[svn-r6099]

Purpose:
    bug fix.
Description:
    h5dump cannot dump data and datatype for VL string.
Platforms tested:
    eirene, arabica
Misc. update:
    MANIFEST, RELEASE.txt

8 files changed, 245 insertions, 36 deletions

diff --git a/tools/h5dump/h5dump.c b/tools/h5dump/h5dump.c
index d751d25..36250d3 100644
--- a/tools/h5dump/h5dump.c
+++ b/tools/h5dump/h5dump.c
@@ -677,12 +677,14 @@ print_datatype(hid_t type)
     char       *fname;
     hid_t       nmembers, mtype, str_type;
     int         i, j, ndims, perm[H5DUMP_MAX_RANK];
-    size_t      size;
+    size_t      size=0;
     hsize_t     dims[H5DUMP_MAX_RANK];
     H5T_str_t   str_pad;
     H5T_cset_t  cset;
     H5G_stat_t  statbuf;
     hid_t       super;
+    hid_t       tmp_type;
+    htri_t      is_vlstr=FALSE;
 
     switch (H5Tget_class(type)) {
     case H5T_INTEGER:
@@ -770,15 +772,23 @@ print_datatype(hid_t type)
 	break;
 
     case H5T_STRING:
-	size = H5Tget_size(type);
-	str_pad = H5Tget_strpad(type);
-	cset = H5Tget_cset(type);
+        /* Make a copy of type in memory in case when TYPE is on disk, the size
+         * will be bigger than in memory.  This makes it easier to compare 
+         * types in memory. */
+        tmp_type = H5Tcopy(type);
+	size = H5Tget_size(tmp_type);
+	str_pad = H5Tget_strpad(tmp_type);
+	cset = H5Tget_cset(tmp_type);
+        is_vlstr = H5Tis_variable_str(tmp_type);
 
 	printf("H5T_STRING %s\n", dump_header_format->strblockbegin);
 	indent += COL;
 
 	indentation(indent + COL);
-	printf("%s %d;\n", STRSIZE, (int) size);
+        if(is_vlstr)
+            printf("%s H5T_VARIABLE;\n", STRSIZE);
+        else
+	    printf("%s %d;\n", STRSIZE, (int) size);
 
 	indentation(indent + COL);
 	printf("%s ", STRPAD);
@@ -800,14 +810,17 @@ print_datatype(hid_t type)
 	    printf("unknown_cset;\n");
 
 	str_type = H5Tcopy(H5T_C_S1);
+        if(is_vlstr)
+            H5Tset_size(str_type, H5T_VARIABLE);
+        else
+	    H5Tset_size(str_type, size);
 	H5Tset_cset(str_type, cset);
-	H5Tset_size(str_type, size);
 	H5Tset_strpad(str_type, str_pad);
 
 	indentation(indent + COL);
 	printf("%s ", CTYPE);
 
-	if (H5Tequal(type, str_type)) {
+	if (H5Tequal(tmp_type, str_type)) {
 	    printf("H5T_C_S1;\n");
 	    H5Tclose(str_type);
 	} else {
@@ -817,7 +830,7 @@ print_datatype(hid_t type)
 	    H5Tset_size(str_type, size);
 	    H5Tset_strpad(str_type, str_pad);
 
-	    if (H5Tequal(type, str_type)) {
+	    if (H5Tequal(tmp_type, str_type)) {
 		printf("H5T_FORTRAN_S1;\n");
 	    } else {
 		printf("unknown_one_character_type;\n ");
@@ -827,6 +840,8 @@ print_datatype(hid_t type)
 	    H5Tclose(str_type);
 	}
 
