Implement improved string literals with Swift-style syntax for interpolation, hex, & unicode characters, plus C & Python style formatting and Unicode, HTML, LaTeX, and Emoji entities



Build Status

Coverage Status


The StringLiterals package is an attempt to bring a cleaner string literal syntax to Julia, as well as having an easier way of producing formatted strings, borrowing from both Python and C formatted printing syntax. It also adds support for using LaTex, Emoji, HTML, or Unicode entity names that are looked up at compile-time.

Currently, it adds a Swift style string macro, f"...", which uses the Swift syntax for interpolation, i.e. \(expression). This means that you never have to worry about strings with the $ character in them, which is rather frequent in some applications. Also, Unicode sequences are represented as in Swift, i.e. as \u{hexdigits}, where there can be from 1 to 6 hex digits. This syntax eliminates having to worry about always outputting 4 or 8 hex digits, to prevent problems with 0-9,A-F,a-f characters immediately following. Finally, I have added four ways of representing characters in the literal string, \:emojiname:, julia-observer-quote-cut-paste-4__worklt;latexname>, julia-observer-quote-cut-paste-5__workamp;htmlname; and \N{UnicodeName}. This makes life a lot easier when you want to keep the text of a program in ASCII, and also to be able to write programs using those characters that might not even display correctly in their editor.

It also adds a string macro that instead of building a string, can print the strings and interpolated values directly, without having to create a string out of all the parts. Finally, there are uppercase versions of the macros, which also supports the legacy sequences, $ for string interpolation, \x followed by 1 or 2 hex digits, \u followed by 1 to 4 hex digits, and \U followed by 1 to 8 hex digits.

This uses a fork of the https://github.com/JuliaIO/Formatting.jl package to provide formatting capability, as well as Tom Breloff's PR https://github.com/JuliaIO/Formatting.jl/pull/10, which provides the capability of using settable printing defaults based on the types of the argument. The formatting code has been extensively modified, see https://github.com/JuliaString/Format.jl.

  • \ can be followed by: 0, $, ", ', \, a, b, e, f, n, r, t, u, v, N, %, (, <, {, : or &. In the legacy modes, x and U are also allowed after the \. Unlike standard Julia string literals, unsupported characters give an error (as in Swift).

  • julia-observer-quote-cut-paste-12__work outputs a nul byte (0x00) (note: as in Swift, octal sequences are not supported, just the nul byte)

  • \a outputs the "alarm" or "bell" control code (0x07)

  • \b outputs the "backspace" control code (0x08)

  • \e outputs the "escape" control code (0x1b)

  • \f outputs the "formfeed" control code (0x0c)

  • \n outputs the "newline" or "linefeed" control code (0x0a)

  • \r outputs the "return" (carriage return) control code (0x0d)

  • \t outputs the "tab" control code (0x09)

  • \v outputs the "vertical tab" control code (0x0b)

  • \u{<hexdigits>} is used to represent a Unicode character, with 1-6 hex digits.

  • julia-observer-quote-cut-paste-22__worklt; followed by a LaTeX entity name followed by > outputs that character or sequence if the name is valid.

  • \: followed by an Emoji name followed by : outputs that character or sequence (if a valid name)

  • julia-observer-quote-cut-paste-26__workamp; followed by an HTML entity name followed by ; outputs that character or sequence (if a valid name)

  • \N{ followed by a Unicode entity name (case-insensitive!) followed by a } outputs that Unicode character (if a valid name)

  • \(expression) simply interpolates the value of the expression, the same as $(expression) in standard Julia string literals.

  • \%<ccc><formatcode>(arguments) is interpolated as a call to cfmt("<cccc><formatcode>",arguments), where <ccc><formatcode> is a C-style format string.

  • \%(arguments) is interpolated as a call to fmt(arguments). This is especially useful when defaults have been set for the type of the first argument.

  • fmt_default!{T}(::Type{T}, syms::Symbol...; kwargs...) sets the defaults for a particular type.

  • fmt_default!(syms::Symbol...; kwargs...) sets the defaults for all types.

Symbols that can currently be used are: :ljust or :left, :rjust or :right, :commas, :zpad or :zeropad, and :ipre or :prefix.

  • reset!{T}(::Type{T}) resets the defaults for a particular type.
  • defaultSpec(x) will return the defaults for the type of x, and
  • defaultSpec{T}(::Type{T}) will return the defaults for the given type.

There is currently support for Python style formatting, although that is a work-in-progress, and I am intending to improve the syntax to make it as close as possible to Python's 3.6 format strings. Currently, the syntax is \{<formatstring>}(expression), however I plan on changing it shortly to \{expression} (equivalent to pyfmt("", expression), and \{expression;formatstring} (equivalent to pyfmt("formatstring", expression).

First Commit


Last Touched

4 months ago


38 commits

Used By: