| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
Make `compute-addresses` less clever. Stop striding from the middle of
one row to the next. This way we'll also obviate the need for indexing
backwards from a pointer in the next commit.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Kinda hacky, but might scale enough for machine code.
This was really hard to debug. Single tests passed, but when I ran all
tests I got breakage because tests long before (from the 056trace layer)
were not cleaning up properly.
My instinct was to call clear-stream on Trace-stream, which was wrong
(the trace didn't have the wrong contents, it was literally a bad
object). It was also wrong in a counter-productive way: calling
clear-stream on a real Trace stream (which is the size of a page of
memory) takes a long time in emulated mode.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When rerunning most recently run test, also load the same files as
before. Try the following commands in sequence:
$ ./run_one_test.sh 056trace.subx test-compute-addresses # 1
0 error: missing reference to global 'test-compute-addresses'
$ ./run_one_test.sh apps/survey.subx test-compute-addresses # 2
$ ./run_one_test.sh 056trace.subx trace # run previous test # 3
Now command 3 will work, and it will behave identically to command 2.
This way I can switch to another file in Vim and hit `\t` in a random
place in the file and still have the previously run test be rerun.
|
| |
|
| |
|
|
|
|
|
| |
Simplify run_one_test.sh. We already create /tmp/run_one_test.subx, we
don't need to track more state across runs.
|
|
|
|
| |
Fix some comments.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
No trace statements yet, so we don't know if it works.
|
| |
|
|
|
|
| |
Clean up CI.
|
|
|
|
|
| |
Flesh out final test some more. We also now have a new family of
primitives for writing non-strings to input streams in tests.
|
|
|
|
| |
When two assertions don't share context they're clearer when separated.
|
|
|
|
|
|
|
| |
Move tests adjacent to function being tested.
Originally I didn't expect the helpers to get their own tests. But even
so, arguably the original layout was less clear.
|
| |
|
|
|
|
|
| |
Make a test a little less confusing; stream was initialized with data
filled in but outside bounds.
|
|
|
|
|
| |
1. skip-next-line should skip newline
2. trace-scan was falling into an infinite loop on non-matching lines
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
This missing line was giving the illusion that all was well when running
a single test at a time.
Now trace-scan still hangs, but it hangs reliably. And it's clear what
the logic error is.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Fix a bug when running tests singly in the 'SubX standard library'.
|
|
|
|
|
| |
Now that we don't have to edit code to run a single test, delete that
commented out fragment everywhere.
|
| |
|
| |
|
|
|
|
|
| |
I checked if we needed to bring back anything since commit 3976, but the
only difference is dropping the :(scenario) DSL.
|
| |
|
|
|
|
| |
Runs test under cursor and opens its trace.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|