+        H5Tclose(tmp_type);
+
 	indent -= COL;
 	indentation(indent + COL);
 	printf("%s", dump_header_format->strblockend);
@@ -1887,13 +1902,14 @@ dump_data(hid_t obj_id, int obj_data, struct subset_t *sset)
 
         alloc_size = nelmts * MAX(H5Tget_size(type), H5Tget_size(p_type));
         assert(alloc_size == (hsize_t)((size_t)alloc_size)); /*check for overflow*/
+        
         buf = malloc((size_t)alloc_size);
         assert(buf);
 
         if (H5Aread(obj_id, p_type, buf) >= 0)
             status = h5tools_dump_mem(stdout, outputformat, obj_id, p_type,
-                                      space, buf, depth);
-
+                                    space, buf, depth);
+        
         free(buf);
         H5Tclose(p_type); 
         H5Sclose(space);
@@ -2857,10 +2873,25 @@ done:
 	d_status = EXIT_FAILURE;
 
     free_handler(hand, argc);
-
+    
+    for(i=0; i<info.group_table->nobjs; i++) {
+        if(info.group_table->objs[i].objname)
+            free(info.group_table->objs[i].objname);
+    }
     free(group_table->objs);
+    
+    for(i=0; i<info.dset_table->nobjs; i++) {
+        if(info.dset_table->objs[i].objname)
+            free(info.dset_table->objs[i].objname);
+    }
     free(dset_table->objs);
+    
+    for(i=0; i<info.type_table->nobjs; i++) {
+        if(info.type_table->objs[i].objname)
+            free(info.type_table->objs[i].objname);
+    }
     free(type_table->objs);
+    
     free(prefix);
     free(info.prefix);
 
@@ -4557,12 +4588,14 @@ xml_print_strs(hid_t did, int source)
 {
     herr_t                  e;
     hid_t                   type, space;
-    char                   *buf;
+    void                   *buf;
     char                   *bp;
     char                   *onestring;
     hsize_t                 ssiz;
-    size_t                  tsiz;
+    size_t                  tsiz, str_size;
     size_t                  i;
+    htri_t                  is_vlstr;
+
     if (source == DATASET_DATA) {
 	type = H5Dget_type(did);
     } else if (source == ATTRIBUTE_DATA) {
@@ -4575,12 +4608,14 @@ xml_print_strs(hid_t did, int source)
 	/* return an error */
 	return FAIL;
     }
+    is_vlstr = H5Tis_variable_str(type);
+
     if (source == DATASET_DATA) {
 	space = H5Dget_space(did);
 	ssiz = H5Sget_simple_extent_npoints(space);
 	ssiz *= H5Tget_size(type);
 
-	buf = calloc((size_t)ssiz, sizeof(char));
+	buf = malloc((size_t)ssiz);
 
 	if (buf == NULL) {
 	    return FAIL;
@@ -4596,11 +4631,12 @@ xml_print_strs(hid_t did, int source)
 	space = H5Aget_space(did);
 	ssiz = H5Sget_simple_extent_npoints(space);
 	ssiz *= H5Tget_size(type);
-
-	buf = calloc((size_t)ssiz, sizeof(char));
+        
+	buf = malloc((size_t)ssiz);
 	if (buf == NULL) {
 	    return FAIL;
 	}
+
 	e = H5Aread(did, type, buf);
 	if (e < 0) {
 	    free(buf);
@@ -4611,28 +4647,36 @@ xml_print_strs(hid_t did, int source)
 	return FAIL;
     }
 
-/*  pull out each string... */
+    /* pull out each string... */
     ssiz = H5Sget_simple_extent_npoints(space);
 
     tsiz = H5Tget_size(type);
-    onestring = (char *) calloc((size_t)tsiz, sizeof(char));
-    bp = buf;
+    bp = (char*)buf;
+    if(!is_vlstr)
+        onestring = (char *) calloc(tsiz, sizeof(char));
 
     for (i = 0; i < ssiz; i++) {
-	strncpy(onestring, bp, tsiz);
+        if(is_vlstr) {
+            onestring = *(char **)bp;
+            str_size = (size_t)strlen(onestring);
+        } else {    
+    	    strncpy(onestring, bp, tsiz);
+            str_size = tsiz;
+        }
 	indentation(indent + COL);
 
 	if (!onestring) {
 	    printf("\"%s\"\n", "NULL");
 	} else {
-            char *t_onestring = xml_escape_the_string(onestring, (int)tsiz);
+            char *t_onestring = xml_escape_the_string(onestring, (int)str_size);
 
