path: root/lib/dg.exp

# `dg' general purpose testcase driver.
# Copyright (C) 94, 95, 96, 97, 98, 1999 Free Software Foundation, Inc.

# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
# 
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
# 
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.

# Please email any bugs, comments, and/or additions to this file to:
# dje@cygnus.com.

# This file was written by Doug Evans (dje@cygnus.com).

# This file is based on old-dejagnu.exp.  It is intended to be more extensible
# without incurring the overhead that old-dejagnu.exp can.  All test framework
# commands appear in the testcase as "{ dg-xxx args ... }".  We pull them out
# with one grep, and then run the function(s) named by "dg-xxx".  When running
# dg-xxx, the line number that it occurs on is always passed as the first
# argument.  We also support different kinds of tools via callbacks.
#
# The currently supported options are:
#
# dg-prms-id N
#	set prms_id to N
#
# dg-options "options ..." [{ target selector }]
#	specify special options to pass to the tool (eg: compiler)
#
# dg-do do-what-keyword [{ target/xfail selector }]
#	`do-what-keyword' is tool specific and is passed unchanged to
#	${tool}-dg-test.  An example is gcc where `keyword' can be any of:
#	preprocess|compile|assemble|link|run
#	and will do one of: produce a .i, produce a .s, produce a .o,
#	produce an a.out, or produce an a.out and run it (the default is
#	compile).
#
# dg-error regexp comment [{ target/xfail selector } [{.|0|linenum}]]
#	indicate an error message <regexp> is expected on this line
#	(the test fails if it doesn't occur)
#	Linenum=0 for general tool messages (eg: -V arg missing).
#	"." means the current line.
#
# dg-warning regexp comment [{ target/xfail selector } [{.|0|linenum}]]
#	indicate a warning message <regexp> is expected on this line
#	(the test fails if it doesn't occur)
#
# dg-bogus regexp comment [{ target/xfail selector } [{.|0|linenum}]]
#	indicate a bogus error message <regexp> use to occur here
#	(the test fails if it does occur)
#
# dg-build regexp comment [{ target/xfail selector }]
#	indicate the build use to fail for some reason
#	(errors covered here include bad assembler generated, tool crashes,
#	and link failures)
#	(the test fails if it does occur)
#
# dg-excess-errors comment [{ target/xfail selector }]
#	indicate excess errors are expected (any line)
#	(this should only be used sparingly and temporarily)
#
# dg-output regexp [{ target selector }]
#	indicate the expected output of the program is <regexp>
#	(there may be multiple occurrences of this, they are concatenated)
#
# dg-final { tcl code }
#	add some tcl code to be run at the end
#	(there may be multiple occurrences of this, they are concatenated)
#	(unbalanced braces must be \-escaped)
#
# "{ target selector }" is a list of expressions that determine whether the
# test succeeds or fails for a particular target, or in some cases whether the
# option applies for a particular target.  If the case of `dg-do' it specifies
# whether the testcase is even attempted on the specified target.
#
# The target selector is always optional.  The format is one of:
#
# { xfail *-*-* ... } - the test is expected to fail for the given targets
# { target *-*-* ... } - the option only applies to the given targets
#
# At least one target must be specified, use *-*-* for "all targets".
# At present it is not possible to specify both `xfail' and `target'.
# "native" may be used in place of "*-*-*".
#
# Example:
#
#       [ ... some complicated code ... ]
#	return a; /* { dg-build "fatal" "ran out of spill regs" { xfail i386-*-* } } */
#
# In this example, the compiler use to crash on the "return a;" for some
# target and that it still does crash on i386-*-*.  Admittedly, this is a
# contrived example.
#
# ??? It might be possible to add additional optional arguments by having
# something like: { dg-error ".*syntax.*" "syntax error" { { foo 1 } ... } }
#
# Callbacks
#
# ${tool}-dg-test testfile do-what-keyword extra-flags
#
#	Run the test, be it compiler, assembler, or whatever.
#
# ${tool}-dg-prune target_triplet text
#
#	Optional callback to delete output from the tool that can occur
#	even in successful ("pass") situations and interfere with output
#	pattern matching.  This also gives the tool an opportunity to review
#	the output and check for any conditions which indicate an "untested"
#	or "unresolved" state.  An example is if a testcase is too big and
#	fills all available ram (which can happen for 16 bit cpus).  The
#	result is either the pruned text or
#	"::untested|unresolved|unsupported::message"
#	(eg: "::unsupported::memory full").
#
# Notes:
# 1) All runnable testcases must return 0 from main() for success.
#    You can't rely on getting any return code from target boards, and the
#    `exec' command says a program fails if it returns non-zero.
#
# Language independence is (theoretically) achieved by:
#
# 1) Using global $tool to indicate the language (eg: gcc, g++, gas, etc.).
#    This should only be used to look up other objects.  We don't want to
#    have to add code for each new language that is supported.  If this is
#    done right, no code needs to be added here for each new language.
#
# 2) Passing tool options in as arguments.
#
#    Earlier versions of ${tool}_start (eg: gcc_start) would only take the name
#    of the file to compile as an argument.  Newer versions accept a list of
#    one or two elements, the second being a string of *all* options to pass
#    to the tool.  We require this facility.
#
# 3) Callbacks.
#
# Try not to do anything else that makes life difficult.
#
# The normal way to write a testsuite is to have a .exp file containing:
#
# load_lib ${tool}-dg.exp
# dg-init
# dg-runtest [lsort [glob -nocomplain $srcdir/$subdir/foo*]] ...
# dg-finish

