From 80ab2fddfdf30f09f0a0a637654cbb3cd5c7baa6 Mon Sep 17 00:00:00 2001
From: Rich Turner
Date: Fri, 12 Aug 1983 17:53:34 -0700
Subject: MS-DOS v2.0 Release

---
 v2.0/bin/PROHST.PAS | 403 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 403 insertions(+)
 create mode 100644 v2.0/bin/PROHST.PAS

(limited to 'v2.0/bin/PROHST.PAS')

diff --git a/v2.0/bin/PROHST.PAS b/v2.0/bin/PROHST.PAS
new file mode 100644
index 0000000..4ee13fc
--- /dev/null
+++ b/v2.0/bin/PROHST.PAS
@@ -0,0 +1,403 @@
+PROGRAM prohst(input,output);
+{$debug- $line- $symtab+}
+
+{**********************************************************************}
+{*                                                                    *}
+{*                           prohst                                   *}
+{*                                                                    *}
+{* This program produces a histogram from the profile file produced   *}
+{* by the MS-DOS profile utility. It optionally reads the map file    *}
+{* generated when the program being profiled was linked, and writes   *}
+{* either the module address or, if available, the line number as     *}
+{* a prefix to the line of the graph which describes a particular     *}
+{* bucket.                                                            *}
+{*                                                                    *}
+{* After using filbm (derived from the Pascal and Fortran front end   *}
+{* command scanner) to parse its parameters, prohst opens the map     *}
+{* file if specified, searches for the heading line, and then reads   *}
+{* the lines giving the names and positions of the modules. It builds *}
+{* a linked list of module names and start addresses.                 *}
+{*                                                                    *}
+{* It then reads the bucket file header and and bucket array elements *}
+{* into a variable created on the heap. It simultaneously calculates  *}
+{* a normalization factor. It writes the profile listing header and   *}
+{* starts to write the profile lines. For each bucket, the address    *}
+{* is calculated. The first entry in the address/name linked list     *}
+{* is the lowest addressed module. This is initially the 'current'    *}
+{* module. The bucket address is compared with the current module     *}
+{* address. When it becomes the greater, the module name is written   *}
+{* to the listing and the next entry in the address/name list becomes *}
+{* the current module. If line numbers are available, the bucket      *}
+{* address is also compared to the current line/address. This is      *}
+{* read and calculated directly from the file. Since there may be     *}
+{* more than one line per bucket, several entries may be read until   *}
+{* the addresses compare within the span of addresses encompassed by  *}
+{* a bucket (its 'width'). Note that the idiosyncracies of Pascal i/o *}
+{* make it necessary to continually check for the end of the map file *}
+{* and the complexity of this code is mainly due to an attempt to     *}
+{* make it reasonably resilient to changes in the format of the map   *}
+{* file.                                                              *}
+{*                                                                    *}
+{**********************************************************************}
+
+
+CONST
+  max_file = 32;
+
+
+TYPE
+  filenam = LSTRING (max_file);
+  sets = SET OF 0..31;
+  address_pointer = ^address_record;
+  address_record = RECORD
+                     next: address_pointer;
+                     name: STRING (15);
+                     address: WORD;
+                   END;
+
+        
+
+
+
+VAR
+
+  i: INTEGER;
+  bucket: FILE OF WORD;
+  hist: TEXT;
+  map: TEXT;
+
+  first_address,
+  this_address: address_pointer; 
+  current_base: WORD;
+  bucket_name,
+  hist_name,
+  map_name: filenam;
+
+  switches: sets;
+
+  line: LSTRING (100);
+
+  map_avail: BOOLEAN;
+  line_nos_avail: BOOLEAN;
+