-	    printf("\"%s\"\n", xml_escape_the_string(onestring, (int)tsiz));
+	    printf("\"%s\"\n", t_onestring);
             free(t_onestring);
 	}
-
-	bp += tsiz;
+        
+       bp += tsiz;
     }
+
     return SUCCEED;
 }
 
diff --git a/tools/h5dump/h5dumpgentest.c b/tools/h5dump/h5dumpgentest.c
index 8f6e3bc..f0e9515 100644
--- a/tools/h5dump/h5dumpgentest.c
+++ b/tools/h5dump/h5dumpgentest.c
@@ -63,6 +63,7 @@
 #define FILE35  "tfamily%05d.h5"
 #define FILE36  "tmulti"
 #define FILE37  "tlarge_objname.h5"
+#define FILE38  "tvlstr.h5"
 
 #define LENSTR		50
 #define LENSTR2		11
@@ -108,6 +109,9 @@ typedef struct s1_t {
 #define ARRAY3_DIM1 6
 #define ARRAY3_DIM2 3
 
+/* VL string datatype name */
+#define VLSTR_TYPE      "vl_string_type"
+
 static void gent_group(void)
 {
     hid_t fid, group;
@@ -2838,6 +2842,55 @@ static void gent_large_objname(void)
     H5Fclose(fid);
 }
 
+static void gent_vlstr(void)
+{
+    const char *wdata[SPACE1_DIM1]= {
+        "Four score and seven years ago our forefathers brought forth on this continent a new nation,",
+        "conceived in liberty and dedicated to the proposition that all men are created equal.",
+        "Now we are engaged in a great civil war,",
+        "testing whether that nation or any nation so conceived and so dedicated can long endure."
+        };   /* Information to write */
+    char *string_att= "This is the string for the attribute";
+    hid_t		fid1;		/* HDF5 File IDs		*/
+    hid_t		dataset, root;	/* Dataset ID			*/
+    hid_t		sid1, dataspace;/* Dataspace ID			*/
+    hid_t		tid1, att;      /* Datatype ID			*/
+    hsize_t		dims1[] = {SPACE1_DIM1};
+
+    /* Create file */
+    fid1 = H5Fcreate(FILE38, H5F_ACC_TRUNC, H5P_DEFAULT, H5P_DEFAULT);
+
+    sid1 = H5Screate_simple(SPACE1_RANK, dims1, NULL);
+
+    /* Create a VL string datatype to refer to */
+    tid1 = H5Tcopy (H5T_C_S1);
+    H5Tset_size (tid1,H5T_VARIABLE);
+
+    /* Create a dataset and write VL string to it. */
+    dataset=H5Dcreate(fid1,"Dataset1",tid1,sid1,H5P_DEFAULT);
+    H5Dwrite(dataset,tid1,H5S_ALL,H5S_ALL,H5P_DEFAULT,wdata);
+    H5Dclose(dataset);
+
+    /* Create a named VL string type.  Change padding of datatype */
+    H5Tset_strpad(tid1, H5T_STR_NULLPAD);
+    H5Tcommit(fid1, "vl_string_type", tid1);
+
+    /* Create an group attribute of VL string type */
+    root = H5Gopen(fid1, "/");
+    dataspace = H5Screate(H5S_SCALAR);
+
+    att = H5Acreate(root, "test_scalar", tid1, dataspace, H5P_DEFAULT);
+    H5Awrite(att, tid1, &string_att);
+
+    /* Close */
+    H5Tclose(tid1);
+    H5Sclose(sid1);
+    H5Sclose(dataspace);
+    H5Aclose(att);
+    H5Gclose(root);
+    H5Fclose(fid1);
+}
+
 int main(void)
 {
     gent_group();
@@ -2888,6 +2941,7 @@ int main(void)
     gent_multi();
 
     gent_large_objname();
+    gent_vlstr();
 
     return 0;
 }
diff --git a/tools/h5dump/testh5dump.sh b/tools/h5dump/testh5dump.sh
index ec22362..99c1a89 100755
--- a/tools/h5dump/testh5dump.sh
+++ b/tools/h5dump/testh5dump.sh
@@ -144,6 +144,9 @@ TOOLTEST tvldtypes2.ddl tvldtypes2.h5
 TOOLTEST tvldtypes3.ddl tvldtypes3.h5
 TOOLTEST tvldtypes4.ddl tvldtypes4.h5
 
