adding one more case to ndr_pull_string_array/ndr_push_string_array

[Jelmer Vernooij]

On Sat, 2010-09-04 at 23:46 +0400, Matthieu Patou wrote:
> On 04/09/2010 22:42, Jelmer Vernooij wrote:
> > On Sat, 2010-09-04 at 19:54 +0400, Matthieu Patou wrote:
> >> It seems that I have a small pb with dfs
> >>
> >> The expanded names is an array of null terminated utf-16 strings but
> >> it's not null terminated that I declared like this:
> >>
> >>        [relative_short,subcontext(0),flag(STR_NOTERM|NDR_REMAINING)]
> >> string *expanded_names
> >>
> >> Although with this I'm able to parse message made from windows it causes
> >> a pb when s4 should act as a server for windows clients as the last
> >> expanded_names is not null terminated.
> >>
> >> If I set the time to be nstring then w7 is happy but I'm not able to
> >> parse the expoted bytes that I attached.
> >>
> >> I tried to use array to take advantage of the fact that I have the size
> >> with things like:
> >>
> >>
> >> [relative_short] nstring tab[size]
> >> or
> >> [relative_short, size_is(size)] nstring *tab
> >> or
> >> [relative_short, size_is(size)] nstring tab[]
> >>
> >> But either it didn't compile, or it didn't pass the test of ndrdump with
> >> the attached file.
> >>
> >>
> >>
> >> I propose to add a kind of exception like:
> >>
> >> diff --git a/librpc/ndr/ndr_string.c b/librpc/ndr/ndr_string.c
> >> index 2e04633..8de2257 100644
> >> --- a/librpc/ndr/ndr_string.c
> >> +++ b/librpc/ndr/ndr_string.c
> >> @@ -469,6 +469,11 @@ _PUBLIC_ enum ndr_err_code
> >> ndr_pull_string_array(struct ndr_pull *ndr, int ndr_f
> >>                           tmp_ctx = ndr->current_mem_ctx;
> >>                           ndr->current_mem_ctx = a;
> >>                           NDR_CHECK(ndr_pull_string(ndr, ndr_flags,&s));
> >> +                       if ((ndr->data_size - ndr->offset) == 0&&
> >> ndr->flags&  LIBNDR_FLAG_REMAINING)
> >> +                       {
> >> +                               a[count] = s;
> >> +                               break;
> >> +                       }
> >>                           ndr->current_mem_ctx = tmp_ctx;
> >>                           if (strcmp("", s)==0) {
> >>                                   a[count] = NULL;
> > Do we already generate correct data from the push function ? The
> > ndr_push/ndr_pull functions should be consistent in how they handle
> > flags.
> No that's why I made a more complete patch, as it is in my dfs branch on 
> git.samba.org.
Ah, great. This isn't particularly urgent, but something to watch out
for as it can be a source of confusion for others.

Cheers,

Jelmer
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20100905/ecc39ed1/attachment.pgp>