bytes/string literal concatenation

[axic]

In some languages subsequent string literals are concatenated prior to compilation, which proves useful for containing large amounts of verbatim text in source code, while splitting it across multiple lines or logical boundaries.

One good example for Solidity is NFT metadata-JSON and SVG generation, where text is built up programatically. Current best practice is use string.concat and break logical parts.

Here's an excerpt from Uniswap NFT:

                '<path id="text-path-a" d="M40 12 H250 A28 28 0 0 1 278 40 V460 A28 28 0 0 1 250 488 H40 A28 28 0 0 1 12 460 V40 A28 28 0 0 1 40 12 z" />',
                '<path id="minimap" d="M234 444C234 457.949 242.21 463 253 463" />',
                '<filter id="top-region-blur"><feGaussianBlur in="SourceGraphic" stdDeviation="24" /></filter>',
                '<linearGradient id="grad-up" x1="1" x2="0" y1="1" y2="0"><stop offset="0.0" stop-color="white" stop-opacity="1" />',
                '<stop offset=".9" stop-color="white" stop-opacity="0" /></linearGradient>',
                '<linearGradient id="grad-down" x1="0" x2="1" y1="0" y2="1"><stop offset="0.0" stop-color="white" stop-opacity="1" /><stop offset="0.9" stop-color="white" stop-opacity="0" /></linearGradient>',
                '<mask id="fade-up" maskContentUnits="objectBoundingBox"><rect width="1" height="1" fill="url(#grad-up)" /></mask>',
                '<mask id="fade-down" maskContentUnits="objectBoundingBox"><rect width="1" height="1" fill="url(#grad-down)" /></mask>',
                '<mask id="none" maskContentUnits="objectBoundingBox"><rect width="1" height="1" fill="white" /></mask>',
                '<linearGradient id="grad-symbol"><stop offset="0.7" stop-color="white" stop-opacity="1" /><stop offset=".95" stop-color="white" stop-opacity="0" /></linearGradient>',
                '<mask id="fade-symbol" maskContentUnits="userSpaceOnUse"><rect width="290px" height="200px" fill="url(#grad-symbol)" /></mask></defs>',
                '<g clip-path="url(#corners)">',

This results in a bunch of items on the stack and a lot of more runtime overhead.

It would be useful allowing 'hello' 'world' to be compounded as helloworld.

[hrkrshnn]

Random comment: we should also be moving these string into code and use codecopy, rather than mstore(...) with 32-byte chunks.

[github-actions]

This issue has been marked as stale due to inactivity for the last 90 days.
It will be automatically closed in 7 days.

[github-actions]

Hi everyone! This issue has been automatically closed due to inactivity.
If you think this issue is still relevant in the latest Solidity version and you have something to contribute, feel free to reopen.
However, unless the issue is a concrete proposal that can be implemented, we recommend starting a language discussion on the forum instead.

[cameel]

Isn't it already implemented? #7524 has been merged ages ago.

Parser already has this covered unless I'm misunderstanding the feature being requested here:

solidity/libsolidity/parsing/Parser.cpp

Lines 2029 to 2039 in e29a68d

	case Token::StringLiteral:
	case Token::UnicodeStringLiteral:
	case Token::HexStringLiteral:
	{
	string literal = m_scanner->currentLiteral();
	Token firstToken = m_scanner->currentToken();
	while (m_scanner->peekNextToken() == firstToken)
	{
	advance();
	literal += m_scanner->currentLiteral();
	}