[gnugo-devel] breakage2tst.py tweak

[Arend Bayer]

Instead of running a separate test file continaing the changed tests, it
seems more convenient to me to invoke regress.pike with the relevant
test cases as arguments (as in
./regress.pike  nngs1.tst:27 strategy.tst:13 endgame.tst:840 ...)

This patch let breakage2tst.py generate such a command line instead of
generating a separate test file. The --update mode is unchanged.

Arend


Index: regression/breakage2tst.py
===================================================================
RCS file: /cvsroot/gnugo/gnugo/regression/breakage2tst.py,v
retrieving revision 1.6
diff -u -p -r1.6 breakage2tst.py
--- regression/breakage2tst.py  27 Feb 2004 21:18:10 -0000      1.6
+++ regression/breakage2tst.py  5 Jan 2005 00:02:29 -0000
@@ -31,102 +31,20 @@ import re
 
 help_string = """
 Usage:
-breakage2tst.py [--pike] <BREAKAGE_FILE >testfile.tst
-       This creates an excerpt of all test cases that appear as unexpected
-       PASS or FAIL in BREAKAGE_FILE and writes these test to testfile.tst.
+breakage2tst.py [--pike] <BREAKAGE_FILE 
+       This creates a command line invoking regress.pike to run all tes
+       cases that appear as unexpected PASS or FAIL in BREAKAGE_FILE.
 breakage2tst.py [--pike] --update <BREAKAGE_FILE
        This changes all .tst files so that the expected results match
        the behaviour of the version that produced BREAKAGE_FILE.
 In both cases, it needs to be run from the regression test directory.
 """
 
-start_tst_string = """
-reset_owl_node_counter
-reset_reading_node_counter
-reset_connection_node_counter
-"""
-               
-finish_tst_string = """
-########### end of tests #####################
-
-# Report number of nodes visited by the tactical reading
-100000 get_reading_node_counter
-#? [0]&
-
-# Report number of nodes visited by the owl code
-100001 get_owl_node_counter
-#? [0]&
-
-# Report number of nodes visited by connection reading
-100002 get_connection_node_counter
-#? [0]&
-"""
-
-
-# This function scans the .tst-file <tstfilename> for tests with numbers in
-# the list testnumbers. These have to be ordered in the same way as they
-# are occuring in the .tst-file
-
-# We copy all non-numbered commands starting from the last loadsgf leading
-# up to each test to be copied, plus all following non-numbered commands
-# until the next loadsgf.
-# This is to cope with trymove/popgo's or increase/decrease_depth_values
+# This prints out the list of tests from testfile in the format
+# <tstfile>:number
 def write_tests(tstfilename, tests):
-       global command_id
-
-       if len(tests) == 0:
-               return
-
-       print
-       print '# Tests from', tstfilename+':',
-       tstfile = open(tstfilename, 'r')
-       commands = ''
-       state = 0
        for number, expected in tests:
-               comment = ''
-               current_line = tstfile.readline()
-               command_pattern = re.compile("^%d " % number)
-
-               # Look for the line containing the command with matching id,
-               # while keeping recent commands and comments
-               while (not command_pattern.match(current_line)):
-                       if re.match(r"#[^\?]", current_line):
-                               comment = comment + current_line
-                       elif (re.match(r"$", current_line)
-                             or re.match(r" +$", current_line)):
-                               comment = ''
-                       elif re.search("loadsgf", current_line):
-                               if state == 1:
-                                       print commands,
-                                       state = 0
-                               commands = current_line
-                       elif (not re.match(r"[0-9]|#|^ *$", current_line)):
-                               commands = commands + current_line
-                       current_line = tstfile.readline()
-               # Found match.
-               state = 1
-               print
-               print '#', tstfilename+':', number
-               print comment + commands, command_id, \
-                       re.sub(command_pattern, '', current_line),
-               commands = ''
-
-               # Now look for the result line:
-               while (not re.match(r"^#\?", current_line)):
-                       current_line = tstfile.readline()
-               print current_line,
-               command_id = command_id + 10
-
-       # We need to do a final scan for some popgo's or similar until
-       # the next loadsgf or EOF:
-       current_line = tstfile.readline()
-       while (current_line != '' and not re.search(" *loadsgf", current_line)):
-               if (not re.match(r"[0-9]|#| *$", current_line)):
-                       print current_line
-               current_line = tstfile.readline()
-
-       tstfile.close
-       print
+               print "%s:%d" % (tstfilename, number),
 
 
 def toggled_result(resultline, expected):
@@ -240,7 +158,6 @@ def parse_pike_input(do_work):
 
 
 def main():
-       global command_id
        mode = 0
        pike = 0
        try:
@@ -265,8 +182,7 @@ def main():
                        pike = 1
        
        if (mode == 0):
-               print start_tst_string
-               command_id = 10
+               print "./regress.pike ",
                do_work = write_tests
        else:
                do_work = update_tstfile
@@ -278,7 +194,7 @@ def main():
                parse_input(do_work)
 
        if (mode == 0):
-               print finish_tst_string
+               print
        else:
                print "Done."