Re: completion inconsistencies involving symlinks

2006-09-09 Thread Chet Ramey
[EMAIL PROTECTED] wrote: >>> I assume my original bug is still open, though? >> >> I thought I answered that. If you prefer physical paths, use `set -o >> physical' to force them. > > I'm not sure I understand. Whether I want to use logical paths or not > seems independent of whether or not the c

Re: completion inconsistencies involving symlinks

2006-09-09 Thread nu6ohq602
I assume my original bug is still open, though? I thought I answered that. If you prefer physical paths, use `set -o physical' to force them. I'm not sure I understand. Whether I want to use logical paths or not seems independent of whether or not the command I'm currently typing only resp

Re: completion inconsistencies involving symlinks

2006-09-09 Thread Chet Ramey
[EMAIL PROTECTED] wrote: > I assume my original bug is still open, though? One solution would be to > be able to specify logical vs physical pathnames in programmable > completion. I thought I answered that. If you prefer physical paths, use `set -o physical' to force them. The completion code

Re: completion inconsistencies involving symlinks

2006-09-09 Thread nu6ohq602
The shell's concept of a logical view of the file system (the default, as specified by POSIX) is maintained by `cd' and communicated to other processes via $PWD. Other processes may or may not choose to use that information. There are environments where support for $PWD is pervasive. Ah, th