+dnl PSPP - a program for statistical analysis.
+dnl Copyright (C) 2017 Free Software Foundation, Inc.
+dnl
+dnl This program is free software: you can redistribute it and/or modify
+dnl it under the terms of the GNU General Public License as published by
+dnl the Free Software Foundation, either version 3 of the License, or
+dnl (at your option) any later version.
+dnl
+dnl This program is distributed in the hope that it will be useful,
+dnl but WITHOUT ANY WARRANTY; without even the implied warranty of
+dnl MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+dnl GNU General Public License for more details.
+dnl
+dnl You should have received a copy of the GNU General Public License
+dnl along with this program. If not, see <http://www.gnu.org/licenses/>.
+dnl
AT_BANNER([syntax segmentation])
m4_define([PSPP_CHECK_SEGMENT],
- [AT_CHECK([segment-test $1 input], [0], [expout])
- AT_CHECK([segment-test -1 $1 input], [0], [expout])
- AT_CHECK([segment-test -0 $1 input])
- AT_CHECK([segment-test -01 $1 input])])
+ [AT_CAPTURE_FILE([input])
+ for strip in "" "-s"; do
+ case $strip in # (
+ '') sed 's/^-//' < expout-base > expout ;; # (
+ -s) sed '/^-/d' < expout-base > expout ;;
+ esac
+ AT_CHECK([segment-test $1 $strip input], [0], [expout])
+ AT_CHECK([segment-test -1 $strip $1 input], [0], [expout])
+ AT_CHECK([segment-test -0 $strip $1 input])
+ AT_CHECK([segment-test -01 $strip $1 input])
+ done])
\f
AT_SETUP([identifiers])
AT_KEYWORDS([segment])
AT_DATA([input], [dnl
-a ab abc abcd
-A AB ABC ABCD
-aB aBC aBcD
-$x $y $z
+a ab abc abcd !abcd
+A AB ABC ABCD !ABCD
+aB aBC aBcD !aBcD
+$x $y $z !$z
grève@<00A0>@Ângstrom@<00A0>@poté
-#a #b #c ## #d
-@efg @ @@. @#@ @&t@
+#a #b #c ## #d !#d
+@efg @ @@. @#@ !@ @&t@
## # #12345 #.#
f@#_.#6
GhIjK
.x 1y _z
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
identifier a space
identifier ab space
identifier abc space
-identifier abcd
+identifier abcd space
+macro_id !abcd
newline \n (later)
identifier A space
identifier AB space
identifier ABC space
-identifier ABCD
+identifier ABCD space
+macro_id !ABCD
newline \n (later)
identifier aB space
identifier aBC space
-identifier aBcD
+identifier aBcD space
+macro_id !aBcD
newline \n (later)
identifier $x space
identifier $y space
-identifier $z
+identifier $z space
+macro_id !$z
newline \n (later)
identifier grève
identifier #b space
identifier #c space
identifier ## space
-identifier #d
+identifier #d space
+macro_id !#d
newline \n (later)
identifier @efg space
identifier @ space
identifier @@. space
identifier @#@ space
+macro_id !@ space
newline \n (later)
identifier ## space
identifier x space
number 1
identifier y space
-unexpected_char \_
+punct \_
identifier z
-newline \n (later)
-
-end <U+0000>
+-newline \n (later)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
WXYZ. /* unterminated end of line comment
WxYz./* unterminated end of line comment @&t@
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
identifier abcd. space
identifier abcd
end_command .
identifier WxYz
end_command .
comment /*_unterminated_end_of_line_comment_
-newline \n (first)
-
-end <U+0000>
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
andx orx notx eqx gex gtx lex ltx nex allx byx tox withx
and. with.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
reserved_word and space
reserved_word or space
reserved_word not space
identifier and. space
reserved_word with
end_command .