+#test for file with variable length string data
+TOOLTEST tvlstr.ddl tvlstr.h5
+
 # test for files with array data
 TOOLTEST tarray1.ddl tarray1.h5
 TOOLTEST tarray2.ddl tarray2.h5
@@ -218,6 +221,7 @@ TOOLTEST tarray7.h5.xml --xml tarray7.h5
 TOOLTEST tvldtypes1.h5.xml --xml tvldtypes1.h5
 TOOLTEST tvldtypes2.h5.xml --xml tvldtypes2.h5
 TOOLTEST tvldtypes3.h5.xml --xml tvldtypes3.h5
+TOOLTEST tvlstr.h5.xml --xml tvlstr.h5
 TOOLTEST tsaf.h5.xml --xml tsaf.h5
 TOOLTEST tempty.h5.xml --xml tempty.h5
 
diff --git a/tools/lib/h5tools.c b/tools/lib/h5tools.c
index b2c4cec..fbdb38a 100644
--- a/tools/lib/h5tools.c
+++ b/tools/lib/h5tools.c
@@ -457,6 +457,7 @@ h5tools_dump_simple_data(FILE *stream, const h5dump_t *info, hid_t container,
                                           *this var to count elements and
                                           *break after we see a number equal
                                           *to the ctx->size_last_dim.   */
+    unsigned char       *tmp=NULL;
 
     /* Setup */
     memset(&buffer, 0, sizeof(h5tools_str_t));
@@ -470,7 +471,12 @@ h5tools_dump_simple_data(FILE *stream, const h5dump_t *info, hid_t container,
     for (i = 0; i < nelmts; i++, ctx->cur_elmt++, elmt_counter++) {
         /* Render the element */
         h5tools_str_reset(&buffer);
-        h5tools_str_sprint(&buffer, info, container, type, mem + i * size, ctx);
+        if(H5Tis_variable_str(type)) {
+            tmp = ((unsigned char**)mem)[i];  
+            h5tools_str_sprint(&buffer, info, container, type, tmp, ctx);
+        
+        } else
+            h5tools_str_sprint(&buffer, info, container, type, mem + i * size, ctx);
 
         if (i + 1 < nelmts || (flags & END_OF_DATA) == 0)
             h5tools_str_append(&buffer, "%s", OPT(info->elmt_suf1, ","));
@@ -785,7 +791,7 @@ h5tools_dump_simple_dset(FILE *stream, const h5dump_t *info, hid_t dset,
     hsize_t		sm_nelmts;		/*elements per stripmine*/
     unsigned char      *sm_buf = NULL;		/*buffer for raw data	*/
     hid_t		sm_space;		/*stripmine data space	*/
-
+    
     /* Hyperslab info */
     hssize_t		hs_offset[H5S_MAX_RANK];/*starting offset	*/
     hsize_t		hs_size[H5S_MAX_RANK];	/*size this pass	*/
@@ -820,9 +826,6 @@ h5tools_dump_simple_dset(FILE *stream, const h5dump_t *info, hid_t dset,
             ctx.p_min_idx[i] = 0;
 
     H5Sget_simple_extent_dims(f_space, total_size, NULL);
-    /* printf("total_size[ctx.ndims-1]%d\n",total_size[ctx.ndims-1]);
-     printf("total_size cast %d\n",(int)(total_size[ctx.ndims-1])); */
-	/*assert(total_size[ctx.ndims - 1]==(hsize_t)((int)(total_size[ctx.ndims - 1])));*/
     ctx.size_last_dim = (total_size[ctx.ndims - 1]);
 
     /* calculate the number of elements we're going to print */
@@ -858,6 +861,7 @@ h5tools_dump_simple_dset(FILE *stream, const h5dump_t *info, hid_t dset,
 
     assert(sm_nbytes == (hsize_t)((size_t)sm_nbytes)); /*check for overflow*/
     sm_buf = malloc((size_t)sm_nbytes);
+            
     sm_nelmts = sm_nbytes / p_type_nbytes;
     sm_space = H5Screate_simple(1, &sm_nelmts, NULL);
 
