GitLab

Add a Makefile here.

Add cpranal tests.

Remove these tests: they used to test the compiler's internal printing
machinery, but too much has changed in that area for the tests to
stand any chance of working.  At some point we should add some new
tests for -ddump-foo.

expect tcrun007 to fail, because it requires generics support in the prelude.

expected failure fixes

tcrun015 doesn't exist (???)

don't forget to include std-macros.T

the stubs shouldn't be in the repository

Add Makefiles for these tests

make this an expected failure

update this and actually use it from vanilla-test.T and multimod-test.T

This test is an expected failure.  We have no idea how to run it.

Allow expected failures.

Allow expected failures, by setting the global var $expect to "fail"
(defaults to "pass").

Add a README file, which explains how to run the testsuite, how to
update the expected output, and how to add new tests.  It assimilates
the contents of tests/ghc-regress/HOW_TO_ADD_A_TEST, which has been
removed.

Use makefiles in the subdirectories to ease running the test driver in
a subdirectory.  It can be run in the following ways:

 	make		-- run all the tests in the current directory
 	make verbose	-- as make test, but up the verbosity
 	make accept	-- run the tests, accepting the current output

and you can also set some variables on the command line, eg.

	make accept TESTS=arr001  -- accept an individual test
	make EXTRA_HC_OPTS='-v'   -- extra flags to the haskell compiler
	make CONFIG=<config>	  -- use a different test config

this will all go in the docs at some point.

Add a simple fptools project, so we don't have to wire in things like
the location of GHC_INPLACE, and so we can use Makefiles in the
various test subdirectories to simplify running the test driver.

add strictness test

Add tc133

update these tests

ignore newlines too: because the compiler pretty-prints error
messages, we sometimes encounter spurious line wrappage.

fix obviously wrong typo

We had file better turn on -fglasgow-exts for this test: it uses local
unversal quantification.

These tests now work

print contents of the file that should have been empty, in verbose mode

Add tcrun016

Knock this into slightly better shape and bring over some more tests
from the old test suite.

Print a helpful message when the program produced some output and it
shouldn't have.

Allow $stdin to be overriden.

Use say_fail_if_noteq() rather than inline version.

Track some minor changes to the syntax.

- Replace the parser with a happy one
- simplify the grammar in a couple of places
- don't automatically print out the argument to 'run'.

Allow compiling to .hc only.

merge from old test suite

Don't try to handle multiple sample stdout/stderr files for now.

Fix these tests.

Bring these tests up to date.

--tool is no longer a formal option: you can get the same effect by
just saying tool=<tool> on the command line.  This means it can also
be set in a config file instead.

- print diffs when output files differ (in verbose mode)

- allow new output to be "accepted" by saying accept=True
  on the cmd line

- abstract some command names (to the top of the file for now)

Add a verbose flag (use ./runtests verbose=True to get the actual
command lines printed out)

- Fix a bug in the 'defined' operator
- Print fewer newlines