Better extraction of lines from JSON (handle newline literals, hard-wrapping)

Currently, the frontend splits the source-code into lines wherever the substring \n appears in the JSON. This means newline literals between single- or double-quotes are parsed as indicating a new line of code, so that a simple expression such as:

std::cout << "\n\n\n";

is treated as multiple lines, with only the ; "line" being coloured as if it was executed. In addition to being a display issue, this results in any subsequent lines of code to have incorrect execution counts displayed, not, since there end up being more lines of source code than execution/debugger steps. See e.g. the third example output produced from running cargo run (json/simple.json).

Likewise, (although I will need to actually test this some actual output from the backend), hard-wrapped lines will (almost certainly) cause similar mismatching and display problems.

The first problem should be solvable (entirely on the front-end) with a simple reg-ex.

The second will probably require a little more thought, but using some simple heuristics that rely on common code styles [e.g. checking for ;\n, }\n, or \n. (i.e. multi-line chained method calls)] would probably still go a long way for the vast majority of cases, without requiring any actual / proper parsing of the source code.

Comments (1)