torque-asm/src/stages, branch v2.4.0

Allow label references in pinned address calculations

2025-04-26T02:45:05Z

This is a relaxation of the rule where a label reference could not be used in any context that could change the length of an assembled program. We implement this in the bytecode stage by naively calculating an initial address for each label as before. If a pinned address is calculated from a label reference, some of the calculated addresses could be incorrect. We then attempt to run the bytecode stage, which will calculate a more accurate address for each label based on how pinned addresses are calculated. If the address of any label was changed by running this stage, we re-run the stage up to three more times until all labels stabilise. If the labels fail to stabilise, we return an error.

Implement first-class string literals

2025-04-26T01:00:51Z

This feature promotes strings to a first-class type in the language. If a string is passed to an invocation via the new string-type argument, the string will be passed as a whole value. String arguments can still be passed to an invocation via an integer-type argument, in which case they'll be broken apart into individual characters with the macro being invoked once per character. String-type macro arguments are declared like "name".

Fix misleading error message

2025-04-25T21:31:58Z

The code here is really expecting a block value, but has received an integer value. This catches the case where an integer value is incorrectly included in a macro definition body alongside block tokens.

Allow a macro to invoke itself safely

2025-04-18T08:47:59Z

A macro can now invoke itself if the invocation is inside a conditional block that will eventually return false. The assembler stack can still overflow if the macro recurses too deeply, or if a macro calls itself without a conditional block.

Allow a macro to invoke itself

2025-04-18T08:45:19Z

This will currently cause the assembler to hang in all situations where it is used.

Implement operator for expressions

2025-04-18T08:34:39Z

The operator returns the width of an integer in bits, using the same calculation as for packing an integer into a bit field.

Implement octal literals

2025-04-17T04:37:45Z

Octal literals use the prefix '0o'.

Fix width checks for negative integers

2025-04-11T22:19:26Z

The width of a negative integer was previously being counted in the same way as for a positive integer, by inverting the bits to make it a positive integer and then finding the placement of the highest-order 1 bit. The actual width of a negative integer will always be one greater than this value however, because the highest-order 1 bit of an inverted negative integer will always have directly above it a significant 0 bit used as the sign bit.

Implement negative integer literals

2025-04-11T22:08:09Z

Negative literals take the forms -29, -0x1D, and -0b11101 for decimal, hexadecimal, and binary.

Report token without prefix in invalid literal errors

2025-03-22T22:37:29Z

Previously, the token quoted in the error message for an invalid literal error included the radix prefix. This is already visible in the highlighted source report, and implied by the named radix in the error message.