-newline \n (first)
-
-end <U+0000>
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
AT_KEYWORDS([segment])
AT_DATA([input], [dnl
~ & | = >= > <= < ~= <> ( ) , - + * / [[ ]] **
-~&|=>=><=<~=<>(),-+*/[[]]**
+~&|=>=><=<~=<>(),-+*/[[]]**!*
+% : ; ? _ ` { } ~ !*
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
punct ~ space
punct & space
punct | space
punct [[
punct ]]
punct **
+macro_id !*
newline \n (later)
-end <U+0000>
+punct % space
+punct : space
+punct ; space
+punct ? space
+punct \_ space
+punct ` space
+punct { space
+punct } space
+punct ~ space
+macro_id !*
+-newline \n (later)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
\f
-AT_SETUP([numbers])
+AT_SETUP([positive numbers])
AT_KEYWORDS([segment])
AT_DATA([input], [dnl
0 1 01 001. 1.
5e1 6E-1 7e+1 6E+01 6e-03
.3E1 .4e-1 .5E+1 .6e+01 .7E-03
1.23e1 45.6E-1 78.9e+1 99.9E+01 11.2e-03
-. 1e e1 1e+ 1e-
+. 1e e1 1e+ 1e- 1.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
number 0 space
number 1 space
number 01 space
expected_exponent 1e space
identifier e1 space
expected_exponent 1e+ space
-expected_exponent 1e-
+expected_exponent 1e- space
+number 1
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([negative numbers])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+ -0 -1 -01 -001. -1.
+ -123. /* comment 1 */ /* comment 2 */
+ -.1 -0.1 -00.1 -00.10
+ -5e1 -6E-1 -7e+1 -6E+01 -6e-03
+ -.3E1 -.4e-1 -.5E+1 -.6e+01 -.7E-03
+ -1.23e1 -45.6E-1 -78.9e+1 -99.9E+01 -11.2e-03
+ -/**/1
+ -. -1e -e1 -1e+ -1e- -1.
+])
+AT_DATA([expout-base], [dnl
+spaces _
+number -0 space
+number -1 space
+number -01 space
+number -001. space
+number -1
+end_command .
+newline \n (first)
+ space
+number -123
+end_command . space
+comment /*_comment_1_*/ space
+comment /*_comment_2_*/
+newline \n (first)
+ space
+number -.1 space
+number -0.1 space
+number -00.1 space
+number -00.10
newline \n (later)
-
-end <U+0000>
+ space
+number -5e1 space
+number -6E-1 space
+number -7e+1 space
+number -6E+01 space
+number -6e-03
+newline \n (later)
+ space
+number -.3E1 space
+number -.4e-1 space
+number -.5E+1 space
+number -.6e+01 space
+number -.7E-03
+newline \n (later)
+ space
+number -1.23e1 space
+number -45.6E-1 space
+number -78.9e+1 space
+number -99.9E+01 space
+number -11.2e-03
+newline \n (later)
+ space
+punct -
+comment /**/
+number 1
+newline \n (later)
+ space
+punct -
+punct . space
+expected_exponent -1e space
+punct -
+identifier e1 space
+expected_exponent -1e+ space
+expected_exponent -1e- space
+number -1
+end_command .
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
+
\f
AT_SETUP([strings])
AT_KEYWORDS([segment])
+ /* also a punctuator on blank line
- 'new command'
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
quoted_string 'x' space
quoted_string "y" space
quoted_string 'abc'
start_command - space
quoted_string 'new_command'
-newline \n (later)
-
-end <U+0000>
+-newline \n (later)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
title my title.
#! /usr/bin/pspp
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
shbang #!_/usr/bin/pspp
newline \n (first)
identifier title space
-unquoted_string my_title
+identifier my space
+identifier title
end_command .
newline \n (first)
identifier #
-unexpected_char ! space
+macro_id ! space
punct /
identifier usr
punct /
identifier bin
punct /
identifier pspp
-newline \n (later)
-
-end <U+0000>
+-newline \n (later)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
next command.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
comment_command *_Comment_commands_"don't
newline \n (COMMENT)
end_command .
newline \n (first)
-separate_commands
+separate_commands
newline \n (first)
spaces ___
end_command .
newline \n (first)
-separate_commands
+separate_commands
newline \n (first)
comment_command *_Comment_ends_with_blank_line
end_command .
newline \n (first)
-separate_commands
-newline \n (first)
-
-end <U+0000>
+-separate_commands
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
second paragraph.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
start_document
document DOCUMENT_one_line.
