* Implement proper number literal parsing
- Scientific notation support
- Correct error handling in cases of invalid digits present in binary or
hex literals (0b123, 0x013FK).
* Fix style linter issues
* Support +/- in number literal exponent, add tests
- Add tests for scientific notation number literal parsing
- Add support for +/- in scientific notation exponents
This commit updates the Makefile to complile each .c source file into
an object before linking. The benefit of this is that only the necessary
files will be recompiled when modifying the code.
We also move the Makefile and build.bat up to the root directory of the
repo. This is so that calling make directly builds the binaries.
Both files are updated to adapt to the new directory structure.
With this change, initial build times are increased from ~3s to ~10s on
my machine. However, subsequent build times are improved.
As a result of this change, SConstruct is somewhat deprecated. We can
update this as well if necessary in a future change.
I verified that the build completes successfully and tests pass for
both ubuntu and windows 10.
A future change could compile src/ files into a static library to be
linked with cli/, if necessary.
Issue: #64
Introduced a new function `readLine()` that utilizes the byte buffer
together with `fgetc()` to read input from the STDIN. This enables
the REPL to read characters one at a time for validation & in a
safer manner where the buffer gets to be resized if it's full.
Currently, it is required for the shell's working directory to be in
the repo's test/ directory in order for the tests to run correctly.
This also applies for the static analysis checker.
The reason for this is that paths in the test orchestration code are
defined relative to the test directory. Instead, and since none of
these scripts are modules that will in the forseeable future be
imported into other files, we update all paths to be relative
to the absolute path of the script's file itself.
┏( ͡❛ ͜ʖ ͡❛)┛
This commit implements list addition. A new list is allocated, and
the contents of l1 and l2 are concatenated onto the list.
This is done by extending the buffer template macro to support a
`concat` operation, which given an `other` buffer, appends all
elements of the `other` buffer to the end of the `self` buffer.
Allocations are done as needed.
We implement this additional operation, since the alternative would
be to call `reserve()` for each `write()` operation, which doesn't
seem ideal.
Issue: #55