+  norm: REAL;
+  per_cent: INTEGER;
+  real_bucket,
+  norm_bucket: REAL;
+  cum_per_cent,
+  real_per_cent: REAL;
+
+  bucket_num,
+  clock_grain,
+  bucket_size,
+  prog_low_pa,
+  prog_high_pa,
+  dos_pa,
+  hit_io,
+  hit_dos,
+  hit_high: WORD;
+
+  seg,
+  offset,
+  parcel: WORD;
+
+  address: WORD;
+  new_line_no,
+  line_no: WORD;
+
+  dummy : LSTRING (8);
+  name: LSTRING (20);
+  line_no_part: LSTRING (17);
+  start: LSTRING (6);
+
+  buckets: ^SUPER ARRAY [1 .. *] OF REAL;
+
+  this_bucket: WORD;
+
+LABEL 1;
+
+
+PROCEDURE filbm (VAR prffil, hstfil, mapfil: filenam;
+                      VAR switches: sets); EXTERN;
+
+FUNCTION realword (w: WORD): REAL;
+BEGIN
+  IF ORD (w) < 0 THEN BEGIN
+    realword := FLOAT (maxint) + FLOAT (ORD (w - maxint));
+    END
+  ELSE BEGIN
+    realword := FLOAT (ORD(w));
+  END {IF};
+END {realword};
+
+
+
+PROCEDURE skip_spaces;
+BEGIN
+  WHILE NOT eof(map) AND THEN map^ = ' ' DO BEGIN
+    get (map);
+  END {WHILE};
+END {skip_spaces};
+
+
+FUNCTION hex_char (ch: CHAR): WORD;
+BEGIN
+  IF ch >= '0' AND THEN ch <= '9' THEN BEGIN
+    hex_char := WRD (ch) - WRD ('0');
+    END
+  ELSE IF ch >= 'A' AND THEN ch <= 'F' THEN BEGIN
+    hex_char := WRD (ch) - WRD ('A') + 10;
+    END
+  ELSE BEGIN
+    WRITELN ('Invalid hex character');
+    hex_char := 0;
+  END {IF};
+END {hex_char};
+
+
+FUNCTION read_hex (i :WORD): WORD;
+VAR
+  hex_val: WORD;
+BEGIN
+  skip_spaces;
+  hex_val := 0;
+  WHILE NOT eof (map) AND THEN i <> 0 DO BEGIN
+    hex_val := hex_val * 16 + hex_char (map^);
+    GET (map);
+    i := i - 1;
+  END {WHILE};
+    read_hex := hex_val;
+END {read_hex};
+
+FUNCTION read_h: WORD;
+BEGIN
+  read_h := read_hex (4);
+  get (map);
+  get (map);
+END;
+
+FUNCTION read_word: WORD;
+VAR 
+  int_value: WORD;
+BEGIN
+  int_value := 0;
+  IF NOT EOF (map) THEN BEGIN
+    READ (map, int_value);
+  END {IF};
+  read_word := int_value;
+END {read_word};
+
+
+FUNCTION map_digit: BOOLEAN;
+BEGIN
+  map_digit := (map^ >= '0') OR (map^ <= '9');
+END {map_digit};
+
+BEGIN {prohst}
+  writeln (output, '    Profile Histogram Utility - Version 1.0');
+  writeln (output);
+  writeln (output, '         Copyright - Microsoft, 1983');
+         
+  start := '      ';
+
+  filbm (bucket_name, hist_name, map_name, switches);
+
+  IF 31 IN switches THEN BEGIN
+    ABORT ('Map file must not be terminal', 0, 0);
+  END {IF};
+
+  IF NOT (28 IN switches) THEN BEGIN
+    ABORT ('No histogram file specified', 0, 0);
+  END {IF};
+
+  ASSIGN (bucket, bucket_name);
+  reset (bucket);
+  ASSIGN (hist, hist_name);
+  rewrite (hist);
+  
+  map_avail := 29 IN switches;
+  line_nos_avail := FALSE;
+
+  IF map_avail THEN BEGIN
+    ASSIGN (map, map_name);
+    RESET (map);
+  
+    WHILE NOT EOF (map) AND THEN start <> ' Start' DO BEGIN
+      READLN (map, start);
+    END {WHILE};
+    
+    NEW (first_address);
+    this_address := NIL;
+
+    WHILE NOT EOF(map) DO BEGIN
+      READLN (map, line);
+      IF line.len < 6 OR ELSE line [2] < '0' OR ELSE
+          line [2] > '9' THEN BEGIN
+        BREAK;
+      END {IF};
+
+      IF this_address <> NIL THEN BEGIN
+        NEW (this_address^.next);
+        this_address := this_address^.next;
+        END
+      ELSE BEGIN
+        this_address := first_address;
+      END {IF};
+      this_address^.next := NIL;
+
+      this_address^.address := (hex_char (line [2]) * 4096) + 
+                               (hex_char (line [3]) * 256) +   
+                               (hex_char (line [4]) * 16) + 
+                               hex_char (line [5]);
+
+      FOR i := 1 TO 15 DO BEGIN
+        this_address^.name [i] := line [22 + i];
+      END {FOR};
+
+    END {WHILE};
+
+    WHILE NOT EOF (map) DO BEGIN
+      READLN (map, line_no_part);
+      IF line_no_part = 'Line numbers for ' THEN BEGIN
+        line_nos_avail := TRUE;
+        BREAK;
+      END {IF};
+    END {WHILE};
+    
+  END {IF};
+
+  read (bucket, clock_grain, bucket_num, bucket_size,