# Global state variables.
# The defaults are for GCC.

# The default do-what keyword.
set dg-do-what-default compile

# When dg-interpreter-batch-mode is 1, no execution test or excess error 
# tests are performed.
set dg-interpreter-batch-mode 0

# Line number format.  This is how line numbers appear in program output.
set dg-linenum-format ":%d:"
proc dg-format-linenum { linenum } {
    global dg-linenum-format
    return [format ${dg-linenum-format} $linenum]
}

# Useful subroutines.

# dg-get-options -- pick out the dg-xxx options in a testcase
#
# PROG is the file name of the testcase.
# The result is a list of options found.
#
# Example: For the following testcase:
#
# /* { dg-prms-id 1234 } */
# int foo { return 0; } /* { dg-build fatal "some comment" } */
#
# we return:
#
# { dg-prms-id 1 1234 } { dg-build 2 fatal "some comment" }

proc dg-get-options { prog } {
    set result ""

    set tmp [grep $prog "{\[ \t\]\+dg-\[-a-z\]\+\[ \t\]\+.*\[ \t\]\+}" line]
    if ![string match "" $tmp] {
	foreach i $tmp {
	    #send_user "Found: $i\n"
	    # FIXME: When to use "+" and "\+" isn't clear.
	    # Seems to me it took awhile to get this to work.
	    regexp "(\[0-9\]\+)\[ \t\]\+{\[ \t\]+(dg-\[-a-z\]+)\[ \t\]\+(.*)\[ \t\]+}\[^\}\]*(\n|$)" $i i line cmd args
	    #send_user "Found: $cmd $line $args\n"
	    append result " { $cmd $line $args }"
	}
    }

    #send_user "Returning: $result\n"
    return $result
}

#
# Process optional xfail/target arguments
#
# SELECTOR is "xfail target-triplet-1 ..." or "target target-triplet-1 ..."
# `target-triplet' may be "native".
# For xfail, the result is "F" (expected to Fail) if the current target is
# affected, otherwise "P" (expected to Pass).
# For target, the result is "S" (target is Selected) if the target is selected,
# otherwise "N" (target is Not selected).
#
proc dg-process-target { selector } {
    global target_triplet

    set isnative [isnative]
    set triplet_match 0

    #send_user "dg-process-target: $selector\n"

    set selector [string trim $selector]
    if [regexp "^xfail " $selector] {
	set what xfail
    } elseif [regexp "^target " $selector] {
	set what target
    } else {
	# The use of error here and in other dg-xxx utilities is intentional.
	# dg-test will catch them and do the right thing.
	error "syntax error in target selector \"$selector\""
    }

    # ??? This should work but it doesn't.  tcl bug?
    #if [regexp "^${what}(( \[^ \]+-\[^ \]+-\[^ \]+)|( native))+$" $selector tmp selector]
    if [regexp "^${what}( \[^ \]+-\[^ \]+-\[^ \]+| native)+$" $selector] {
	regsub "^${what} " $selector "" selector
	#send_user "selector: $selector\n"
	foreach triplet $selector {
	    if [string match $triplet $target_triplet] {
		set triplet_match 1
	    } elseif { $isnative && $triplet == "native" } {
		set triplet_match 1
	    }
	}
    } else {
	error "syntax error in target selector \"$selector\""
    }

    if { $triplet_match } {
	return [expr { $what == "xfail" ? "F" : "S" }]
    } else {
	return [expr { $what == "xfail" ? "P" : "N" }]
    }
}

# Predefined user option handlers.
# The line number is always the first element.
# Note that each of these are varargs procs (they have an `args' argument).
# Tests for optional arguments are coded with ">=" to simplify adding new ones.

proc dg-prms-id { args } {
    global prms_id	;# this is a testing framework variable

    if { [llength $args] > 2 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    set prms_id [lindex $args 1]
}

#
# Set tool options
#
# Different options can be used for different targets by having multiple
# instances, selecting a different target each time.  Since options are
# processed in order, put the default value first.  Subsequent occurrences
# will override previous ones.
#

proc dg-options { args } {
    upvar dg-extra-tool-flags extra-tool-flags

    if { [llength $args] > 3 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    if { [llength $args] >= 3 } {
	switch [dg-process-target [lindex $args 2]] {
	    "S" { set extra-tool-flags [lindex $args 1] }
	    "N" { }
	    "F" { error "[lindex $args 0]: `xfail' not allowed here" }
	    "P" { error "[lindex $args 0]: `xfail' not allowed here" }
	}
    } else {
	set extra-tool-flags [lindex $args 1]
    }
}

#
# Record what to do (compile/run/etc.)
#
# Multiple instances are supported (since we don't support target and xfail
# selectors on one line), though it doesn't make much sense to change the
# compile/assemble/link/run field.  Nor does it make any sense to have
# multiple lines of target selectors (use one line).
#
proc dg-do { args } {
    upvar dg-do-what do-what

    if { [llength $args] > 3 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    set selected [lindex ${do-what} 1]	;# selected? (""/S/N)
    set expected [lindex ${do-what} 2]	;# expected to pass/fail (P/F)

    if { [llength $args] >= 3 } {
	switch [dg-process-target [lindex $args 2]] {
	    "S" {
		set selected "S"
	    }
	    "N" {
		# Don't deselect a target if it's been explicitly selected,
		# but indicate a specific target has been selected (so don't
		# do this testcase if it's not appropriate for this target).
		# The user really shouldn't have multiple lines of target
		# selectors, but try to do the intuitive thing (multiple lines
		# are OR'd together).
		if { $selected != "S" } {
		    set selected "N"
		}
	    }
	    "F" { set expected "F" }
	    "P" {
		# There's nothing to do for "P".  We don't want to clobber a
		# previous xfail for this target.
	    }
	}
    } else {
	# Note: A previous occurrence of `dg-do' with target/xfail selectors
	# is a user mistake.  We clobber previous values here.
	set selected S
	set expected P
    }

    switch [lindex $args 1] {
	"preprocess" { }
	"compile" { }
	"assemble" { }
	"link" { }
	"run" { }
	default {
	    error "[lindex $args 0]: syntax error"
	}
    }
    set do-what [list [lindex $args 1] $selected $expected]
}

proc dg-error { args } {
    upvar dg-messages messages

    if { [llength $args] > 5 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    set xfail ""
    if { [llength $args] >= 4 } {
	switch [dg-process-target [lindex $args 3]] {
	    "F" { set xfail "X" }
	    "P" { set xfail "" }
	    "N" {
		# If we get "N", this error doesn't apply to us so ignore it.
		return
	    }
	}
    }

    if { [llength $args] >= 5 } {
	switch [lindex $args 4] {
	    "." { set line [dg-format-linenum [lindex $args 0]] }
	    "0" { set line "" }
	    "default" { set line [dg-format-linenum [lindex $args 4]] }
	}
    } else {
	set line [dg-format-linenum [lindex $args 0]]
    }

    lappend messages [list $line "${xfail}ERROR" [lindex $args 1] [lindex $args 2]]
}

