1 2000-03-27 Paul D. Smith <psmith@gnu.org>
3 * scripts/features/statipattrules: Test that static pattern rules
4 whose prerequisite patterns resolve to empty strings throw an
5 error (instead of dumping core). Fixes PR/1670.
7 * scripts/features/reinvoke: Make more robust by touching "b"
8 first, to ensure it's not newer than "a".
9 Reported by Marco Franzen <Marco.Franzen@Thyron.com>.
11 * scripts/functions/call: Whoops. The fix to PR/1527 caused
12 recursive invocations of $(call ...) to break. I can't come up
13 with any way to get both working at the same time, so I backed out
14 the fix to 1527 and added a test case for recursive calls. This
15 also tests the fix for PR/1610.
17 * scripts/features/double_colon: Test that circular dependencies
18 in double-colon rule sets are detected correctly (PR/1671).
20 2000-03-26 Paul D. Smith <psmith@gnu.org>
22 * scripts/targets/INTERMEDIATE: Test that make doesn't remove
23 .INTERMEDIATE files when given on the command line (PR/1669).
25 2000-03-08 Paul D. Smith <psmith@gnu.org>
27 * scripts/options/dash-k: Add a test for error detection by
28 multiple targets depending on the same prerequisite with -k.
31 2000-02-07 Paul D. Smith <psmith@gnu.org>
33 * scripts/features/escape: Add a test for backslash-escaped spaces
34 in a target name (PR/1586).
36 2000-02-04 Paul D. Smith <psmith@gnu.org>
38 * scripts/features/patspecific_vars: Add a test for pattern-specific
39 target variables inherited from the parent target (PR/1407).
41 2000-02-02 Paul D. Smith <psmith@gnu.org>
43 * run_make_tests.pl (set_more_defaults): Hard-code the LANG to C
44 to make sure sorting order, etc. is predictable.
45 Reported by Andreas Jaeger <aj@suse.de>.
47 * run_make_tests.pl (set_more_defaults): Set the $wtime variable
48 depending on the OS. Eli Zaretskii <eliz@is.elta.co.il> reports
49 this seems to need to be *4* on DOS/Windows, not just 2. Keep it
51 * scripts/features/vpathplus (touchfiles): Use the $wtime value
52 instead of hardcoding 2.
53 * scripts/targets/SECONDARY: Ditto.
54 * scripts/targets/INTERMEDIATE: Ditto.
56 2000-01-27 Paul D. Smith <psmith@gnu.org>
58 * test_driver.pl (toplevel): Don't try to run test scripts which
59 are really directories.
61 2000-01-23 Paul D. Smith <psmith@gnu.org>
63 * scripts/features/include: Remove a check; the fix caused more
64 problems than the error, so I removed it and removed the test for
67 2000-01-11 Paul D. Smith <psmith@gnu.org>
69 * scripts/functions/call: Add a test for PR/1517 and PR/1527: make
70 sure $(call ...) doesn't eval its arguments and that you can
71 invoke foreach from it without looping forever.
73 1999-12-15 Paul D. Smith <psmith@gnu.org>
75 * scripts/targets/INTERMEDIATE: Add a test for PR/1423: make sure
76 .INTERMEDIATE settings on files don't disable them as implicit
77 intermediate possibilities.
79 1999-12-01 Paul D. Smith <psmith@gnu.org>
81 * scripts/features/double_colon: Add a test for PR/1476: Try
82 double-colon rules as non-goal targets and during parallel builds
83 to make sure they're handled serially.
85 1999-11-17 Paul D. Smith <psmith@gnu.org>
87 * scripts/functions/if: Add a test for PR/1429: put some text
88 after an if-statement to make sure it works.
90 * scripts/features/targetvars: Add a test for PR/1380: handling +=
91 in target-specific variable definitions correctly.
