This avoids redundancy between them (and also having to port fixes and features)
The error messages have been preserved through a string reporting mechanism
Make the bool field an actual bool
Rename `iReloc` to a more exact `isKnown` (as was already pointed out by some
comments)
Make the value of `BANK(symbol)` consistent when the argument is invalid
Stop using that bitfield for everything, including what can be determined otherwise
It also makes it easier to have a sane state, since some bits were (supposedly)
mutually exclusive
This macro hid away the arguments to the underlying call, and served
no purpose beyond saving the programmer some typing. This is 2019,
people have IDEs (even Vim!) with autocompletion.
f29d768 forgot to switch these two `expr->nVal` to `src1->nVal`
This won't break anything, since this wasn't published yet.
I checked, and didn't see any other missed changes.
Reported by pret, I confirmed that `1 && 1` evaluated to 0.
The createpatch() function was using a fixed-size buffer. I've changed it
to be dynamically allocated. I saw that the RPN format used in patches is
slightly different from the one used internally in the assembler, so I
added a new member to the Expression struct to track the patch size.
I've also limited the RPN expression length to 1MB. I realized that the
patch RPN expression could potentially be longer than the internal RPN
expression, so the internal expression would need a limit smaller than
UINT32_MAX. I thought 1MB would be a reasonable limit.
It seemed that the consensus in our discussions of signed integer
overflow, which invokes undefined behavior in C, was that integer
arithmetic should be two's complement and there should be no warning for
overflows. I have implemented that by converting values to unsigned types
when appropriate. These changes will mostly preserve existing behavior,
except for a few cases that were being handled incorrectly before.
The case of dividing INT_MIN by -1 previously resulted in a CPU
exception and program termination. Now, that case is detected and results
in a warning and a value of INT_MIN.
Similarly, INT_MIN % -1 would have resulted in a CPU exception. Since this
is a mathematically valid operation with a result of 0, it now simply
gives that result without a warning.
I noticed that in rpn.c, there were attempts in certain operation handlers
to validate the nVal members of the source expressions even when the
expressions may have been relocatable expressions with meaningless numbers
for the nVal member. This could have caused spurious errors/warnings, so I
made those handlers confirm that isReloc is false before validating nVal.
Also, integer constants that are too large now result in a warning. The
post-conversion values have not been changed, in order to preserve
backward compatibility.
The tests are not exhaustive, there are some conditions that aren't
checked. The tests are based in the C standard rules about undefined
behaviour.
This is a compatibility break but, hopefully, all projects are using
sane values. If not, there is no guarantee that the projects will build
in any platform where RGBDS can be compiled, so it would be better to
fix them.
Even though, technically, the left shift of a negative value is always
undefined, some projects rely on its current behaviour. This is the
reason why this doesn't cause a fatal error.
Signed-off-by: Antonio Niño Díaz <antonio_nd@outlook.com>
With permission from the main authors [1], most of the code has been
relicensed under the MIT license.
SPDX license identifiers are used so that the license headers in source
code files aren't too large.
Add CONTRIBUTORS.rst file.
[1] https://github.com/rednex/rgbds/issues/128
Signed-off-by: Antonio Niño Díaz <antonio_nd@outlook.com>
The bank of a section can be requested with `BANK("Section Name")`, and
the bank of the current section with `BANK(@)`. In both cases, the bank
number is resolved by the linker.
New commands have been added to the list of RPN commands of object
files, and the rest has been moved so that new additions don't force a
new change in the number of the enumerations.
Increase object file version, as it is now incompatible with the old
format.
Update manpages to reflect the new ways of using `BANK()` and the new
format of the object files.
Signed-off-by: Antonio Niño Díaz <antonio_nd@outlook.com>
They work with the 16-bit registers BC, DE and HL, returning the
corresponding 8-bit register. HIGH() works with AF as well, returning A.
They also work with any kind of constant or symbol, generating a RPN
patch in the object file if the value is not defined at assembly time.
They work with macro arguments as well.
Signed-off-by: Antonio Niño Díaz <antonio_nd@outlook.com>
Symbols are created when using a label in the wild, even if they aren't defined. Solely using a symbol as an argument to BANK() skips this, so the symbol is never created.
This evaluates the argument instead of trying to find a symbol. This way, symbols that don't exist are created when passed into BANK().
Merging lai's source with this one is very irritating because
they have different indentation styles. I couldn't find what profile
vegard used for his version, so I used these flags (which should bring
the source close to KNF):
-bap
-br
-ce
-ci4
-cli0
-d0
-di0
-i8
-ip
-l79
-nbc
-ncdb
-ndj
-ei
-nfc1
-nlp
-npcs
-psl
-sc
-sob