+    prog_low_pa, prog_high_pa, dos_pa, hit_io, hit_dos, hit_high);
+
+  NEW (buckets,ORD (bucket_num));
+
+  norm := 0.0;
+  norm_bucket := 0.0;
+
+  FOR i := 1 TO ORD (bucket_num) DO BEGIN
+    read (bucket, this_bucket);
+    real_bucket := realword (this_bucket);
+
+    IF real_bucket > norm_bucket THEN BEGIN
+      norm_bucket := real_bucket;
+    END {IF};
+
+    norm := norm + real_bucket;
+    buckets^[i] := real_bucket;
+  END {FOR};
+  norm_bucket := 45.0/norm_bucket;
+  norm := 100.0/norm;
+
+  WRITELN (hist, 'Microsoft Profiler Output Listing');
+  
+  WRITELN (hist);
+  WRITELN (hist, ORD (bucket_num):6, bucket_size:4,'-byte buckets.');
+
+  WRITELN (hist);
+  WRITELN (hist, 'Profile taken between ', prog_low_pa*16::16,
+    ' and ', prog_high_pa*16::16, '.');
+
+  WRITELN (hist);
+  WRITELN (hist, 'DOS program address:', dos_pa::16);
+
+  WRITELN (hist);
+  WRITELN (hist, 'Number of hits in DOS: ', hit_dos:5, 
+            ' or ', realword (hit_dos) * norm:4:1, '%.');
+  WRITELN (hist, 'Number of hits in I/O: ', hit_io:5,
+            ' or ', realword (hit_io) * norm:4:1, '%.');
+  WRITELN (hist, 'Number of hits high  : ', hit_high:5,
+            ' or ', realword (hit_high) * norm:4:1, '%.');
+  WRITELN (hist);
+  WRITELN (hist, ' Hits  Addr.  Line/ Cumul.  % 0.0               ',
+                              '                         ',
+                              1.0/norm:1:1);
+
+  WRITELN (hist, '              Offset           +----------------',
+                              '----------------------------');
+  WRITELN (hist, name);
+  i := 0;
+  parcel := 0;
+  current_base := 0;
+  line_no := 0;
+  new_line_no := 0;
+  cum_per_cent := 0.0;
+
+  WHILE i < ORD (bucket_num) DO BEGIN
+    i := i + 1;
+    IF buckets^[i] < 0.9 THEN BEGIN
+      WRITELN (hist);
+      REPEAT
+        i := i + 1;
+      UNTIL (i = ORD (bucket_num)) OR ELSE buckets^[i] > 0.0;
+    END {IF};
+
+    address := bucket_size * (WRD (i) - 1);
+    
+    WHILE map_avail AND THEN
+        address >= first_address^.address DO BEGIN
+      WRITELN (hist, '      ', first_address^.name);
+      current_base := first_address^.address;
+      first_address := first_address^.next;
+    END {WHILE};
+
+    WHILE line_nos_avail AND THEN NOT eof (map) AND THEN
+                                       address >= parcel DO BEGIN
+      skip_spaces;
+      WHILE  (map^ < '0') OR (map^ > '9') DO BEGIN
+      
+        IF EOF (map) THEN BEGIN
+          goto 1;
+        END {IF};
+        READLN (map);
+        skip_spaces;
+      END {WHILE};
+
+
+      line_no := new_line_no;
+      new_line_no := read_word;
+      seg := read_hex (4);
+      IF EOF (map) THEN BEGIN
+        GOTO 1;
+      END {IF};
+      IF map^ <> ':' THEN BEGIN
+        WRITELN ('Invalid map file');
+      END {IF};
+      get (map);
+      IF EOF (map) THEN BEGIN
+        GOTO 1;
+      END {IF};
+      offset := read_hex (3) + WRD (hex_char (map^) > 0);
+      get (map);
+      IF map^ <> 'H' THEN BEGIN
+        WRITELN ('Invalid map file');
+      END {IF};
+      IF EOF (map) THEN BEGIN
+        GOTO 1;
+      END {IF};
+      get (map);
+      parcel := seg + offset;
+    END {WHILE};
+1:  real_per_cent := buckets^[i] * norm;
+    cum_per_cent := cum_per_cent + real_per_cent;
+    per_cent := ROUND ( buckets^[i] * norm_bucket);
+
+    WRITE (hist, buckets^ [i]:6:0, ' ',
+               address*16:6:16);
+    IF line_no <> 0 THEN BEGIN 
+      WRITE (hist, line_no:6);
+      line_no := 0;
+      END
+    ELSE IF map_avail AND THEN first_address <> NIL THEN BEGIN
+      WRITE (hist, ' #', address - first_address^.address:4:16);
+      END
+    ELSE BEGIN
+      WRITE (hist, '      ');
+    END {IF};
+    
+    WRITELN (hist, ' ', cum_per_cent:5:1, ' ', real_per_cent:4:1, ' |',
+               '*': per_cent);
+  END {WHILE};
+  WRITELN (hist, '                               +-----------------',
+                              '------------------');
+END.
+
\ No newline at end of file
-- 
cgit v1.2.3