oysteikt/heimdal
0
0
Fork 0
Code Issues 5 Pull Requests Releases Activity

Fix ipropd-slave assert when doing del_enctype

Browse Source 
Change a paranoid heim_assert into something better.  Update block
    comment.

    Next commit: add a test for del_enctype and iprop.
This commit is contained in:
Nicolas Williams
2012-03-15 18:51:38 -05:00
parent ca6a22276e
commit ef9012aef5
1 changed files with 5 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
lib/kadm5/log.c
View File

@@ -701,26 +701,12 @@ kadm5_log_replay_modify (kadm5_server_context *context,
/*
* We don't need to do anything about key history here because
* we always log KADM5_TL_DATA when we change keys/passwords, so
* the code below this will handle key history implicitly.
* However, if we had to, the code to handle key history here
* would look like this:
*
* HDB_extension *ext;
* ...
* ext = hdb_find_extension(&log_ent.entry,
* choice_HDB_extension_data_hist_keys);
* if (ext);
* ret = hdb_replace_extension(context->context, &ent.entry, ext);
* else
* ret = hdb_clear_extension(context->context, &ent.entry,
* choice_HDB_extension_data_hist_keys);
*
* Maybe we should do this here anyways, wasteful as it would
* be, as a defensive programming measure? For now we heim_assert().
* the log entry contains a complete entry, including hdb
* extensions. We do need to make sure that KADM5_TL_DATA is in
* the mask though, since that's what it takes to update the
* extensions (see below).
*/
heim_assert((mask & KADM5_TL_DATA),
"Wouldn't log and replay key history");
mask |= KADM5_TL_DATA;
for (i = 0; i < ent.entry.keys.len; ++i)
free_Key(&ent.entry.keys.val[i]);