Compares the result file produced for each test description in a
collection with its corresponding benchmark file.
Format
COMPARE collection-name [/qualifier...]
1 – Command Qualifiers
1.1 /CHARACTERS
Performs a character-by-character comparison of the results file with the benchmark file. The default is /SCREENS for interactive tests and /RECORDS for noninteractive tests.
1.2 /FULL
For noninteractive and interactive terminal tests, the /FULL qualifier includes a complete listing of the text in the difference file that was identical and a listing of the differences encountered when the result file and benchmark file are compared.
1.3 /IGNORE
/IGNORE=keyword
The /IGNORE qualifier enables you to specify that various aspects
of benchmark and result files are to be ignored during comparison.
The following keywords apply to noninteractive and interactive
terminal tests, only.
Keyword Result
CASE Ignores any differences between the case of
alphabetic characters (A,a,B,b, .. . )
FORM-FEEDS Ignores form-feed characters
LEADING_BLANKS Ignores leading blanks and tabs
SPACING Treats multiple blanks and tabs as a single space
TRAILING_BLANKS Ignores trailing blanks and tabs
For interactive terminal tests, if you specify the /IGNORE and
/SCREENS qualifiers together, Digital Test Manager performs the
comparison screen by screen and ignores the /IGNORE qualifier.
If you specify more than one keyword, separate the keywords with
commas and enclose the list in parentheses. The output file (your
result file) is not changed in any way by the /IGNORE qualifier.
The following keyword applies to DECwindows tests only.
Keyword Result
MASK Ignores masked areas defined on DECwindows
benchmark images
1.4 /LOG
/LOG (D) /NOLOG Controls whether Digital Test Manager displays informational and success messages on your screen.
1.5 /MASK
/MASK=(keyword,...)
The /MASK qualifier selects one or more sets of mask patterns that
will cause certain run-time data to be ignored during the comparison
of screens for interactive terminal tests. Such run-time data is
ignored by substituting blank space for patterns in both the
benchmark and result screens before they are compared.
The valid values for keyword are as follows:
Keyword Mask
ALL Specifies that all the masks in this table be used
DATE The date mask ignores comparison of date stamps
The following list shows some examples of the date
patterns to be ignored; this list is not all
inclusive.
dd-mmm-yyyy ( example: 17-OCT-1998 )
dd mmm yy ( example: 17 OCT 98 )
yy.mmm.dd ( example: 98.OCT.17 )
mm/dd/yy ( example: 10/17/98 )
yyyy/mm/dd ( example: 1998/10/17 )
month day, year ( example: October 17, 1998 )
month day, year ( example: Oct. 17, 1998 )
day.month.year ( example: 17.October.1998 )
year-month-day ( example: 98-October-17 )
TIME Ignores time stamps with the following forms:
hh:mm:ss.xxxx ( example: 15:37:53.22 )
hh:mm:ss ( example: 15:37:53 )
hh:mm ( example: 15:37 )
hh:mm xm ( example: 3:37 PM )
hhHmmm ( example: 15H37m )
hhHmm' ( example: 15H37' )
hh.mm h ( example: 15.37 h )
hh h mm"ss s ( example: 15 h 37"53 s )
hh h mm min ( example: 15 h 37 min )
kl hh.mm ( example: kl 15.37 )
h hh.mm ( example: h 15.37 )
VERSION Ignores file versions on file names
FILE_NAMES Ignores file names of the form FILENAME.EXT
DIRECTORIES Ignores directory specification fields of
the form DISK:[DIRECTORY]
If you specify more than one keyword, separate the keywords with
commas and enclose the list in parentheses. If you specify only
one keyword, omit the parentheses. Masking of is performed in
the order of the keywords shown above.
1.6 /PARALLEL
/PARALLEL /NOPARALLEL (D) Specifies whether the lines that do not match in the result and benchmark files are formatted side by side. If you specify the /[NO]PARALLEL and /SCREENS qualifiers together, Digital Test Manager performs the comparison screen by screen and ignores the /[NO]PARALLEL qualifier.
1.7 /RECORDS
For noninteractive and interactive terminal tests, the /RECORDS qualifier performs a record-by-record comparison of the result and benchmark files. The default is /SCREENS for interactive tests and /RECORDS for noninteractive tests. Records are identical only if they contain the same characters. Use this type of comparison only when you expect the record in which a string appears to be the same each time a comparison is performed. Use caution when specifying the /RECORDS qualifier for an interactive test because the records in the result file are not guaranteed to be written the same way each time the test runs. You might want to use the /RECORDS qualifier for an interactive test whose result file is not generated by Digital Test Manager, for example, if you rename a test output file to be DTM$RESULT.
1.8 /SCREENS
Performs a screen-by-screen comparison of the result and benchmark files for an interactive test. The default is /SCREENS for interactive tests and /RECORDS for noninteractive tests.
1.9 /SENTINEL
/SENTINEL=("begin-delimiter","end-delimiter")
Specifies a pair of strings used to delimit a section of text to
be ignored during the comparison of result and benchmark files for
noninteractive tests. The delimiters can be up to 256 characters
per line, and must be unique. Any text between and including the
delimiters is treated as if it did not exist.
If you do not enclose the sentinel strings in quotation marks,
they are converted to uppercase before the comparison of the
files. Sentinel strings may contain any characters, but if you
include spaces, tabs or commas, they must be enclosed in
quotation marks.
1.10 /WIDTH
/WIDTH=n For noninteractive and interactive terminal tests that were compared with the /CHARACTERS or /RECORDS qualifier, the /WIDTH qualifier specifies the maximum width allowed for the differences report. The minimum width is 48 columns and the maximum width is 511 columns. The default value is 132 columns.
2 – Example
DTM> COMPARE MAIL_COLL %DTM-I-SUCCEEDED, the comparison for the test MAIL_TEST succeeded %DTM-I-SUCCEEDED, the comparison for the test SEND_MAIL_TEST succeeded %DTM-S-COMPARED, collection MAIL_COLL compared DTM> This example compares the results for all tests in the collection MAIL_COLL. For each test, Digital Test Manager deletes the result files for tests whose benchmark and result files match, and it creates a difference file for tests whose benchmark and result files differ.