@@ -896,10 +900,10 @@ h5tools_dump_simple_dset(FILE *stream, const h5dump_t *info, hid_t dset,
         flags = (elmtno == 0) ? START_OF_DATA : 0;
         flags |= ((elmtno + hs_nelmts) >= p_nelmts) ? END_OF_DATA : 0;
         h5tools_dump_simple_data(stream, info, dset, &ctx, flags, hs_nelmts,
-                                 p_type, sm_buf);
+                             p_type, sm_buf);
 
         /* Reclaim any VL memory, if necessary */
-        if (vl_data)
+        if(vl_data)
             H5Dvlen_reclaim(p_type, sm_space, H5P_DEFAULT, sm_buf);
 
         /* Calculate the next hyperslab offset */
@@ -925,7 +929,9 @@ h5tools_dump_simple_dset(FILE *stream, const h5dump_t *info, hid_t dset,
 
     H5Sclose(sm_space);
     H5Sclose(f_space);
+    
     free(sm_buf);
+
     return SUCCEED;
 }
 
@@ -967,7 +973,7 @@ h5tools_dump_simple_mem(FILE *stream, const h5dump_t *info, hid_t obj_id,
 
     ctx.indent_level = indentlevel;
     ctx.need_prefix = 1;
-
+    
     /* Assume entire data space to be printed */
     for (i = 0; i < (hsize_t)ctx.ndims; i++)
         ctx.p_min_idx[i] = 0;
@@ -976,7 +982,7 @@ h5tools_dump_simple_mem(FILE *stream, const h5dump_t *info, hid_t obj_id,
 
     for (i = 0, nelmts = 1; ctx.ndims != 0 && i < (hsize_t)ctx.ndims; i++)
         nelmts *= ctx.p_max_idx[i] - ctx.p_min_idx[i];
- 
+    
     if (nelmts == 0)
         return SUCCEED; /*nothing to print*/
     assert(ctx.p_max_idx[ctx.ndims - 1]==(hsize_t)((int)ctx.p_max_idx[ctx.ndims - 1]));
@@ -1075,8 +1081,13 @@ h5tools_fixtype(hid_t f_type)
 	 * strDUAction == TRUE. if it is false we will do the original action
 	 * here.
 	 */
-	m_type = H5Tcopy(f_type);
-	H5Tset_cset(m_type, H5T_CSET_ASCII);
+        if(H5Tis_variable_str(f_type)) {
+	   m_type = H5Tcopy(H5T_C_S1);
+	   H5Tset_size(m_type, H5T_VARIABLE);
+        } else {
+	   m_type = H5Tcopy(f_type);
+	   H5Tset_cset(m_type, H5T_CSET_ASCII);
+        }
 	break;
 
     case H5T_COMPOUND:
diff --git a/tools/lib/h5tools_str.c b/tools/lib/h5tools_str.c
index 0820cfb..3cacff0 100644
--- a/tools/lib/h5tools_str.c
+++ b/tools/lib/h5tools_str.c
@@ -545,7 +545,11 @@ h5tools_str_sprint(h5tools_str_t *str, const h5dump_t *info, hid_t container,
         unsigned int i;
 
         quote = '\0';
-        size = H5Tget_size(type);
+        if(H5Tis_variable_str(type)) {
+            size = HDstrlen(cp_vp);
+        } else {
+            size = H5Tget_size(type);
+        }
         pad = H5Tget_strpad(type);
 
