json/include/nlohmann at 24de5b2f5cfa0f0bf832ec07fc8c89632aa654bd - json

amorfo77/json

0 Fork 0

mirror of https://github.com/nlohmann/json.git synced 2026-10-18 03:58:12 +00:00

Files

T

Niels Lohmann 24de5b2f5c Count lines incrementally when a recovering parse reports many errors

With seekable input, the lexer derives line and column on demand (#5809)
by counting the newlines in all input consumed so far. A SAX parser that
recovers from errors asks for the position once per error, so input with
an error at every character took quadratic time: "{" followed by 100000
commas took 85 s at -O0, and the parser tests timed out under Valgrind
(1500 s, was 69 s on develop).

get_position() now remembers what it counted and only scans the input
consumed since the previous call; count_consumed_lines() takes the start
of the range. The 100000 commas now take 0.17 s, unit-class_parser 0.9 s
instead of 86 s. Over 69284 error messages from random multi-line inputs,
the messages are identical to those before the change.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

2026-10-17 20:10:10 +02:00

detail

Count lines incrementally when a recovering parse reports many errors

2026-10-17 20:10:10 +02:00

thirdparty/hedley

Undefine the four JSON_HEDLEY_* macros that leak after including json.hpp (#5475 )

2026-09-09 09:48:17 +02:00

adl_serializer.hpp

📄 adjust year (#5044 )

2026-01-01 20:00:39 +01:00

byte_container_with_subtype.hpp

📄 adjust year (#5044 )