| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
Fixed deepToString to put spaces after commas
(like regular Array toString, and List, etc.).
Fixed all the test cases depending on this.
Fixed the interpreter to keep my mind from wandering.
|
|
|
|
|
| |
Added test for array pretty-printing.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Problem was calling incompleteInputError multiple times would escalate
the problem to be treated as a "real" error. This does not appear to be
the correct semantics. Tried to fix this up as best as possible, but
Sean needs to investigate the implications for the Eclipse plugin.
|
| |
|
|
|
|
|
|
|
|
|
| |
Generalizes the allowed inputs to the interpreter to be any combination
of phrases that can be placed inside a template. Instead of having one
"request" object try to understand the entire combination of phrases,
the code now uses multiple "member handlers" for each request, one
member handler for each phrase.
|
|
|
|
|
| |
Tweaked whitespace in the test.
|
| |
|
| |
|
|
|
|
|
| |
there are type errors in the user's code
|
|
|
|
|
|
|
|
| |
Fixed interpreter error recovery mode and something in Namers that was
unlinking module symbols (removed unlink, hope resident compiler is ok,
but there are no tests for this in the test suite, better to fix the
tests we do have then worry about tests we don't right?)
|
| |
|
| |
|
|
|
|
|
| |
variable names generated by the interpreter.
|
|
|
|
|
| |
Skip the first line of output, because now it includes a version number
|
|
|
|
|
|
| |
The parser no longer signals "incomplete input" in cases where the input
already has an outright syntax error.
|
| |
|
| |
|
|
|
|
|
|
| |
added a new line at the end, so that hopefully the test is more robust
on different machines
|
| |
|
|
|