[svn-r10170] Purpose:

Bug Fixes

Description:
Fixes for several bugs, including dumping of excess output to a temporary file, fix for printing
hsize_t datatype, and the long awaited fix for intermixed output.

Solution:
Fix 1:  Overflow file
Previously, any output that a worker task made was buffered locally in memory, up to a point.  Any
output beyond the size of the buffer (used to be 10k) was discarded.  Now, the memory buffer size has been
changed to 1k and any output beyond this amount is sent a temporary file.  This way, no output is lost
and memory usage is kept under control.  The temporary file is deleted as soon as a worker task finishes
sending its contents to the manager.

Fix 2:  hsize_t printing
Printing of the hsize_t datatype used to be handled by %Hu passed to HDfprintf.  However, there is no corresponding HDvsnprintf that
is able to print hsize_t types.  These are now printed with the aid of H5_PRINTF_LL_WIDTH.

Fix 3:  Intermixed output fix
Intermixed output would occur on some machines (although I haven't seen it happen for a while) due to the unpredictability of the underlying network
and the speed at which various message would travel.  This has been fixed by having all output send to the manager
for printing.  The worker tasks no longer print the output themselves upon receipt of a token, but instead
send that data to the manager.



Platforms tested:
heping, eirene, tg-login (the only place that seems to still experience intermixed output every now and then)

Misc. update:

1 files changed, 38 insertions, 9 deletions

diff --git a/tools/h5diff/ph5diff_main.c b/tools/h5diff/ph5diff_main.c
index cd252b9..3d34af8 100644
--- a/tools/h5diff/ph5diff_main.c
+++ b/tools/h5diff/ph5diff_main.c
@@ -98,7 +98,7 @@ int main(int argc, const char *argv[])
 	MPI_Barrier(MPI_COMM_WORLD);
 
 	print_results(nfound, &options);
-	print_manager_output();
+	print_manager_output();  
 
 	MPI_Finalize();
 
@@ -131,8 +131,10 @@ ph5diff_worker(int nID)
     struct diff_args args;
     hid_t file1_id, file2_id;	
     char	filenames[2][1024];
+    char	out_data[PRINT_DATA_MAX_SIZE] = {0};
     hsize_t    nfound=0;
     MPI_Status Status;
+    int 	i;
 
     MPI_Comm_rank(MPI_COMM_WORLD, &nID);
     outBuffOffset = 0;
@@ -140,8 +142,6 @@ ph5diff_worker(int nID)
     MPI_Recv(filenames, 1024*2, MPI_CHAR, 0, MPI_ANY_TAG, MPI_COMM_WORLD, &Status);
     if(Status.MPI_TAG == MPI_TAG_PARALLEL)
     {
-	/*printf("We're in parallel mode...opening the files\n"); */
-
 	/* disable error reporting */
 	H5E_BEGIN_TRY
 	{
@@ -179,9 +179,38 @@ ph5diff_worker(int nID)
 
 		    /*Wait for print token. */
 		    MPI_Recv(NULL, 0, MPI_BYTE, 0, MPI_TAG_PRINT_TOK, MPI_COMM_WORLD, &Status);
-
-		    /*When get token, print stuff out and return token */
-		    printf("%s", outBuff);
+	
+		    /*When get token, send all of our output to the manager task and then return the token */
+		    for(i=0; i<outBuffOffset; i+=PRINT_DATA_MAX_SIZE)
+			MPI_Send(outBuff+i, PRINT_DATA_MAX_SIZE, MPI_BYTE, 0, MPI_TAG_PRINT_DATA, MPI_COMM_WORLD);
+
+			
+		    /* An overflow file exists, so we send it's output to the manager too and then delete it */
+		    if(overflow_file)
+		    {
+    			int	tmp;
+			memset(out_data, 0, PRINT_DATA_MAX_SIZE);
+			i=0;
+			
+			rewind(overflow_file);
+			while((tmp = getc(overflow_file)) >= 0)
+			{
+			    *(out_data + i++) = (char)tmp;
+			    if(i==PRINT_DATA_MAX_SIZE)
+			    {
+				MPI_Send(out_data, PRINT_DATA_MAX_SIZE, MPI_BYTE, 0, MPI_TAG_PRINT_DATA, MPI_COMM_WORLD);
+				i=0;
+				memset(out_data, 0, PRINT_DATA_MAX_SIZE);
+			    }
+			}
+
+			if(i>0)
+			    MPI_Send(out_data, PRINT_DATA_MAX_SIZE, MPI_BYTE, 0, MPI_TAG_PRINT_DATA, MPI_COMM_WORLD);
+
+			fclose(overflow_file);
+			overflow_file = NULL;
+		    }
+		    
 		    fflush(stdout);
 		    memset(outBuff, 0, OUTBUFF_SIZE);
 		    outBuffOffset = 0;
@@ -196,17 +225,17 @@ ph5diff_worker(int nID)
 		}
 		else
 		    MPI_Send(&nfound, sizeof(nfound), MPI_BYTE, 0, MPI_TAG_DONE, MPI_COMM_WORLD);
-
 	    }
 	    else if(Status.MPI_TAG == MPI_TAG_END)
 	    {
 		MPI_Recv(NULL, 0, MPI_BYTE, 0, MPI_TAG_END, MPI_COMM_WORLD, &Status);
-	/*	printf("exiting..., task: %d\n", nID); */
+	/*	printf("exiting..., task: %d\n", nID); 
+		fflush(stdout);*/
 		break;
 	    }
 	    else
 	    {
-		printf("ERROR....invalid tag received\n");
+		printf("ph5diff_worker: ERROR: invalid tag (%d) received\n", Status.MPI_TAG);
 		MPI_Abort(MPI_COMM_WORLD, 0);
 	    }