93 1999-10-15 Paul D. Smith <psmith@gnu.org>
95 * scripts/variables/MAKEFILES: This was really broken: it didn't
96 test anything at all, really. Rewrote it, plus added a test for
99 1999-10-13 Paul D. Smith <psmith@gnu.org>
101 * scripts/options/dash-n: Add a test for PR/1379: "-n doesn't
102 behave properly when used with recursive targets".
104 1999-10-08 Paul D. Smith <psmith@gnu.org>
106 * scripts/features/targetvars: Add a check for PR/1378:
107 "Target-specific vars don't inherit correctly"
109 1999-09-29 Paul D. Smith <psmith@gnu.org>
111 * test_driver.pl (get_osname): Change $fancy_file_names to
112 $short_filenames and reverse the logic.
113 (run_each_test): Change test of non-existent $port_host to use
114 $short_filenames--problem reported by Eli Zaretskii.
116 1999-09-23 Paul D. Smith <psmith@gnu.org>
118 * scripts/features/parallelism: Add a check to ensure that the
119 jobserver works when we re-invoke. Also cleaned up the tests a
120 little, reducing the number of rules we use so the test won't need
121 as many "sleep" commands.
123 1999-09-16 Paul D. Smith <psmith@gnu.org>
125 * scripts/features/reinvoke: Remove invocations of "touch" in
126 makefiles. See the comments on the touch function rewrite below.
127 Note that UNIX touch behaves the same way if the file already
128 exists: it sets the time to the _local_ time. We don't want
129 this. This is probably a good tip for makefile writers in
130 general, actually... where practical.
131 * scripts/options/dash-l: Ditto.
132 * scripts/options/dash-n: Ditto.
134 * test_driver.pl (run_each_test): In retrospect, I don't like the
135 .lN/.bN/.dN postfix required by DOS. So, for non-DOS systems I
136 changed it back to use .log, .base, and .diff.
138 * run_make_tests.pl (set_more_defaults): Move the check for the
139 make pathname to here from set_defaults (that's too early since it
140 happens before the command line processing).
141 Create a new variable $port_type, calculated from $osname, to
142 specify what kind of system we're running on. We should integrate
143 the VOS stuff here, too.
144 (valid_option): Comment out the workdir/-work stuff so people
145 won't be fooled into thinking it works... someone needs to fix
148 * scripts/functions/origin: Use $port_type instead of $osname.
149 * scripts/functions/foreach: Ditto.
150 * scripts/features/default_names: Ditto.
152 1999-09-15 Paul D. Smith <psmith@gnu.org>
154 * test_driver.pl (touch): Rewrite this function. Previously it
155 used to use utime() to hard-set the time based on the current
156 local clock, or, if the file didn't exist, it merely created it.
157 This mirrors exactly what real UNIX touch does, but it fails badly
158 on networked filesystems where the FS server clock is skewed from
159 the local clock: normally modifying a file causes it to get a mod
160 time based on the _server's_ clock. Hard-setting it based on the
161 _local_ clock causes gratuitous errors and makes the tests
162 unreliable except on local filesystems. The new function will
163 simply modify the file, allowing the filesystem to set the mod
166 * scripts/features/parallelism: The second test output could
167 change depending on how fast some scripts completed; use "sleep"
168 to force the order we want.
170 * test_driver.pl (toplevel): A bug in Perl 5.000 to Perl 5.004
171 means that "%ENV = ();" doesn't do the right thing. This worked
172 in Perl 4 and was fixed in Perl 5.004_01, but use a loop to delete
173 the environment rather than require specific versions.
175 * run_make_tests.pl (set_more_defaults): Don't use Perl 5 s///
176 modifier "s", so the tests will run with Perl 4.
177 (set_more_defaults): Set $pure_log to empty if there's no -logfile
178 option in PURIFYOPTIONS.
179 (setup_for_test): Don't remove any logs unless $pure_log is set.
181 1999-09-15 Eli Zaretskii <eliz@is.elta.co.il>
183 * scripts/features/reinvoke: Put the SHELL definition in the right
186 1999-09-15 Paul D. Smith <psmith@gnu.org>
188 ChangeLog file for the test suite created.