[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: (ITS#5408) back-ldif bugs



Another question:

How should backends handle the return value from, and rs->sr_err
to/from, send_search_entry() and send_search_reference()?  The
back-ldif handling differs from back-bdb/search.c.

At one point ldapsearch got insufficientAccessRights when an entry
_below_ baseDN in the search scope was ACL-protected.  But I can't
reproduce it, and don't know if the error was mine or back-ldif's.


Other details:

ldif_back_search() modified op->ors_scope.  That looked improper,
I removed it.

The ENOENT special case for rename() in spew_entry() made little
sense, since it was just moving the file it just created, and to
the same directory.  I deleted it, but I wonder why it was there.

I wrote:
>     + Sorting (in ldif_r_enum_tree()):
>       Should "a={1}x" sort before or after "a=x"?  Currently it sorts
>       like "a={".  "a=" (normally before) or "a=<CHAR_MAX>" (normally
>       after) would be better.

No, UCHAR_MAX.  strcmp compares the strings as unsigned char*.

-- 
Hallvard