The following 'hash -d' statements have different exit statuses.
bash --noprofile --norc -c \
'hash -d ls; echo $?' ## 0
bash --noprofile --norc -c \
'cat </dev/null; hash -d ls 2>/dev/null; echo $?' ## 1
This depends on a hashable command being given (e.g. cat </dev/null)
prior to the 'hash -d'.
Bash' documentation says:
"The return status is true unless a name is not found ..."
So, I'd say, both commands should have exit status 1 - as 'ls' was not hashed.
POSIX (https://pubs.opengroup.org/onlinepubs/9699919799/utilities/hash.html)
only has -r as option.
FWIW:
https://github.com/search?q=language%3AShell+content%3A%2F%28%5E%7C%5B%5B%3Ablank%3A%5D%5D%29hash%5B%5B%3Ablank%3A%5D%5D-d%2F+NOT+zsh&type=code
Simple patch attached, making 'hash -d ls' exit with 1 in the first case also.
But ...
Questions
1) Why is 'hash -d x' exiting with 1 at all? If the goal is to `unhash x',
'hash -d x' will do the job. Even if x was not hashed at all.
But I am not sure that such a change is feasible (backwards compatible).
2) 'hash x' and 'hash -l x' seem to do the same
(i.c. path search and add to hash if found).
It would be more useful to let 'hash -l x' give the 'hash -l' output for x.
Moreover, I find it unexpected that 'hash -l x' adds x to the hash.
A quick look on github, shows no use of 'hash -l x',
so changes might be ok.
'hash -lt x' outputs what I would have expected 'hash -l x' to do.
Note that 'hash -lt x' doesn't add x to the hash (unexpected).
3) These option combinations should error/exit 1:
Ignoring the '-p path' part:
hash -p path
hash -p path -t name
hash -p path -l
hash -p path -r
This adds name, using path, but ignores the other option:
hash -p path -l name
hash -p path -d name
Using -t ignores -d:
hash ls; hash -td ls
And -d ignores -l:
hash ls; hash -ld ls
Conclusion
I would say that:
- Combining options should error;
- Except for an added -r, which first clears;
- 'hash -l x' shouldn't add, but print what 'hash -lt x' currently prints;
- 'hash -d x' always exits 0, quietly.
Let me know if any of this stuff is wanted, I think I can implement this.
--
Regards, Mike Jonkmans
hashcmd.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git i/hashcmd.c w/hashcmd.c
index ebb92346..62e080be 100644
--- i/hashcmd.c
+++ w/hashcmd.c
@@ -66,9 +66,12 @@ phash_remove (const char *filename)
{
register BUCKET_CONTENTS *item;
- if (hashing_enabled == 0 || hashed_filenames == 0)
+ if (hashing_enabled == 0)
return 0;
+ if (hashed_filenames == 0)
+ return 1;
+
item = hash_remove (filename, hashed_filenames, 0);
if (item)
{
