summary refs log tree commit diff
path: root/FEATURES
diff options
context:
space:
mode:
authorMikael Magnusson <mikachu@gmail.com>2014-11-06 16:06:23 +0100
committerMikael Magnusson <mikachu@gmail.com>2014-11-07 11:10:30 +0100
commita9004ce32f79f2244f0882bd0e69248ea0c9c791 (patch)
tree9cf4e4131ad0132d07c5e5b0bb48e51a2989d84d /FEATURES
parent7abd611a2396bad9d93d18681a2c59cb1ea0e158 (diff)
downloadzsh-a9004ce32f79f2244f0882bd0e69248ea0c9c791.tar.gz
zsh-a9004ce32f79f2244f0882bd0e69248ea0c9c791.tar.xz
zsh-a9004ce32f79f2244f0882bd0e69248ea0c9c791.zip
33607: Fix some minor problems in zattr module
Remove the listattr call in zgetattr, it only caused no error to be output
when trying to retrieve an xattr from a file with no xattrs. When a file
had xattrs, it would just add an extra syscall for no good reason.

Always set an array parameter in zlistattr, this makes the returned value
much easier to use; in fact the _zattr completer didn't account for this
and zlistattr foo <tab> on a file with only one attribute did not work.

Almost all of the patch is only reindent, the only modification that's
not a pure deletions is:
 -        ret = 1 + (attr_len > val_len || attr_len < 0);
 +        ret = 1 + ((val_len > 0 && attr_len > val_len) || attr_len < 0);
which makes sure we return the correct error in the new path due to the
removed listattr call. (If val_len is -1 due to no attribute existing,
it doesn't mean the user should retry the call because the attribute
grew in size).
Diffstat (limited to 'FEATURES')
0 files changed, 0 insertions, 0 deletions