Issue 10341 - ICE on case-range statement without an associated switch statement

[ghost]

http://d.puremagic.com/issues/show_bug.cgi?id=10341

[ghost]

I can't seem to return this here, it just segfaults at a different location. For expressions I can usually return new ErrorExp(), but we don't seem to have an equivalent for statements?

[WalterBright]

I plan to eventually create an ErrorStatement, but haven't done so yet.

[yebblies]

You could possibly use an empty CompoundStatement as a no-op statement.

[ghost]

Like so?:

if (sw == NULL)
{
    error("case range not in switch statement");
    return new CompoundStatement(loc, new Statements());
}

[yebblies]

I think so. Alternatively, some statements return NULL if they are removed by semantic, like static ifs/asserts that are not triggered. This should work at least in theory.

[ghost]

Returning NULL seems to work too. I'll give it a go and see what the tester says.

[9rnsr]

LGTM.

[9rnsr]

Mmm, Currently, the following error has occurred frequently on Linux platform in Auto-tester.

Creating output directory: test_results
Building d_do_test tool
OS: posix
Error: cannot find source code for runtime library file 'object.d'
       dmd might not be correctly installed. Run 'dmd -man' for installation instructions.
Specify path to file 'object.d' with -I switch
make: *** [test_results/d_do_test] Error 1
Building d_do_test tool
OS: posix
Error: cannot find source code for runtime library file 'object.d'
       dmd might not be correctly installed. Run 'dmd -man' for installation instructions.
Specify path to file 'object.d' with -I switch
make: *** [test_results/d_do_test] Error 1

I'm really not sure what's wrong, but from the circumstance merging this pull request is the most doubtful.
@AndrejMitrovic, sorry but I'll revert this change from master branch.

[9rnsr]

The snapshot of failures which had been displayed by auto tester:

[ghost]

@9rnsr I see no relationship between case range statement and dmd's inability to find object.d. You have reverted the pull, but failures still occur.

[9rnsr]

@mxfm Right...

[ghost]

@9rnsr: Could this be the problem?: cf03d8e

I didn't think about MSYS when I implemented that pull.

[braddr]

Sorry abut those tester errors.. entirely my fault. Ignore them. I've deprecated all the tests from that build host so those tests are being re-run.

[ghost]

@9rnsr: Should I reopen this pull?

[9rnsr]

@braddr Thanks, I confirmed.
@AndrejMitrovic Unfortunately, we cannot re-open the pull request which has once merged. I pushed the merge commit to upstream once again (095b58a).