bpo-37587: Make json.loads faster for long strings (GH-14752)

mpaolini · ncoghlan · commit 8a758f5b99c5 · 2019-07-31T00:16:34.000+10:00
When scanning the string, most characters are valid, so
checking for invalid characters first means never needing
to check the value of strict on valid strings, and only
needing to check it on invalid characters when doing
non-strict parsing of invalid strings.

This provides a measurable reduction in per-character
processing time (~11% in the pre-merge patch testing).
diff --git a/Misc/NEWS.d/next/Library/2019-07-13-16-02-48.bpo-37587.fd-1aF.rst b/Misc/NEWS.d/next/Library/2019-07-13-16-02-48.bpo-37587.fd-1aF.rst
@@ -0,0 +1 @@
+Make json.loads faster for long strings. (Patch by Marco Paolini)
diff --git a/Modules/_json.c b/Modules/_json.c
@@ -439,7 +439,7 @@ scanstring_unicode(PyObject *pystr, Py_ssize_t end, int strict, Py_ssize_t *next
             if (c == '"' || c == '\\') {
                 break;
             }
-            else if (strict && c <= 0x1f) {
+            else if (c <= 0x1f && strict) {
                 raise_errmsg("Invalid control character at", pystr, next);
                 goto bail;
             }

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1 @@`
	`1`	`+Make json.loads faster for long strings. (Patch by Marco Paolini)`
Original file line number	Diff line number	Diff line change
`@@ -439,7 +439,7 @@ scanstring_unicode(PyObject pystr, Py_ssize_t end, int strict, Py_ssize_t next`
`439`	`439`	`if (c == '"' \|\| c == '\\') {`
`440`	`440`	`break;`
`441`	`441`	`}`
`442`		`- else if (strict && c <= 0x1f) {`
	`442`	`+ else if (c <= 0x1f && strict) {`
`443`	`443`	`raise_errmsg("Invalid control character at", pystr, next);`
`444`	`444`	`goto bail;`
`445`	`445`	`}`