Configure compiler and silence warnings of Clang-17 (#1)

.github/workflows/ci.yml

@@ -0,0 +1,49 @@
+name: CI
+
+on:
+  pull_request:
+  push:
+    branches:
+      - master
+
+jobs:
+  build:
+    name: ${{matrix.name}}
+    runs-on: ${{matrix.os}}
+    env:
+      CC: ${{matrix.cc}}
+    defaults:
+      run:
+        shell: bash
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - name: Ubuntu-GCC
+            cc: gcc
+            os: ubuntu-latest
+
+          - name: Ubuntu-Clang
+            cc: clang
+            os: ubuntu-latest
+
+          - name: macOS-AppleClang
+            cc: clang
+            os: macos-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: System Information
+        run: |
+          awk -v a=$(uname) 'BEGIN { a == "Linux" ? system("cat /etc/issue") : system("sw_vers") }'
+          ${CC} --version
+
+      - name: Make
+        run: |
+          make
+
+      - name: Generate SF 0.01 Data
+        run: |
+          mkdir generated_data && cd generated_data
+          ../dbgen -f -s 0.01 -T a -b ../dists.dss

README → README.md

@@ -1,3 +1,8 @@
+![Build Status](https://github.com/hyrise/ssb-dbgen/actions/workflows/ci.yml/badge.svg?branch=master)
+### Note: This fork is adapted for compilation using recent versions of Clang and GCC, which we ensure with a [GitHub action](https://github.com/hyrise/ssb-dbgen/actions).
+
+# SSB Data Generator
+
 Note: In our research paper we use the SSB instead of SSBM
 Version of 2/28/10: 
 Cardinality of supplier fixed to follow benchmark spec: now 2000*SF

bm_utils.c

@@ -575,7 +575,20 @@ set_state(int table, long sf, long procs, long step, long *extra_rows)
 		}
 	*extra_rows = remainder % procs;
 	if (step > procs)	/* moving to the end to generate updates */
-		tdefs[table].gen_seed(*extra_rows);
+        {
+        // HYRISE: The following line raises a compilation error when adding function prototypes the the tdef gen_seed
+        // member, as they epxect two arguments. All seed generators passed in driver.c (L 230-249) take either two
+        // parameters or are simply 0. Thus, calling this line would always lead to an error. With this assumption, we
+        // simply ensure we fail if this should ever be called (did not happen after generating tables for multiple
+        // scale factors).
+        //
+        // tdefs[table].gen_seed(*extra_rows);
+        //
+        fprintf(stderr,
+            "Did not expect to reach this line. Re-check function prototypes and remove compiler check if necessary. See %s:%d\n",
+            __FILE__, __LINE__);
+        exit(1);
+        }
 
 	return(result);
 }

build.c

@@ -76,11 +76,12 @@ gen_phone(long ind, char *target, long seed)
 
     RANDOM(acode, 100, 999, seed);
     RANDOM(exchg, 100, 999, seed);
-    RANDOM(number, 1000, 9999, seed);	
-    sprintf(target, "%02d", 10 + (ind % NATIONS_MAX));
-    sprintf(target + 3, "%03d", acode);
-    sprintf(target + 7, "%03d", exchg);
-    sprintf(target + 11, "%04d", number);
+    RANDOM(number, 1000, 9999, seed);
+    // HYRISE: Change format specifier.
+    sprintf(target, "%02ld", 10 + (ind % NATIONS_MAX));
+    sprintf(target + 3, "%03ld", acode);
+    sprintf(target + 7, "%03ld", exchg);
+    sprintf(target + 11, "%04ld", number);
     target[2] = target[6] = target[10] = '-';	
     return;
 }
@@ -91,8 +92,9 @@ gen_category(char *target, long seed){
   RANDOM(num1,1,5,seed);
   RANDOM(num2,1,5,seed);
   strcpy(target,"MFGR");
-  sprintf(target + 4, "%01d", num1);
-  sprintf(target + 5, "%01d", num2);
+  // HYRISE: Change format specifier.
+  sprintf(target + 4, "%01ld", num1);
+  sprintf(target + 5, "%01ld", num2);
   return;
 } 
 
@@ -410,14 +412,17 @@ long mk_part(long index, part_t *p)
 
 
 	RANDOM(mfgr, P_MFG_MIN, P_MFG_MAX, P_MFG_SD);
-	sprintf(p->mfgr, "%s%d", "MFGR#", mfgr);
+	// HYRISE: Change format specifier.
+	sprintf(p->mfgr, "%s%ld", "MFGR#", mfgr);
 
     RANDOM(cat, P_CAT_MIN, P_CAT_MAX, P_CAT_SD);
-    sprintf(p->category, "%s%d", p->mfgr,cat);
+    // HYRISE: Change format specifier.
+    sprintf(p->category, "%s%ld", p->mfgr,cat);
 
