From d4c145179420e6de0d2189059ea988072a7e7e45 Mon Sep 17 00:00:00 2001 From: Kim Alvefur Date: Fri, 10 Dec 2021 22:48:45 +0100 Subject: util.format: Escape invalid UTF-8 by passing trough serialization Should prevent invalid UTF-8 from making it into the logs, which can cause trouble with terminals or log viewers or other tools, such as when grep determines that log files are binary. --- util/format.lua | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'util') diff --git a/util/format.lua b/util/format.lua index 43097e6a..d452fd3d 100644 --- a/util/format.lua +++ b/util/format.lua @@ -5,6 +5,7 @@ local tostring = tostring; local unpack = table.unpack or unpack; -- luacheck: ignore 113/unpack local pack = require "util.table".pack; -- TODO table.pack in 5.2+ +local valid_utf8 = require "util.encodings".utf8.valid; local type = type; local dump = require "util.serialization".new("debug"); local num_type = math.type or function (n) @@ -60,10 +61,18 @@ local function format(formatstring, ...) args[i] = dump(arg); spec = "%s"; elseif option == "s" then - args[i] = tostring(arg):gsub("[%z\1-\8\11-\31\127]", control_symbols):gsub("\n\t?", "\n\t"); + arg = tostring(arg); + if arg:find("[\128-\255]") and not valid_utf8(arg) then + args[i] = dump(arg); + else + args[i] = arg:gsub("[%z\1-\8\11-\31\127]", control_symbols):gsub("\n\t?", "\n\t"); + end elseif type(arg) ~= "number" then -- arg isn't number as expected? args[i] = tostring(arg); spec = "[%s]"; + option = "s"; + spec = "[%s]"; + t = "string"; elseif expects_integer[option] and num_type(arg) ~= "integer" then args[i] = tostring(arg); spec = "[%s]"; -- cgit v1.2.3