proc dg-warning { args } {
    upvar dg-messages messages

    if { [llength $args] > 5 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    set xfail ""
    if { [llength $args] >= 4 } {
	switch [dg-process-target [lindex $args 3]] {
	    "F" { set xfail "X" }
	    "P" { set xfail "" }
	    "N" {
		# If we get "N", this warning doesn't apply to us so ignore it.
		return
	    }
	}
    }

    if { [llength $args] >= 5 } {
	switch [lindex $args 4] {
	    "." { set line [dg-format-linenum [lindex $args 0]] }
	    "0" { set line "" }
	    "default" { set line [dg-format-linenum [lindex $args 4]] }
	}
    } else {
	set line [dg-format-linenum [lindex $args 0]]
    }

    lappend messages [list $line "${xfail}WARNING" [lindex $args 1] [lindex $args 2]]
}

proc dg-bogus { args } {
    upvar dg-messages messages

    if { [llength $args] > 5 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    set xfail ""
    if { [llength $args] >= 4 } {
	switch [dg-process-target [lindex $args 3]] {
	    "F" { set xfail "X" }
	    "P" { set xfail "" }
	    "N" {
		# If we get "N", this message doesn't apply to us so ignore it.
		return
	    }
	}
    }

    if { [llength $args] >= 5 } {
	switch [lindex $args 4] {
	    "." { set line [dg-format-linenum [lindex $args 0]] }
	    "0" { set line "" }
	    "default" { set line [dg-format-linenum [lindex $args 4]] }
	}
    } else {
	set line [dg-format-linenum [lindex $args 0]]
    }

    lappend messages [list $line "${xfail}BOGUS" [lindex $args 1] [lindex $args 2]]
}

proc dg-build { args } {
    upvar dg-messages messages

    if { [llength $args] > 4 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    set xfail ""
    if { [ llength $args] >= 4 } {
	switch [dg-process-target [lindex $args 3]] {
	    "F" { set xfail "X" }
	    "P" { set xfail "" }
	    "N" {
		# If we get "N", this lossage doesn't apply to us so ignore it.
		return
	    }
	}
    }

    lappend messages [list [lindex $args 0] "${xfail}BUILD" [lindex $args 1] [lindex $args 2]]
}

proc dg-excess-errors { args } {
    upvar dg-excess-errors-flag excess-errors-flag

    if { [llength $args] > 3 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    if { [llength $args] >= 3 } {
	switch [dg-process-target [lindex $args 2]] {
	    "F" { set excess-errors-flag 1 }
	    "S" { set excess-errors-flag 1 }
	}
    } else {
	set excess-errors-flag 1
    }
}

#
# Indicate expected program output
#
# We support multiple occurrences, but we do not implicitly insert newlines
# between them.
#
# Note that target boards don't all support this kind of thing so it's a good
# idea to specify the target all the time.  If one or more targets are
# explicitly selected, the test won't be performed if we're not one of them
# (as long as we were never mentioned).
#
# If you have target dependent output and want to set an xfail for one or more
# of them, use { dg-output "" { xfail a-b-c ... } }.  The "" won't contribute
# to the expected output.
#
proc dg-output { args } {
    upvar dg-output-text output-text

    if { [llength $args] > 3 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    # Allow target dependent output.

    set expected [lindex ${output-text} 0]
    if { [llength $args] >= 3 } {
	switch [dg-process-target [lindex $args 2]] {
	    "N" { return }
	    "S" { }
	    "F" { set expected "F" }
	    # Don't override a previous xfail.
	    "P" { }
	}
    }

    if { [llength ${output-text}] == 1 } {
	# First occurrence.
	set output-text [list $expected [lindex $args 1]]
    } else {
	set output-text [list $expected "[lindex ${output-text} 1][lindex $args 1]"]
    }
}

proc dg-final { args } {
    upvar dg-final-code final-code

    if { [llength $args] > 2 } {
	error "[lindex $args 0]: too many arguments"
	return
    }

    #send_user "dg-final: $args\n"
    append final-code "[lindex $args 1]\n"
}

#
# Set up our environment
#
# There currently isn't much to do, but always calling it allows us to add
# enhancements without having to update our callers.
# It must be run before calling `dg-test'.

