changeset
add support for line continuation in the properties parser - resolves #457
Wed, 10 Dec 2025 13:12:27 +0100
- author
- Olaf Wintermann <olaf.wintermann@gmail.com>
- date
- Wed, 10 Dec 2025 13:12:27 +0100
- changeset 1562
- f2b63cad2142
- parent 1561
- fcebf53de51c
- child 1563
- 6e2f83ef6f5a
add support for line continuation in the properties parser - resolves #457
| src/cx/properties.h | file | annotate | diff | comparison | revisions | |
| src/properties.c | file | annotate | diff | comparison | revisions | |
| tests/test_properties.c | file | annotate | diff | comparison | revisions |
--- a/src/cx/properties.h Tue Dec 09 19:05:35 2025 +0100 +++ b/src/cx/properties.h Wed Dec 10 13:12:27 2025 +0100 @@ -77,9 +77,6 @@ * The character, when appearing at the end of a line, continues that line. * This is '\' by default. */ - /** - * Reserved for future use. - */ char continuation; };
--- a/src/properties.c Tue Dec 09 19:05:35 2025 +0100 +++ b/src/properties.c Wed Dec 10 13:12:27 2025 +0100 @@ -31,12 +31,13 @@ #include <assert.h> #include <stdio.h> #include <string.h> +#include <ctype.h> const CxPropertiesConfig cx_properties_config_default = { - '=', - '#', - '\0', - '\0', + '=', + '#', + '\0', + '\0', '\\', }; @@ -96,13 +97,30 @@ // a pointer to the buffer we want to read from CxBuffer *current_buffer = &prop->input; - + + char comment1 = prop->config.comment1; + char comment2 = prop->config.comment2; + char comment3 = prop->config.comment3; + char delimiter = prop->config.delimiter; + char continuation = prop->config.continuation; + // check if we have rescued data if (!cxBufferEof(&prop->buffer)) { // check if we can now get a complete line cxstring input = cx_strn(prop->input.space + prop->input.pos, prop->input.size - prop->input.pos); cxstring nl = cx_strchr(input, '\n'); + while (nl.length > 0) { + // check for line continuation + char previous = nl.ptr > input.ptr ? nl.ptr[-1] : prop->buffer.space[prop->buffer.size-1]; + if (previous == continuation) { + // this nl is a line continuation, check the next newline + nl = cx_strchr(cx_strsubs(nl, 1), '\n'); + } else { + break; + } + } + if (nl.length > 0) { // we add as much data to the rescue buffer as we need // to complete the line @@ -129,12 +147,7 @@ return CX_PROPERTIES_INCOMPLETE_DATA; } } - - char comment1 = prop->config.comment1; - char comment2 = prop->config.comment2; - char comment3 = prop->config.comment3; - char delimiter = prop->config.delimiter; - + // get one line and parse it while (!cxBufferEof(current_buffer)) { const char *buf = current_buffer->space + current_buffer->pos; @@ -147,6 +160,7 @@ size_t delimiter_index = 0; size_t comment_index = 0; bool has_comment = false; + bool has_continuation = false; size_t i = 0; char c = 0; @@ -161,6 +175,9 @@ if (delimiter_index == 0 && !has_comment) { delimiter_index = i; } + } else if (delimiter_index > 0 && c == continuation && i+1 < len && buf[i+1] == '\n') { + has_continuation = true; + i++; } else if (c == '\n') { break; } @@ -225,10 +242,53 @@ k = cx_strtrim(k); val = cx_strtrim(val); if (k.length > 0) { + current_buffer->pos += i + 1; + assert(current_buffer->pos <= current_buffer->size); + assert(current_buffer != &prop->buffer || current_buffer->pos == current_buffer->size); + + if (has_continuation) { + char *ptr = (char*)val.ptr; + if (current_buffer != &prop->buffer) { + // move value to the rescue buffer + if (prop->buffer.space == NULL) { + cxBufferInit(&prop->buffer, NULL, 256, NULL, CX_BUFFER_AUTO_EXTEND); + } + prop->buffer.size = 0; + prop->buffer.pos = 0; + if (cxBufferWrite(val.ptr, 1, val.length, &prop->buffer) != val.length) { + return CX_PROPERTIES_BUFFER_ALLOC_FAILED; + } + val.ptr = prop->buffer.space; + ptr = prop->buffer.space; + } + // value.ptr is now inside the rescue buffer and we can + // remove the continuation character from the value + bool trim = false; + size_t x = 0; + for(size_t j=0;j<val.length;j++) { + c = ptr[j]; + if (j+1 < val.length && c == '\\' && ptr[j+1] == '\n') { + // skip continuation and newline character + j++; + trim = true; // enable trim in the next line + continue; + } + if (j > x) { + if (trim) { + if (isspace((unsigned char)c)) { + continue; + } + trim = false; + } + ptr[x] = c; + } + x++; + } + val.length = x; + } *key = k; *value = val; - current_buffer->pos += i + 1; - assert(current_buffer->pos <= current_buffer->size); + return CX_PROPERTIES_NO_ERROR; } else { return CX_PROPERTIES_INVALID_EMPTY_KEY;