 
 	RANDOM(brnd, P_BRND_MIN, P_BRND_MAX, P_BRND_SD);
-	sprintf(p->brand,"%s%d",p->category,brnd);
+	// HYRISE: Change format specifier.
+	sprintf(p->brand,"%s%ld",p->category,brnd);
 
 	p->tlen = pick_str(&p_types_set, P_TYPE_SD, p->type);
 	p->tlen = strlen(p_types_set.list[p->tlen].text);
@@ -610,7 +615,8 @@ int gen_city(char *cityName, char *nationName){
     }
     RANDOM(randomPick, 0, 9, 98);
 
-    sprintf(cityName+CITY_FIX-1,"%d",randomPick);
+    // HYRISE: Change format specifier.
+    sprintf(cityName+CITY_FIX-1,"%ld",randomPick);
     cityName[CITY_FIX] = '\0';
     return 0; 
 }

driver.c

@@ -108,10 +108,12 @@ char *spawn_args[25];
 * flat file print functions; used with -F(lat) option
 */
 #ifdef SSBM
-int pr_cust (customer_t * c, int mode);
-int pr_part (part_t * p, int mode);
-int pr_supp (supplier_t * s, int mode);
-int pr_line (order_t * o, int mode);
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int pr_cust (void * c, int mode);
+int pr_part (void * p, int mode);
+int pr_supp (void * s, int mode);
+int pr_line (void * t, int mode);
 #else
 int pr_cust (customer_t * c, int mode);
 int pr_line (order_t * o, int mode);
@@ -129,13 +131,15 @@ int pr_region (code_t * c, int mode);
 * inline load functions; used with -D(irect) option
 */
 #ifdef SSBM
-int ld_cust (customer_t * c, int mode);
-int ld_part (part_t * p, int mode);
-int ld_supp (supplier_t * s, int mode);
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int ld_cust (void * c, int mode);
+int ld_part (void * p, int mode);
+int ld_supp (void * s, int mode);
 
 /*todo: get rid of ld_order*/
-int ld_line (order_t * o, int mode);
-int ld_order (order_t * o, int mode);
+int ld_line (void * o, int mode);
+int ld_order (void * o, int mode);
 
 #else
 int ld_cust (customer_t * c, int mode);
@@ -198,12 +202,14 @@ int hd_region (FILE * f);
 * data verfication functions; used with -O v option
 */
 #ifdef SSBM
-int vrf_cust (customer_t * c, int mode);
-int vrf_part (part_t * p, int mode);
-int vrf_supp (supplier_t * s, int mode);
-int vrf_line (order_t * o, int mode);
-int vrf_order (order_t * o, int mode);
-int vrf_date (date_t,int mode);
+// HYRISE: change first parameter to void pointer so the tdef verify member can have a function prototype
+// (C2x compatibility).
+int vrf_cust (void * c, int mode);
+int vrf_part (void * p, int mode);
+int vrf_supp (void * s, int mode);
+int vrf_line (void * o, int mode);
+int vrf_order (void * o, int mode);
+int vrf_date (void * d, int mode);
 #else
 int vrf_cust (customer_t * c, int mode);
 int vrf_line (order_t * o, int mode);
@@ -604,7 +610,8 @@ partial (int tbl, int s)
 
 	if (verbose > 0)
 	{
-		fprintf (stderr, "\tStarting to load stage %d of %d for %s...",
+		// HYRISE: Change format specifier.
+		fprintf (stderr, "\tStarting to load stage %d of %ld for %s...",
 			s, children, tdefs[tbl].comment);
 	}
 
@@ -633,7 +640,8 @@ pload (int tbl)
 
 	if (verbose > 0)
 	{
-		fprintf (stderr, "Starting %d children to load %s",
+		// HYRISE: Change format specifier.
+		fprintf (stderr, "Starting %ld children to load %s",
 			children, tdefs[tbl].comment);
 	}
 	for (c = 0; c < children; c++)
@@ -1017,8 +1025,9 @@ main (int ac, char **av)
 			{
 			if (verbose > 0)
 #ifdef SSBM
+				// HYRISE: Change format specifier.
 				fprintf (stderr,
-				"Generating update pair #%d for %s [pid: %d]",
+				"Generating update pair #%ld for %s [pid: %d]",
 				upd_num + 1, tdefs[LINE].comment, DSS_PROC);
 #else
 				fprintf (stderr,
@@ -1117,14 +1126,16 @@ main (int ac, char **av)
 					}
 #endif
 					if (verbose > 0)
-						fprintf (stderr, "%s data for %s [pid: %ld]",
+						// HYRISE: Change format specifier.
+						fprintf (stderr, "%s data for %s [pid: %d]",
 						(validate)?"Validating":"Generating", tdefs[i].comment, DSS_PROC);
 					gen_tbl (i, minrow, rowcnt, upd_num);
 					if (verbose > 0)
 						fprintf (stderr, "done.\n");
 				}
 				if (validate)
-					printf("Validation checksum for %s at %d GB: %0x\n", 
+					// HYRISE: Change format specifier.
+					printf("Validation checksum for %s at %ld GB: %0lx\n",
 						 tdefs[i].name, scale, tdefs[i].vtotal);
 		}
 

