This is an automated email from the ASF dual-hosted git repository.
leginee pushed a commit to branch trunk
in repository https://gitbox.apache.org/repos/asf/openoffice.git
The following commit(s) were added to refs/heads/trunk by this push:
new a9afaeb67b Fix sal encoding-dispatch off-by-one (ISCII removal) +
ZapfDingbats table bugs (#488)
a9afaeb67b is described below
commit a9afaeb67b88027bd0a60034e84d56dc7c60c68b
Author: Peter <[email protected]>
AuthorDate: Sun Jul 5 17:15:23 2026 +0000
Fix sal encoding-dispatch off-by-one (ISCII removal) + ZapfDingbats table
bugs (#488)
Impl_getTextEncodingData() is a positional array indexed by the ABI-frozen
rtl_TextEncoding enum. #i119141 removed the ISCII converter but deleted its
array slot with no NULL placeholder, shifting JAVA_UTF8, ADOBE_STANDARD,
ADOBE_SYMBOL, PT154 and ADOBE_DINGBATS onto the wrong converter. Restore the
NULL slot and add a compile-time size guard.
Wiring ADOBE_DINGBATS correctly surfaced three latent ZapfDingbats table
errors (0x6C→U+25CF; reverse range to U+2797; circled-digits reverse start
0xAC); fix all three. Update rtl_textcvt qa: park the now-unimplemented
ISCII
behind TEST_ISCII_DEVANAGARI, correct the ISCII Windows codepage to 57002.
---
main/sal/qa/rtl/textenc/rtl_textcvt.cxx | 20 +++++++++++++++++---
main/sal/textenc/convertadobe.tab | 6 +++---
main/sal/textenc/textenc.cxx | 14 ++++++++++++++
3 files changed, 34 insertions(+), 6 deletions(-)
diff --git a/main/sal/qa/rtl/textenc/rtl_textcvt.cxx
b/main/sal/qa/rtl/textenc/rtl_textcvt.cxx
index dd5f68e999..ec1148f08a 100644
--- a/main/sal/qa/rtl/textenc/rtl_textcvt.cxx
+++ b/main/sal/qa/rtl/textenc/rtl_textcvt.cxx
@@ -36,6 +36,13 @@
#include "sal/types.h"
#include "gtest/gtest.h"
+/* The ISCII Devanagari converter was removed for licensing reasons (its table
+ was derived from LGPL code; see #i119141), so
RTL_TEXTENCODING_ISCII_DEVANAGARI
+ is currently unimplemented and has no converter. The test data that
exercises
+ it is kept below but compiled out. Define this macro to re-enable those
+ checks once an Apache-licensed ISCII converter is reintroduced. */
+// #define TEST_ISCII_DEVANAGARI
+
namespace {
struct SingleByteCharSet {
@@ -1062,6 +1069,7 @@ TEST_F(Test, testSingleByte) {
0x0425,0x0418,0x0419,0x041A,0x041B,0x041C,0x041D,0x041E,
0x041F,0x042F,0x0420,0x0421,0x0422,0x0423,0x0416,0x0412,
0x042C,0x042B,0x0417,0x0428,0x042D,0x0429,0x0427,0x042A } },
+#ifdef TEST_ISCII_DEVANAGARI
{ RTL_TEXTENCODING_ISCII_DEVANAGARI,
{ 0x0000,0x0001,0x0002,0x0003,0x0004,0x0005,0x0006,0x0007,
0x0008,0x0009,0x000A,0x000B,0x000C,0x000D,0x000E,0x000F,
@@ -1095,6 +1103,7 @@ TEST_F(Test, testSingleByte) {
0x094D,0x093C,0x0964,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,
0xFFFF,0x0966,0x0967,0x0968,0x0969,0x096A,0x096B,0x096C,
0x096D,0x096E,0x096F,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF } },
+#endif
{ RTL_TEXTENCODING_ADOBE_STANDARD,
{ 0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,
0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,0xFFFF,
@@ -1178,7 +1187,7 @@ TEST_F(Test, testSingleByte) {
0x2738,0x2739,0x273A,0x273B,0x273C,0x273D,0x273E,0x273F,
// 60
0x2740,0x2741,0x2742,0x2743,0x2744,0x2745,0x2746,0x2747,
- 0x2748,0x2749,0x274A,0x274B,0x27CF,0x274D,0x25A0,0x274F,
+ 0x2748,0x2749,0x274A,0x274B,0x25CF,0x274D,0x25A0,0x274F,
0x2750,0x2751,0x2752,0x25B2,0x25BC,0x25C6,0x2756,0x25D7,
0x2758,0x2759,0x275A,0x275B,0x275C,0x275D,0x275E,0xFFFF,
// 80
@@ -1194,7 +1203,7 @@ TEST_F(Test, testSingleByte) {
// C0
0x2780,0x2781,0x2782,0x2783,0x2784,0x2785,0x2786,0x2787,
0x2788,0x2789,0x278A,0x278B,0x278C,0x278D,0x278E,0x278F,
- 0x2790,0x2791,0x2792,0x2793,0x2794,0x2795,0x2796,0x2797,
+ 0x2790,0x2791,0x2792,0x2793,0x2794,0x2192,0x2194,0x2195,
0x2798,0x2799,0x279A,0x279B,0x279C,0x279D,0x279E,0x279F,
// E0
0x27A0,0x27A1,0x27A2,0x27A3,0x27A4,0x27A5,0x27A6,0x27A7,
@@ -2775,7 +2784,10 @@ TEST_F(Test, testWindows) {
{ 0, RTL_TEXTENCODING_DONTKNOW, true },
{ 0, RTL_TEXTENCODING_UCS4, true },
{ 0, RTL_TEXTENCODING_UCS2, true },
- { 0, RTL_TEXTENCODING_ISCII_DEVANAGARI, true }
+ // 57002 is the genuine Windows code page for ISCII Devanagari; the
+ // codepage<->encoding mapping is a fixed table independent of whether
+ // a converter is implemented (the converter was removed, see
#i119141).
+ { 57002, RTL_TEXTENCODING_ISCII_DEVANAGARI, true }
};
for (std::size_t i = 0; i < sizeof data / sizeof data[0]; ++i) {
OSL_ASSERT(data[i].codePage != 0 || data[i].reverse);
@@ -2832,8 +2844,10 @@ TEST_F(Test, testInfo) {
{ RTL_TEXTENCODING_IBM_861, RTL_TEXTENCODING_INFO_MIME, true },
{ RTL_TEXTENCODING_IBM_863, RTL_TEXTENCODING_INFO_MIME, true },
{ RTL_TEXTENCODING_IBM_865, RTL_TEXTENCODING_INFO_MIME, true },
+#ifdef TEST_ISCII_DEVANAGARI
{ RTL_TEXTENCODING_ISCII_DEVANAGARI, RTL_TEXTENCODING_INFO_ASCII, true
},
{ RTL_TEXTENCODING_ISCII_DEVANAGARI, RTL_TEXTENCODING_INFO_MIME, false
},
+#endif
{ RTL_TEXTENCODING_ADOBE_STANDARD, RTL_TEXTENCODING_INFO_ASCII, false
},
{ RTL_TEXTENCODING_ADOBE_STANDARD, RTL_TEXTENCODING_INFO_MIME, true },
{ RTL_TEXTENCODING_ADOBE_STANDARD, RTL_TEXTENCODING_INFO_SYMBOL, false
},
diff --git a/main/sal/textenc/convertadobe.tab
b/main/sal/textenc/convertadobe.tab
index 2a59313889..1191211da2 100644
--- a/main/sal/textenc/convertadobe.tab
+++ b/main/sal/textenc/convertadobe.tab
@@ -233,7 +233,7 @@ adobeSymbolEncodingConverterData = {
rtl::textenc::BmpUnicodeToSingleByteRange const
unicodeToAdobeDingbatsEncoding[]
= { { 0x0020, 0x0020 - 0x0020, 0x20 }, { 0x00A0, 0x00A0 - 0x00A0, 0x20 },
{ 0x2192, 0x2192 - 0x2192, 0xD5 }, { 0x2194, 0x2195 - 0x2194, 0xD6 },
- { 0x2460, 0x2469 - 0x2460, 0xAB },
+ { 0x2460, 0x2469 - 0x2460, 0xAC },
{ 0x25A0, 0x25A0 - 0x25A0, 0x6E },