proc dg-init { } {
}

# dg-runtest -- simple main loop useful to most testsuites
#
# FLAGS is a set of options to always pass.
# DEFAULT_EXTRA_FLAGS is a set of options to pass if the testcase doesn't
# specify any (with dg-option).
# ??? We're flipping between "flag" and "option" here.

proc dg-runtest { testcases flags default-extra-flags } {
    global runtests

    foreach testcase $testcases {
	# If we're only testing specific files and this isn't one of them, skip it.
	if ![runtest_file_p $runtests $testcase] {
	    continue
	}
	verbose "Testing [file tail [file dirname $testcase]]/[file tail $testcase]"
	dg-test $testcase $flags ${default-extra-flags}
    }
}

# dg-trim-dirname -- rip DIR_NAME out of FILE_NAME
# 
# Syntax: dg-trim-dirname dir_name file_name
# We need to go through this contorsion in order to properly support
# directory-names which might have embedded regexp special characters.

proc dg-trim-dirname { dir_name file_name } {
    set special_character "\[\?\+\-\.\(\)\$\|\]"
    regsub -all $special_character $dir_name "\\\\&" dir_name
    regsub "^$dir_name/?" $file_name "" file_name
    return $file_name
}

# dg-test -- runs a new style DejaGnu test
#
# Syntax: dg-test [-keep-output] prog tool_flags default_extra_tool_flags
#
# PROG is the full path name of the file to pass to the tool (eg: compiler).
# TOOL_FLAGS is a set of options to always pass.
# DEFAULT_EXTRA_TOOL_FLAGS are additional options if the testcase has none.

#proc dg-test { prog tool_flags default_extra_tool_flags } {
proc dg-test { args } {
    global dg-do-what-default dg-interpreter-batch-mode dg-linenum-format
    global errorCode errorInfo
    global tool
    global srcdir		;# eg: /calvin/dje/devo/gcc/./testsuite/
    global host_triplet target_triplet

    set keep 0
    set i 0

    if { [string index [lindex $args 0] 0] == "-" } {
	for { set i 0 } { $i < [llength $args] } { incr i } {
	    if { [lindex $args $i] == "--" } {
		incr i
		break
	    } elseif { [lindex $args $i] == "-keep-output" } {
		set keep 1
	    } elseif { [string index [lindex $args $i] 0] == "-" } {
		clone_output "ERROR: dg-test: illegal argument: [lindex $args $i]"
		return
	    } else {
		break
	    }
	}
    }

    if { $i + 3 != [llength $args] } {
	clone_output "ERROR: dg-test: missing arguments in call"
	return
    }
    set prog [lindex $args $i]
    set tool_flags [lindex $args [expr $i + 1]]
    set default_extra_tool_flags [lindex $args [expr $i + 2]]

    set text "\[- A-Za-z0-9\.\;\"\_\:\'\`\(\)\!\#\=\+\?\&\*]*"

    set name [dg-trim-dirname $srcdir $prog]
    # If we couldn't rip $srcdir out of `prog' then just do the best we can.
    # The point is to reduce the unnecessary noise in the logs.  Don't strip
    # out too much because different testcases with the same name can confuse
    # `test-tool'.
    if [string match "/*" $name] {
	set name "[file tail [file dirname $prog]]/[file tail $prog]"
    }

    # Process any embedded dg options in the testcase.

    # Use "" for the second element of dg-do-what so we can tell if it's been
    # explicitly set to "S".
    set dg-do-what [list ${dg-do-what-default} "" P]
    set dg-excess-errors-flag 0
    set dg-messages ""
    set dg-extra-tool-flags $default_extra_tool_flags
    set dg-final-code ""

    # `dg-output-text' is a list of two elements: pass/fail and text.
    # Leave second element off for now (indicates "don't perform test")
    set dg-output-text "P"

    # Define our own "special function" `unknown' so we catch spelling errors.
    # But first rename the existing one so we can restore it afterwards.
    catch {rename dg-save-unknown ""}
    rename unknown dg-save-unknown
    proc unknown { args } {
	return -code error "unknown dg option: $args"
    }

    set tmp [dg-get-options $prog]
    foreach op $tmp {
	verbose "Processing option: $op" 3
	set status [catch "$op" errmsg]
	if { $status != 0 } {
	    if { 0 && [info exists errorInfo] } {
		# This also prints a backtrace which will just confuse
		# testcase writers, so it's disabled.
		perror "$name: $errorInfo\n"
	    } else {
		perror "$name: $errmsg for \"$op\"\n"
	    }
	    # ??? The call to unresolved here is necessary to clear `errcnt'.
	    # What we really need is a proc like perror that doesn't set errcnt.
	    # It should also set exit_status to 1.
	    unresolved "$name: $errmsg for \"$op\""
	    return
	}
    }