dss.h

@@ -161,10 +161,11 @@ typedef struct
    char     *name;
    char     *comment;
    long      base;
-   int       (*header) ();
-   int       (*loader[2]) ();
-   long      (*gen_seed)();
-   int       (*verify) ();
+   // HYRISE: add function prototypes (C2x compatibility).
+   int       (*header) (FILE *);
+   int       (*loader[2]) (void *, int);
+   long      (*gen_seed)(int, long);
+   int       (*verify) (void *, int);
    int       child;
    unsigned long vtotal;
 }         tdef;
@@ -359,7 +360,8 @@ extern tdef tdefs[];
  */
 #define  C_SIZE       165
 #define  C_NAME_TAG   "Customer#"
-#define  C_NAME_FMT   "%s%09d"
+// HYRISE: Change format specifier.
+#define  C_NAME_FMT   "%s%09ld"
 #define  C_MSEG_MAX    5
 #define  C_ABAL_MIN   -99999
 #define  C_ABAL_MAX    999999
@@ -515,16 +517,17 @@ int dbg_print(int dt, FILE *tgt, void *data, int len, int eol);
 #define  PR_STRT(fp)   /* any line prep for a record goes here */
 #define  PR_END(fp)    fprintf(fp, "\n")   /* finish the record here */
 
+// HYRISE: Change format specifier.
 #ifdef SSBM
 #define  PR_DATE(tgt, yr, mn, dy)	\
-   sprintf(tgt, "19%02d%02d%02d", yr, mn, dy)
+   sprintf(tgt, "19%02ld%02ld%02ld", yr, mn, dy)
 #else
 #ifdef MDY_DATE
 #define  PR_DATE(tgt, yr, mn, dy)	\
-   sprintf(tgt, "%02d-%02d-19%02d", mn, dy, yr)
+   sprintf(tgt, "%02ld-%02ld-19%02ld", mn, dy, yr)
 #else
 #define  PR_DATE(tgt, yr, mn, dy)	\
-sprintf(tgt, "19%02d-%02d-%02d", yr, mn, dy)
+sprintf(tgt, "19%02ld-%02ld-%02ld", yr, mn, dy)
 #endif /* DATE_FORMAT */
 #endif
 /*

dsstypes.h

@@ -43,8 +43,10 @@ typedef struct
 
 /* customers.c */
 long mk_cust   PROTO((long n_cust, customer_t * c));
-int pr_cust    PROTO((customer_t * c, int mode));
-int ld_cust    PROTO((customer_t * c, int mode));
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int pr_cust    PROTO((void * c, int mode));
+int ld_cust    PROTO((void * c, int mode));
 
 #ifdef SSBM
 
@@ -125,7 +127,9 @@ typedef struct
 /* order.c */
 long	mk_order	PROTO((long index, order_t * o, long upd_num));
 int		pr_order	PROTO((order_t * o, int mode));
-int		ld_order	PROTO((order_t * o, int mode));
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int		ld_order	PROTO((void * o, int mode));
 void	ez_sparse	PROTO((long index, DSS_HUGE *ok, long seq));
 #ifndef SUPPORT_64BITS
 void	hd_sparse	PROTO((long index, DSS_HUGE *ok, long seq));
@@ -182,8 +186,10 @@ typedef struct
 
 /* parts.c */
 long mk_part   PROTO((long index, part_t * p));
-int pr_part    PROTO((part_t * part, int mode));
-int ld_part    PROTO((part_t * part, int mode));
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int pr_part    PROTO((void * part, int mode));
+int ld_part    PROTO((void * part, int mode));
 
 #ifdef SSBM
 typedef struct
@@ -216,8 +222,10 @@ typedef struct
 
 /* supplier.c */
 long mk_supp   PROTO((long index, supplier_t * s));
-int pr_supp    PROTO((supplier_t * supp, int mode));
-int ld_supp    PROTO((supplier_t * supp, int mode));
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int pr_supp    PROTO((void * supp, int mode));
+int ld_supp    PROTO((void * supp, int mode));
 
 #ifdef SSBM
 /*todo: add new date table*/
@@ -247,8 +255,10 @@ typedef struct
 /* date.c */
 
 long mk_date   PROTO((long index, date_t * d));
-int pr_date    PROTO((date_t * date, int mode));
-int ld_date    PROTO((date_t * date, int mode));
+// HYRISE: change first parameter to void pointer so the tdef loader member can have a function prototype
+// (C2x compatibility).
+int pr_date    PROTO((void * date, int mode));
+int ld_date    PROTO((void * date, int mode));
 
 #endif