the "underlying" translator. There are issues with executing script. I think for scripts it is the normal case that they are reopened by name, but if that is not possible, they are reopened by giving the interpreter the name /dev/fd/NR > This might not be the case, but if it is I think this would be reason > enough to use a separate directory. /hurd is provided by a translator as well, and there can be translators on / or /hurd, so that doesn't save you. If you start to talk about the details it is quickly getting pretty weird :), but for example the bootstrap is done by GRUB, which can read filesystems itself, so it reads /hurd/ext2fs from the disk directly, and loads that into memory, in addition to the microkernel and the run time linker. Thanks, Marcus -- `Rhubarb is no Egyptian god.' Debian http://www.debian.org [EMAIL PROTECTED] Marcus Brinkmann GNU http://www.gnu.org [EMAIL PROTECTED] [EMAIL PROTECTED] http://www.marcus-brinkmann.de -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