{ 0x25B2, 0x25B2 - 0x25B2, 0x73 }, { 0x25BC, 0x25BC - 0x25BC, 0x74 },
{ 0x25C6, 0x25C6 - 0x25C6, 0x75 }, { 0x25CF, 0x25CF - 0x25CF, 0x6C },
@@ -266,7 +266,7 @@ adobeDingbatsEncodingConverterData = {
0x2730, 0x2731, 0x2732, 0x2733, 0x2734, 0x2735, 0x2736, 0x2737, // 0x50
0x2738, 0x2739, 0x273A, 0x273B, 0x273C, 0x273D, 0x273E, 0x273F,
0x2740, 0x2741, 0x2742, 0x2743, 0x2744, 0x2745, 0x2746, 0x2747, // 0x60
- 0x2748, 0x2749, 0x274A, 0x274B, 0x27CF, 0x274D, 0x25A0, 0x274F,
+ 0x2748, 0x2749, 0x274A, 0x274B, 0x25CF, 0x274D, 0x25A0, 0x274F,
0x2750, 0x2751, 0x2752, 0x25B2, 0x25BC, 0x25C6, 0x2756, 0x25D7, // 0x70
0x2758, 0x2759, 0x275A, 0x275B, 0x275C, 0x275D, 0x275E, 0xFFFF,
0xF8D7, 0xF8D8, 0xF8D9, 0xF8DA, 0xF8DB, 0xF8DC, 0xF8DD, 0xF8DE, // 0x80
@@ -279,7 +279,7 @@ adobeDingbatsEncodingConverterData = {
0x2778, 0x2779, 0x277A, 0x277B, 0x277C, 0x277D, 0x277E, 0x277F,
0x2780, 0x2781, 0x2782, 0x2783, 0x2784, 0x2785, 0x2786, 0x2787, // 0xC0
0x2788, 0x2789, 0x278A, 0x278B, 0x278C, 0x278D, 0x278E, 0x278F,
- 0x2790, 0x2791, 0x2792, 0x2793, 0x2794, 0x2795, 0x2796, 0x2797, // 0xD0
+ 0x2790, 0x2791, 0x2792, 0x2793, 0x2794, 0x2192, 0x2194, 0x2195, // 0xD0
0x2798, 0x2799, 0x279A, 0x279B, 0x279C, 0x279D, 0x279E, 0x279F,
0x27A0, 0x27A1, 0x27A2, 0x27A3, 0x27A4, 0x27A5, 0x27A6, 0x27A7, // 0xE0
0x27A8, 0x27A9, 0x27AA, 0x27AB, 0x27AC, 0x27AD, 0x27AE, 0x27AF,
diff --git a/main/sal/textenc/textenc.cxx b/main/sal/textenc/textenc.cxx
index 746d97ce07..ddb207be39 100644
--- a/main/sal/textenc/textenc.cxx
+++ b/main/sal/textenc/textenc.cxx
@@ -210,11 +210,25 @@ Impl_getTextEncodingData(rtl_TextEncoding nEncoding)
SAL_THROW_EXTERN_C()
&aImplBig5HkscsTextEncodingData, /* BIG5_HKSCS */
&aImplTis620TextEncodingData, /* TIS_620 */
&aImplKoi8UTextEncodingData, /* KOI8_U */
+ NULL, /* ISCII_DEVANAGARI: converter removed (#i119141), but the
+ slot MUST remain. This array is indexed positionally by
+ the (ABI-frozen) rtl_TextEncoding enum value, so a missing
+ slot shifts every following encoding onto the wrong
+ converter. */
&aImplJavaUtf8TextEncodingData, /* JAVA_UTF8 */
&adobeStandardEncodingData, /* ADOBE_STANDARD */
&adobeSymbolEncodingData, /* ADOBE_SYMBOL */
&aImplPT154TextEncodingData, /* PT154 */
&adobeDingbatsEncodingData }; /* ADOBE_DINGBATS */
+ /* aData[] is indexed positionally by the (ABI-frozen) rtl_TextEncoding
+ enum, so its length must equal the highest enum value plus one. Assert
+ this at compile time: dropping a slot (as happened to ISCII_DEVANAGARI,
+ #i119141) would otherwise silently shift every following encoding onto
+ the wrong converter, undetectable at run time because the bounds check
+ below just maps the now-out-of-range top encoding to NULL. */
+ typedef int Impl_textEncodingDataSizeCheck[
+ sizeof aData / sizeof aData[0]
+ == RTL_TEXTENCODING_ADOBE_DINGBATS + 1 ? 1 : -1];
return
nEncoding < sizeof aData / sizeof aData[0] ? aData[nEncoding] : NULL;
}