--- a/tests/test_properties.c Tue Dec 09 19:05:35 2025 +0100 +++ b/tests/test_properties.c Wed Dec 10 13:12:27 2025 +0100 @@ -41,6 +41,7 @@ CX_TEST_ASSERT(prop.config.comment1 == '#'); CX_TEST_ASSERT(prop.config.comment2 == 0); CX_TEST_ASSERT(prop.config.comment3 == 0); + CX_TEST_ASSERT(prop.config.continuation == '\\'); CX_TEST_ASSERT(prop.input.space == NULL); CX_TEST_ASSERT(prop.buffer.space == NULL); @@ -383,6 +384,108 @@ free(long_value); } +CX_TEST(test_properties_next_line_continuation) { + const char *str = + "key1 = multiline \\\nvalue\n" + "key2 = normal\n" + "key3 = multiline \\\n trim \n" + "key4 = m1\\\nm2\\\n m3\\\nm4 \n" + "key5 = no\\continuation\n"; + + CxProperties prop; + cxPropertiesInitDefault(&prop); + + cxstring key; + cxstring value; + + CX_TEST_DO { + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, str)); + + CX_TEST_ASSERT(cxPropertiesNext(&prop, &key, &value) == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key1")); + CX_TEST_ASSERT(!cx_strcmp(value, "multiline value")); + + CX_TEST_ASSERT(cxPropertiesNext(&prop, &key, &value) == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key2")); + CX_TEST_ASSERT(!cx_strcmp(value, "normal")); + + CX_TEST_ASSERT(cxPropertiesNext(&prop, &key, &value) == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key3")); + CX_TEST_ASSERT(!cx_strcmp(value, "multiline trim")); + + CX_TEST_ASSERT(cxPropertiesNext(&prop, &key, &value) == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key4")); + CX_TEST_ASSERT(!cx_strcmp(value, "m1m2m3m4")); + + CX_TEST_ASSERT(cxPropertiesNext(&prop, &key, &value) == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key5")); + CX_TEST_ASSERT(!cx_strcmp(value, "no\\continuation")); + } + + cxPropertiesDestroy(&prop); +} + +CX_TEST(test_properties_next_line_continuation_part) { + CxProperties prop; + cxPropertiesInitDefault(&prop); + + cxstring key; + cxstring value; + CxPropertiesStatus result; + const char *str; + + CX_TEST_DO { + // key1 = continue continue ...line + str = "key1 "; + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, str)); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_INCOMPLETE_DATA); + + str = "= continue \\"; + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, str)); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_INCOMPLETE_DATA); + + str = "\ncontinue \\\n"; + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, str)); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_INCOMPLETE_DATA); + + str = "..."; + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, str)); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_INCOMPLETE_DATA); + + str = "line\nkey2 = value2\n"; + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, str)); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key1")); + CX_TEST_ASSERT(!cx_strcmp(value, "continue continue ...line")); + + // key2 = value2 + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(!cx_strcmp(key, "key2")); + CX_TEST_ASSERT(!cx_strcmp(value, "value2")); + + // key3 = continue-line + str = "key3=\\\ncontinue-\\\n line"; + size_t len = strlen(str); + for(size_t i=0;i<len;i++) { + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, cx_strn(str+i, 1))); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_INCOMPLETE_DATA); + } + CX_TEST_ASSERT(0 == cxPropertiesFill(&prop, "\n")); + result = cxPropertiesNext(&prop, &key, &value); + CX_TEST_ASSERT(result == CX_PROPERTIES_NO_ERROR); + CX_TEST_ASSERT(!cx_strcmp(key, "key3")); + CX_TEST_ASSERT(!cx_strcmp(value, "continue-line")); + } + + cxPropertiesDestroy(&prop); +} + CX_TEST(test_properties_load) { CxTestingAllocator talloc; cx_testing_allocator_init(&talloc); @@ -710,6 +813,8 @@ cx_test_register(suite, test_properties_next_multi); cx_test_register(suite, test_properties_next_part); cx_test_register(suite, test_properties_next_long_lines); + cx_test_register(suite, test_properties_next_line_continuation); + cx_test_register(suite, test_properties_next_line_continuation_part); cx_test_register(suite, test_properties_load); cx_test_register(suite, test_properties_load_empty_file); cx_test_register(suite, test_properties_load_only_comments);