-end_command
+end_command
separate_commands
newline \n (first)
newline \n (DOCUMENT)
document ____________line.
-end_command
+end_command
separate_commands
newline \n (first)
newline \n (DOCUMENT)
document second_paragraph.
-end_command
-separate_commands
-newline \n (first)
-
-end <U+0000>
+-end_command
+-separate_commands
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
\f
-AT_SETUP([TITLE, SUBTITLE, FILE LABEL commands])
+AT_SETUP([FILE LABEL command])
AT_KEYWORDS([segment])
AT_DATA([input], [dnl
-title/**/'Quoted string title'.
-tit /*
-"Quoted string on second line".
-sub "Quoted string subtitle"
- .
-
-TITL /* Not a */ quoted string title.
-SUBT Not a quoted string /* subtitle
-
FIL label isn't quoted.
FILE
lab 'is quoted'.
/**/ lab not quoted here either
])
-AT_DATA([expout], [dnl
-identifier title
-comment /**/
-quoted_string 'Quoted_string_title'
-end_command .
-newline \n (first)
-
-identifier tit space
-comment /*
-newline \n (later)
-
-quoted_string "Quoted_string_on_second_line"
-end_command .
-newline \n (first)
-
-identifier sub space
-quoted_string "Quoted_string_subtitle"
-newline \n (later)
- space
-end_command .
-newline \n (first)
-
-separate_commands
-newline \n (first)
-
-identifier TITL space
-unquoted_string /*_Not_a_*/_quoted_string_title
-end_command .
-newline \n (first)
-
-identifier SUBT space
-unquoted_string Not_a_quoted_string_/*_subtitle
-newline \n (later)
-
-separate_commands
-newline \n (first)
-
+AT_DATA([expout-base], [dnl
identifier FIL space
identifier label space
unquoted_string isn't_quoted
unquoted_string not_quoted_here_either
newline \n (later)
-separate_commands
-newline \n (first)
-
-end <U+0000>
+-separate_commands
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
begin data 123.
not data
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
identifier begin space
identifier data
end_command .
inline_data 5_6_7_/*_x
newline \n (data)
-inline_data
+inline_data
newline \n (data)
inline_data end__data
reserved_word not space
identifier data
-newline \n (later)
-
-end <U+0000>
+-newline \n (later)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
inner command.
end repeat.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
identifier do space
identifier repeat space
identifier x
identifier end space
identifier repeat
end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DO REPEAT command in batch mode])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+do repeat x=a b c
+ y=d e f
+do repeat a=1 thru 5
+another command
+second command
++ third command
+end /* x */ /* y */ repeat print
+end
+ repeat
+do
+ repeat #a=1
+
+ inner command
+end repeat
+])
+AT_DATA([expout-base], [dnl
+identifier do space
+identifier repeat space
+identifier x
+punct =
+identifier a space
+identifier b space
+identifier c
+newline \n (later)
+
+spaces __________
+identifier y
+punct =
+identifier d space
+identifier e space
+identifier f
+newline \n (later)
+
+start_command
+do_repeat_command do_repeat_a=1_thru_5
+newline \n (DO REPEAT)
+
+do_repeat_command another_command
+newline \n (DO REPEAT)
+
+do_repeat_command second_command
+newline \n (DO REPEAT)
+
+do_repeat_command +_third_command
+newline \n (DO REPEAT)
+
+do_repeat_command end_/*_x_*/_/*_y_*/_repeat_print
+newline \n (DO REPEAT)
+
+identifier end
+newline \n (later)
+ space
+identifier repeat
+newline \n (later)
+
+start_command
+identifier do
+newline \n (later)
+
+spaces __
+identifier repeat space
+identifier #a
+punct =
+number 1
+newline \n (later)
+
+separate_commands
+newline \n (DO REPEAT)
+
+do_repeat_command __inner_command
+newline \n (DO REPEAT)
+
+identifier end space
+identifier repeat
+-newline \n (later)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-b])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - simple])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1()
+var1 var2 var3 "!enddefine"
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_body var1_var2_var3_"!enddefine"
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - no newline after parentheses])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1() var1 var2 var3 /* !enddefine
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+macro_body _var1_var2_var3_/*_!enddefine
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - no newline before !ENDDEFINE])
+AT_KEYWORDS([segment ENDDEFINE])
+AT_DATA([input], [dnl
+define !macro1()
+var1 var2 var3!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_body var1_var2_var3
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - all on one line])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1()var1 var2 var3!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+macro_body var1_var2_var3
