Hi all,
I've tested the attached patch, which seems to do the right thing. Can I
ask for a review and application please?
This does not fix the issue with no result row being returned - that looks
like a separate problem to me.
Thanks, Shaheed
On 25 January 2012 18:04, Shaheed Haque <srha...@theiet.org> wrote:
> Hi,
>
> I'm having a little difficulty with the usage of nspi_SeekEntries. I an
> calling it apparently successfully in that it does not return an error, but
> a subsequent call to GetGalTable always seems to return data from the
> beginning. FWIW, the code is equivalent to:
>
> key.ulPropTag = (MAPITAGS)PR_DISPLAY_NAME_UNICODE;
> key.dwAlignPad = 0;
> key.value.lpszW = talloc_strdup(mem.ctx(), displayName.toUtf8().data());
> if (MAPI_E_SUCCESS != nspi_SeekEntries(nspi, mem.ctx(),
> SortTypeDisplayName, &key, tags, NULL, results)) {
> error() << "cannot seek to GAL entry" << displayName << mapiError();
> return false;
> }
> if (MAPI_E_SUCCESS != GetGALTable(m_session, tags, results,
> requestedCount, TABLE_CUR)) {
> error() << "cannot read GAL entries" << mapiError();
> return false;
> }
>
> [ Interestingly, note that the "results" value from the first call is
> always NULL, rather than the relevant row, but I'm not sure if that is
> another problem or not. ]
>
> Now, looking at the implementation of GetGALTable, when TABLE_CUR is
> specified, I think it expects to use the values previously stored in
> the nspi->pStat to do the fetch. And indeed, after the fetch, it updates
> the stored value. But AFAICS, nspi_SeekEntries does not update the stored
> value - that seems like the problem to me. In other words, the bottom of
> this function should look like the bottom of GetGALTable:
>
> OPENCHANGE_RETVAL_IF(!NT_STATUS_IS_OK(status), retval, NULL);
> OPENCHANGE_RETVAL_IF(retval, retval, NULL);
>
> + nspi_ctx->pStat->CurrentRec = r.out.pStat->CurrentRec;
> + nspi_ctx->pStat->Delta = r.out.pStat->Delta;
> + nspi_ctx->pStat->NumPos = r.out.pStat->NumPos;
> + nspi_ctx->pStat->TotalRecs = r.out.pStat->TotalRecs;
>
> return MAPI_E_SUCCESS;
>
> Assuming that sounds reasonable, would a patch be acceptable?
>
> Thanks, Shaheed
>
>
diff -p -r -C 5 openchange-0.11-TRANSPORTER/libmapi/nspi.c openchange-0.11-TRANSPORTER-fixed/libmapi/nspi.c
*** openchange-0.11-TRANSPORTER/libmapi/nspi.c 2011-06-19 01:47:21.000000000 +0100
--- openchange-0.11-TRANSPORTER-fixed/libmapi/nspi.c 2012-01-25 21:39:35.120230908 +0000
*************** _PUBLIC_ enum MAPISTATUS nspi_SeekEntrie
*** 377,386 ****
--- 377,391 ----
status = dcerpc_NspiSeekEntries_r(nspi_ctx->rpc_connection->binding_handle, mem_ctx, &r);
retval = r.out.result;
OPENCHANGE_RETVAL_IF(!NT_STATUS_IS_OK(status), retval, pStat);
OPENCHANGE_RETVAL_IF(retval, retval, pStat);
+ nspi_ctx->pStat->CurrentRec = r.out.pStat->CurrentRec;
+ nspi_ctx->pStat->Delta = r.out.pStat->Delta;
+ nspi_ctx->pStat->NumPos = r.out.pStat->NumPos;
+ nspi_ctx->pStat->TotalRecs = r.out.pStat->TotalRecs;
+
return MAPI_E_SUCCESS;
}
/**
_______________________________________________
devel mailing list
devel@lists.openchange.org
http://mailman.openchange.org/listinfo/devel
