From 579c31e4254b7d8d211bbd03c159c8cd41804f43 Mon Sep 17 00:00:00 2001 From: Yutaro Sakamoto Date: Tue, 10 Dec 2024 05:42:40 +0000 Subject: [PATCH 1/2] fix: convert SJIS string literals to ordinary Java string literals --- cobj/codegen.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/cobj/codegen.c b/cobj/codegen.c index 7ea8f3f4..95a4d7df 100644 --- a/cobj/codegen.c +++ b/cobj/codegen.c @@ -505,7 +505,8 @@ static void joutput_string_write(const unsigned char *s, int size, enum cb_string_category category) { int i; - if (category == CB_STRING_CATEGORY_ALL_ASCII) { + if (category == CB_STRING_CATEGORY_ALL_ASCII || + category == CB_STRING_CATEGORY_ALL_SJIS) { if (param_wrap_string_flag) { joutput("new CobolDataStorage("); } else { From bddb3321a9fde2cff40bc95a33ba7365946187e1 Mon Sep 17 00:00:00 2001 From: Yutaro Sakamoto Date: Thu, 12 Dec 2024 07:28:57 +0000 Subject: [PATCH 2/2] test: add test suites --- tests/i18n_sjis.src/pic-x.at | 59 ++++++++++++++++++++++++++++++++++++ 1 file changed, 59 insertions(+) diff --git a/tests/i18n_sjis.src/pic-x.at b/tests/i18n_sjis.src/pic-x.at index 3311d792..c9c92812 100644 --- a/tests/i18n_sjis.src/pic-x.at +++ b/tests/i18n_sjis.src/pic-x.at @@ -483,3 +483,62 @@ AT_CHECK([cobj prog.cob], [0]) AT_CHECK([java prog], [0], [02]) AT_CLEANUP + +AT_SETUP([Readable string literals]) +# Older compilers converts string literals "日本語" in COBOL source code +# to `CobolUtil.toBytes((byte)0x93, (byte)0xfa, (byte)0x96, (byte)0x7b, (byte)0x8c, (byte)0xea)` in Java source code. +# The following tests check that the compiler converts the string literals to readable ones. + +AT_DATA([prog1.cob], [ + IDENTIFICATION DIVISION. + PROGRAM-ID. prog1. + DATA DIVISION. + WORKING-STORAGE SECTION. + 01 F0 PIC X(30) VALUE "東京1". + PROCEDURE DIVISION. + MOVE "東京2" TO F0. + DISPLAY "東京3". +]) + +AT_CHECK([cobj prog1.cob]) +AT_CHECK([grep '東京1' < prog1.java > /dev/null]) +AT_CHECK([grep '東京2' < prog1.java > /dev/null]) +AT_CHECK([grep '東京3' < prog1.java > /dev/null]) + +# '　' is the first multi-byte Shift-JIS character with respect to the byte order +# see http://charset.7jp.net/sjis.html +AT_DATA([prog2.cob], [ + IDENTIFICATION DIVISION. + PROGRAM-ID. prog2. + DATA DIVISION. + WORKING-STORAGE SECTION. + 01 F0 PIC X(30) VALUE "　1". + PROCEDURE DIVISION. + MOVE "　2" TO F0. + DISPLAY "　3". +]) + +AT_CHECK([cobj prog2.cob]) +AT_CHECK([grep '　1' < prog2.java > /dev/null]) +AT_CHECK([grep '　2' < prog2.java > /dev/null]) +AT_CHECK([grep '　3' < prog2.java > /dev/null]) + +# '熙' is the last printable Shift-JIS character with respect to the byte order. +# See http://charset.7jp.net/sjis.html +AT_DATA([prog3.cob], [ + IDENTIFICATION DIVISION. + PROGRAM-ID. prog3. + DATA DIVISION. + WORKING-STORAGE SECTION. + 01 F0 PIC X(30) VALUE "熙1". + PROCEDURE DIVISION. + MOVE "熙2" TO F0. + DISPLAY "熙3". +]) + +AT_CHECK([cobj prog3.cob]) +AT_CHECK([grep '熙1' < prog3.java > /dev/null]) +AT_CHECK([grep '熙2' < prog3.java > /dev/null]) +AT_CHECK([grep '熙3' < prog3.java > /dev/null]) + +AT_CLEANUP \ No newline at end of file