| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
Clean up primitives for converting from/to hex chars.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
We want slice-equal? for length-prefixed strings, not null-terminated "kernel"
strings.
|
| | |
|
| |
|
|
|
| |
In the process of building slice primitives I found an out-of-bounds access
in write-byte.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
Fix CI.
a) Update canonical binaries.
b) Fix an out-of-bounds access in `clear-stream`. This also required supporting
a new instruction in `subx run` to load an imm8 into rm8.
|
| |
|
|
| |
We only can't use rm32=5 when mod=0. Totally fine when it's mod=1.
|
| | |
|
| | |
|
| |
|
|
| |
More mnemonic register usage in write-stream.
|
| |
|
|
| |
Our first buffer overflow!
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Clean up a few things:
a) Call scan-next-byte in hex.subx with the right number of args. Turns
out tests continue to work fine if they never use the other args.
b) Tear down a test for 'stop' in the right order. Not important since we
have no EBP to restore. But can still be misleading.
c) Have 'check-ints-equal' return nothing. Handy for it to not mess up
EAX. I never use the result anyway, and the name also is imperative suggesting
callers won't expect a return value.
|
| | |
|
| |
|
|
| |
New helper: print an error message, then a numeric byte, then abort.
|
| |
|
|
| |
New helper: printing a string to a buffered file.
|
| |
|
|
| |
Better to use EDI as a mnemonic for 'destination'.
|
| |
|
|
| |
Let's standardize to use opcode 39 rather than 3b by default.
|
| |
|
|
| |
Fix CI since 4827.
|
| | |
|
| |
|
|
|
| |
This is likely a sub-optimal interface, but I'm trying not to agonize.
The whole point of Mu is to permit radical changes at any point in time.
|
| |
|
|
|
|
|
|
|
| |
I was 'returning' a phantom value from 'write' when the underlying '_write'
returns nothing.
In general, returning counts of bytes written is not so useful for error
checking when my primitives abstract away from that. We'll come back to
error signalling later.
|
| |
|
|
|
|
|
| |
Fix CI.
It's kind of a hassle (and wasteful) that I need to redefine 'main' in
every single layer.
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
Start with an exactly corresponding version to Crenshaw 2-1: single-digit
numbers. The only change: we assume the number is in hex.
The next version now supports multi-digit hex numbers.
|
| |
|
|
| |
Simplification.
|
| |
|
|
| |
In the process I had to fix a couple more bugs in support for disp16 instructions.
|
| | |
|
| | |
|
| |
|
|
|
| |
This time I've ported (and test-driven) 'GetChar' and 'GetNum'. The new
tests bring together our new testable interfaces for read() and exit().
|
| |
|
|
|
| |
Long-standing and long-copied typo has been messing with our exit status
on test failures.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
Extract a helper that we'll need for 'read'.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
We'll use a common stream data structure for input and output streams.
Having separate types makes more sense in a more high-level language, where
we have type checking and where functions for handling the different types
are more concise. But in machine code the sweet spot is more toward fewer
types.
|
| | |
|
| |
|
|
|
|
| |
Start injecting all dependencies in the Crenshaw compiler app.
In the process I realized the non-fake code path of 'stop' had a bug.
|
| | |
|