         for (i = 0; i < size && (cp_vp[i] != '\0' || pad != H5T_STR_NULLTERM); i++) {
diff --git a/tools/testfiles/tvlstr.ddl b/tools/testfiles/tvlstr.ddl
new file mode 100644
index 0000000..93b392e
--- /dev/null
+++ b/tools/testfiles/tvlstr.ddl
@@ -0,0 +1,41 @@
+#############################
+Expected output for 'h5dump tvlstr.h5'
+#############################
+HDF5 "tvlstr.h5" {
+GROUP "/" {
+   ATTRIBUTE "test_scalar" {
+      DATATYPE  H5T_STRING {
+            STRSIZE H5T_VARIABLE;
+            STRPAD H5T_STR_NULLPAD;
+            CSET H5T_CSET_ASCII;
+            CTYPE H5T_C_S1;
+         }        
+      DATASPACE  SCALAR  
+      DATA {
+         "This is the string for the attribute"
+      } 
+   } 
+   DATASET "Dataset1" {
+      DATATYPE  H5T_STRING {
+            STRSIZE H5T_VARIABLE;
+            STRPAD H5T_STR_NULLTERM;
+            CSET H5T_CSET_ASCII;
+            CTYPE H5T_C_S1;
+         }        
+      DATASPACE  SIMPLE { ( 4 ) / ( 4 ) } 
+      DATA {
+         "Four score and seven years ago our forefathers brought forth on this continent a new nation,",
+         "conceived in liberty and dedicated to the proposition that all men are created equal.",
+         "Now we are engaged in a great civil war,",
+         "testing whether that nation or any nation so conceived and so dedicated can long endure."
+      } 
+   } 
+   DATATYPE "vl_string_type"       H5T_STRING {
+         STRSIZE H5T_VARIABLE;
+         STRPAD H5T_STR_NULLPAD;
+         CSET H5T_CSET_ASCII;
+         CTYPE H5T_C_S1;
+      };
+    
+} 
+} 
diff --git a/tools/testfiles/tvlstr.h5 b/tools/testfiles/tvlstr.h5
new file mode 100644
index 0000000..53f34ef
--- /dev/null
+++ b/tools/testfiles/tvlstr.h5
diff --git a/tools/testfiles/tvlstr.h5.xml b/tools/testfiles/tvlstr.h5.xml
new file mode 100644
index 0000000..3e5b07e
--- /dev/null
+++ b/tools/testfiles/tvlstr.h5.xml
@@ -0,0 +1,51 @@
+#############################
+Expected output for 'h5dump --xml tvlstr.h5'
+#############################
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE HDF5-File PUBLIC "HDF5-File.dtd" "http://hdf.ncsa.uiuc.edu/DTDs/HDF5-File.dtd">
+<HDF5-File>
+<RootGroup OBJ-XID="root">
+   <Attribute Name="test_scalar">
+      <Dataspace>
+         <ScalarDataspace />
+      </Dataspace>
+      <DataType> 
+         <AtomicType>
+            <StringType Cset="H5T_CSET_ASCII" StrSize="4" StrPad="H5T_STR_NULLPAD"/>
+         </AtomicType>
+      </DataType>
+      <Data>
+         <DataFromFile>
+         "This is the string for the attribute"
+         </DataFromFile>
+      </Data>
+   </Attribute>
+   <NamedDataType Name="vl_string_type" OBJ-XID="/vl_string_type" Parents="root">
+      <DataType>
+         <AtomicType>
+            <StringType Cset="H5T_CSET_ASCII" StrSize="16" StrPad="H5T_STR_NULLPAD"/>
+         </AtomicType>
+      </DataType>
+   </NamedDataType>
+   <Dataset Name="Dataset1" OBJ-XID="/Dataset1" Parents="root">
+      <Dataspace>
+         <SimpleDataspace Ndims="1">
+            <Dimension  DimSize="4" MaxDimSize="4"/>
+         </SimpleDataspace>
+      </Dataspace>
+      <DataType> 
+         <AtomicType>
+            <StringType Cset="H5T_CSET_ASCII" StrSize="4" StrPad="H5T_STR_NULLTERM"/>
+         </AtomicType>
+      </DataType>
+      <Data>
+         <DataFromFile>
+         "Four score and seven years ago our forefathers brought forth on this continent a new nation,"
+         "conceived in liberty and dedicated to the proposition that all men are created equal."
+         "Now we are engaged in a great civil war,"
+         "testing whether that nation or any nation so conceived and so dedicated can long endure."
+         </DataFromFile>
+      </Data>
+   </Dataset>
+</RootGroup>
+</HDF5-File>