    # Restore normal error handling.
    rename unknown ""
    rename dg-save-unknown unknown

    # If we're not supposed to try this test on this target, we're done.
    if { [lindex ${dg-do-what} 1] == "N" } {
	unsupported "$name"
	verbose "$name not supported on this target, skipping it" 3
	return
    }

    # Run the tool and analyze the results.
    # The result of ${tool}-dg-test is in a bit of flux.
    # Currently it is the name of the output file (or "" if none).
    # If we need more than this it will grow into a list of things.
    # No intention is made (at this point) to preserve upward compatibility
    # (though at some point we'll have to).

    set results [${tool}-dg-test $prog [lindex ${dg-do-what} 0] "$tool_flags ${dg-extra-tool-flags}"];

    set comp_output [lindex $results 0];
    set output_file [lindex $results 1];

    #send_user "\nold_dejagnu.exp: comp_output1 = :$comp_output:\n\n"
    #send_user "\nold_dejagnu.exp: message = :$message:\n\n"
    #send_user "\nold_dejagnu.exp: message length = [llength $message]\n\n"

    foreach i ${dg-messages} {
	verbose "Scanning for message: $i" 4

	# Remove all error messages for the line [lindex $i 0]
	# in the source file.  If we find any, success!
	set line [lindex $i 0]
	set pattern [lindex $i 2]
	set comment [lindex $i 3]
	#send_user "Before:\n$comp_output\n"
	if [regsub -all "(^|\n)(\[^\n\]+$line\[^\n\]*($pattern)\[^\n\]*\n?)+" $comp_output "\n" comp_output] {
            set comp_output [string trimleft $comp_output]
	    set ok pass
	    set uhoh fail
	} else {
	    set ok fail
	    set uhoh pass
	}
	#send_user "After:\n$comp_output\n"

	# $line will either be a formatted line number or a number all by
	# itself.  Delete the formatting.
	scan $line ${dg-linenum-format} line
	switch [lindex $i 1] {
	    "ERROR" {
		$ok "$name $comment (test for errors, line $line)"
	    }
	    "XERROR" {
		x$ok "$name $comment (test for errors, line $line)"
	    }
	    "WARNING" {
		$ok "$name $comment (test for warnings, line $line)"
	    }
	    "XWARNING" {
		x$ok "$name $comment (test for warnings, line $line)"
	    }
	    "BOGUS" {
		$uhoh "$name $comment (test for bogus messages, line $line)"
	    }
	    "XBOGUS" {
		x$uhoh "$name $comment (test for bogus messages, line $line)"
	    }
	    "BUILD" {
		$uhoh "$name $comment (test for build failure, line $line)"
	    }
	    "XBUILD" {
		x$uhoh "$name $comment (test for build failure, line $line)"
	    }
	    "EXEC" { }
	    "XEXEC" { }
	}
	#send_user "\nold_dejagnu.exp: comp_output2= :$comp_output:\n\n"
    }
    #send_user "\nold_dejagnu.exp: comp_output3 = :$comp_output:\n\n"

    # Remove messages from the tool that we can ignore.
    #send_user "comp_output: $comp_output\n"
    set comp_output [prune_warnings $comp_output]

    if { [info proc ${tool}-dg-prune] != "" } {
	set comp_output [${tool}-dg-prune $target_triplet $comp_output]
	switch -glob $comp_output {
	    "::untested::*" {
		regsub "::untested::" $comp_output "" message
		untested "$name: $message"
		return
	    }
	    "::unresolved::*" {
		regsub "::unresolved::" $comp_output "" message
		unresolved "$name: $message"
		return
	    }
	    "::unsupported::*" {
		regsub "::unsupported::" $comp_output "" message
		unsupported "$name: $message"
		return
	    }
	}
    }

