[GHC] #11737: Make validate --testsuite-only behave correctly w.r.t. bindisttest
#11737: Make validate --testsuite-only behave correctly w.r.t. bindisttest -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: feature | Status: new request | Priority: normal | Milestone: Component: Build System | Version: 8.1 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- I'm in the process of validating. One of the error messages is a bit wrong, so I change the error message code. But the change is more than just the phrasing of the message; it's conceivable that a mistake in this edit will cause misbehavior, but only on rejected programs. So I determine that I don't need to re-run the full validation, but just recompile and run the testsuite. I recompile. I say `validate --testsuite-only`. And then, after I read the results, I realize that validation uses the bindist version of the GHC, which isn't the one that I just built. So have to do it all again. Thus, request: have `--testsuite-only` not use the bindist version. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11737 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#11737: Make validate --testsuite-only behave correctly w.r.t. bindisttest -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Build System | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by thomie): I think it would be confusing if `./validate --testsuite-only` didn't run the tests in the same way as `./validate`. Maybe one of the following solutions can serve you already? Instead of `./validate --testsuite-only`, run: * `make test THREADS=<n>`. It doesn't use the binary distribution. or * `./validate --no-clean`. It will call `make`, which should be a no-op, then recreate the binary distribution, and then call `make test BINDIST=YES` for you. or * `./validate --testuite-only --fast`. Drawback is that it doesn't run all tests. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11737#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#11737: Make validate --testsuite-only behave correctly w.r.t. bindisttest -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Build System | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by goldfire): Replying to [comment:1 thomie]:
I think it would be confusing if `./validate --testsuite-only` didn't run the tests in the same way as `./validate`.
Point well-taken.
Maybe one of the following solutions can serve you already? Instead of
`./validate --testsuite-only`, run:
* `make test THREADS=<n>`. It doesn't use the binary distribution.
I thought of this but wasn't sure it would run the same set of tests in the same way.
or * `./validate --no-clean`. It will call `make`, which should be a no- op, then recreate the binary distribution, and then call `make test BINDIST=YES` for you.
This wouldn't have been a no-op, because I edited the GHC source code. So it would have rebuilt everything from stage1 on. Perhaps the answer I'm looking for is `validate --no-clean --lax-dependencies` or `validate --no- clean --stage-2`. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11737#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#11737: Make validate --testsuite-only behave correctly w.r.t. bindisttest -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Build System | Version: 8.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by thomie):
wasn't sure it would run the same set of tests in the same way It should. If you want to know more, it's probably best to look at validate script itself. The line that calls the tests is: {{{ $make $MAKE_TEST_TARGET stage=2 $BINDIST $TEST_VERBOSITY THREADS=$threads 2>&1 | tee testlog }}}
`validate --no-clean --lax-dependencies` or `validate --no-clean --stage-2`. You can create a file `mk/validate.mk` containing `stage=2` and `LAX_DEPENDENCIES = YES`.
When running `./validate` (or when running `make` with a leftover `mk/are- validating.mk` file), the build system reads settings from `mk/validate.mk` instead of `mk/build.mk`. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11737#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
#11737: Make validate --testsuite-only behave correctly w.r.t. bindisttest -------------------------------------+------------------------------------- Reporter: goldfire | Owner: Type: feature request | Status: closed Priority: normal | Milestone: Component: Build System | Version: 8.1 Resolution: invalid | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by goldfire): * status: new => closed * resolution: => invalid Comment: Ah right. OK. I guess this is a case of a misguided/uninformed user. Thanks for fixing the bug. :) -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11737#comment:4 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
participants (1)
-
GHC