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - empty])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1()
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - blank lines])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1()
+
+
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_body
+newline \n (DEFINE)
+
+macro_body
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - arguments])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1(a(), b(), c())
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+identifier a
+punct (
+punct )
+punct , space
+identifier b
+punct (
+punct )
+punct , space
+identifier c
+punct (
+punct )
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - multiline arguments])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1(
+ a(), b(
+ ),
+ c()
+)
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+newline \n (later)
+
+spaces __
+identifier a
+punct (
+punct )
+punct , space
+identifier b
+punct (
+newline \n (later)
+
+spaces __
+punct )
+punct ,
+newline \n (later)
+
+spaces __
+identifier c
+punct (
+punct )
+newline \n (later)
+
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - arguments start on second line])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1
+(x,y,z
+)
+content 1
+content 2
+!enddefine.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+newline \n (later)
+
+punct (
+identifier x
+punct ,
+identifier y
+punct ,
+identifier z
+newline \n (later)
+
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_body content_1
+newline \n (DEFINE)
+
+macro_body content_2
+newline \n (DEFINE)
+
+macro_id !enddefine
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - early end of command 1])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1.
+data list /x 1.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+end_command .
+newline \n (first)
+
+identifier data space
+identifier list space
+punct /
+identifier x space
+number 1
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - early end of command 2])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1
+x.
+data list /x 1.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+newline \n (later)
+
+identifier x
+end_command .
+newline \n (first)
+
+identifier data space
+identifier list space
+punct /
+identifier x space
+number 1
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - early end of command 3])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1(.
+x.
+data list /x 1.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+end_command .
+newline \n (first)
+
+identifier x
+end_command .
+newline \n (first)
+
+identifier data space
+identifier list space
+punct /
+identifier x space
+number 1
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - early end of command 4])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+dnl Notice the command terminator at the end of the DEFINE command,
+dnl which should not be there and ends it early.
+define !macro1.
+data list /x 1.
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+end_command .
newline \n (first)
-end <U+0000>
+identifier data space
+identifier list space
+punct /
+identifier x space
+number 1
+end_command .
+-newline \n (first)
+-
+end
+])
+PSPP_CHECK_SEGMENT([-i])
+AT_CLEANUP
+\f
+AT_SETUP([DEFINE command - missing !ENDDEFINE])
+AT_KEYWORDS([segment])
+AT_DATA([input], [dnl
+define !macro1()
+content line 1
+content line 2
+])
+AT_DATA([expout-base], [dnl
+identifier define space
+macro_name !macro1
+punct (
+punct )
+spaces
+newline \n (DEFINE)
+
+macro_body content_line_1
+newline \n (DEFINE)
+
+macro_body content_line_2
+-newline \n (DEFINE)
+-
+end
])
PSPP_CHECK_SEGMENT([-i])
AT_CLEANUP
fourth command.
fifth command.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
identifier first space
identifier command
newline \n (later)
identifier fifth space
identifier command
end_command .
-newline \n (first)
-
-end <U+0000>
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-b])
AT_CLEANUP
fourth command.
fifth command.
])
-AT_DATA([expout], [dnl
+AT_DATA([expout-base], [dnl
identifier command
newline \n (later)
identifier fifth space
identifier command
end_command .
-newline \n (first)
-
-end <U+0000>
+-newline \n (first)
+-
+end
])
PSPP_CHECK_SEGMENT([-a])
AT_CLEANUP
+\f
+# This checks for regression against bug #61253. To see the read of
+# uninitialized data, run with valgrind. The test will pass either
+# way. (The bug report has a more complicated crashing case.)
+AT_SETUP([input ends in carriage return])
+printf '\r' > input
+AT_DATA([expout-base], [dnl
+separate_commands
+spaces \r
+end
+])
+PSPP_CHECK_SEGMENT
+AT_CLEANUP