    # See if someone forgot to delete the extra lines.
    regsub -all "\n+" $comp_output "\n" comp_output
    regsub "^\n+" $comp_output "" comp_output
    #send_user "comp_output: $comp_output\n"

    # Don't do this if we're testing an interpreter.
    # FIXME: why?
    if { ${dg-interpreter-batch-mode} == 0 } {
	# Catch excess errors (new bugs or incomplete testcases).
	if ${dg-excess-errors-flag} {
	    setup_xfail "*-*-*"
	}
	if ![string match "" $comp_output] {
	    fail "$name (test for excess errors)"
	    send_log "Excess errors:\n$comp_output\n"
	} else {
	    pass "$name (test for excess errors)"
	}
    }

    # Run the executable image if asked to do so.
    # FIXME: This is the only place where we assume a standard meaning to
    # the `keyword' argument of dg-do.  This could be cleaned up.
    if { [lindex ${dg-do-what} 0] == "run" } {
	if ![file exists $output_file] {
	    warning "$name compilation failed to produce executable"
	} else {
	    set status -1
	    set result [${tool}_load $output_file]
	    set status [lindex $result 0];
	    set output [lindex $result 1];
	    #send_user "After exec, status: $status\n"
	    if { [lindex ${dg-do-what} 2] == "F" } {
		setup_xfail "*-*-*"
	    }
	    if { "$status" == "pass" } {
		pass "$name execution test"
		verbose "Exec succeeded." 3
		if { [llength ${dg-output-text}] > 1 } {
		    #send_user "${dg-output-text}\n"
		    if { [lindex ${dg-output-text} 0] == "F" } {
			setup_xfail "*-*-*"
		    }
		    set texttmp [lindex ${dg-output-text} 1]
		    if { ![regexp $texttmp ${output}] } {
			fail "$name output pattern test, is ${output}, should match $texttmp"
			verbose "Failed test for output pattern $texttmp" 3
		    } else {
			pass "$name output pattern test, $texttmp"
			verbose "Passed test for output pattern $texttmp" 3
		    }
		    unset texttmp
		}
	    } elseif { "$status" == "fail" } {
		# It would be nice to get some info out of errorCode.
		if [info exists errorCode] {
		    verbose "Exec failed, errorCode: $errorCode" 3
		} else {
		    verbose "Exec failed, errorCode not defined!" 3
		}
		fail "$name execution test"
	    } else {
		$status "$name execution test"
	    }
	}
    }

    # Are there any further tests to perform?
    # Note that if the program has special run-time requirements, running
    # of the program can be delayed until here.  Ditto for other situations.
    # It would be a bit cumbersome though.

    if ![string match ${dg-final-code} ""] {
	regsub -all "\\\\(\[{}\])" ${dg-final-code} "\\1" dg-final-code
	# Note that the use of `args' here makes this a varargs proc.
	proc dg-final-proc { args } ${dg-final-code}
	verbose "Running dg-final tests." 3
	verbose "dg-final-proc:\n[info body dg-final-proc]" 4
	if [catch "dg-final-proc $prog" errmsg] {
	    perror "$name: error executing dg-final: $errmsg"
	    # ??? The call to unresolved here is necessary to clear `errcnt'.
	    # What we really need is a proc like perror that doesn't set errcnt.
	    # It should also set exit_status to 1.
	    unresolved "$name: error executing dg-final: $errmsg"
	}
    }

    # Do some final clean up.
    # When testing an interpreter, we don't compile something and leave an
    # output file.
    if { ! ${keep} && ${dg-interpreter-batch-mode} == 0 } {
	catch "exec rm -f $output_file"
    }
}

#
# Do any necessary cleanups
#
# This is called at the end to undo anything dg-init did (that needs undoing).
#
proc dg-finish { } {
    # Reset this in case caller wonders whether s/he should.
    global prms_id
    set prms_id 0

    # The framework doesn't like to see any error remnants, so remove them.
    global errorInfo
    if [info exists errorInfo] {
	unset errorInfo
    }

    # If the tool has a "finish" routine, call it.
    # There may be a bit of duplication (eg: resetting prms_id), leave it.
    # Let's keep these procs robust.
    global tool
    if ![string match "" [info procs ${tool}_finish]] {
	${tool}_finish
    }
}