diff --git a/.agent/consumer.md b/.agent/consumer.md index a6a62c8..7627d92 100644 --- a/.agent/consumer.md +++ b/.agent/consumer.md @@ -23,8 +23,8 @@ - Redis and scheduling are opt-in through their `kri.*.enabled` settings and matching runtime dependencies. - HTTP bypass targets are server-configured. Never accept an arbitrary full target URL from a frontend; retain allowed-path checks and do not forward hop-by-hop headers. - The external API client returns response bodies as documented, while bypass responses preserve raw bytes and target `Content-Type`. -- `ExternalApiClient` multipart requests generate their own boundary and ignore a caller-supplied `Content-Type`. Multipart boundaries and part headers use US-ASCII, `formField` values use UTF-8, and file bodies preserve the supplied bytes. -- ASCII multipart filenames are sent unchanged in `filename`. Non-ASCII filenames use a collision-free ASCII `filename` fallback plus the original name in UTF-8 percent-encoded `filename*`; spaces are encoded as `%20`, not `+`. +- `ExternalApiClient` multipart requests generate their own boundary and ignore a caller-supplied `Content-Type`. Multipart headers and `formField` values use UTF-8, while file bodies preserve the supplied bytes. +- Multipart filenames, including non-ASCII names, are sent unchanged in the quoted `filename` parameter as raw UTF-8. No ASCII fallback or `filename*` parameter is added; CR/LF rejection and quote/backslash escaping still apply. - `JsonUtil` requires the consumer's explicit `ObjectMapper`; shared date/time helpers define KST as `Asia/Seoul` and also expose UTC semantics. ### Upgrade and Verification diff --git a/README.md b/README.md index de01b67..2eb24ea 100644 --- a/README.md +++ b/README.md @@ -186,15 +186,12 @@ ExternalApiResponse res = client.postMultipart(url, headers, parts); multipart 요청에서는 전달한 headers의 `Content-Type` 값을 무시하고, 필요한 boundary를 포함한 `Content-Type: multipart/form-data; boundary=...` 값을 자동 생성합니다. -boundary와 multipart 파트 헤더는 US-ASCII로 생성하고, 텍스트 필드 내용은 -UTF-8, 파일 내용은 원본 `byte[]`로 전송합니다. ASCII 파일명은 기존 `filename`을 -그대로 사용합니다. non-ASCII 파일명은 서버 호환성을 위해 충돌하지 않는 ASCII -fallback을 `filename`에 넣고, UTF-8 퍼센트 인코딩한 원본 파일명을 `filename*`에 -함께 전송합니다. 공백은 `+`가 아닌 `%20`으로 인코딩됩니다. +multipart 파트 헤더와 텍스트 필드는 UTF-8로 인코딩하고, 파일 내용은 원본 +`byte[]`로 전송합니다. ASCII와 non-ASCII 파일명 모두 원본 값을 quoted `filename`에 +UTF-8 바이트로 전송하며, ASCII fallback이나 `filename*`는 추가하지 않습니다. ```http -Content-Disposition: form-data; name="file"; filename="attachment.txt"; - filename*=UTF-8''%ED%95%9C%EA%B8%80-%ED%8C%8C%EC%9D%BC.txt +Content-Disposition: form-data; name="file"; filename="한글-파일.txt" ``` 지원 메서드: @@ -773,5 +770,8 @@ Endpoint: ## 변경 이력 -- 2026-08-06: non-ASCII multipart 파일명을 ASCII `filename` fallback과 UTF-8 `filename*`로 함께 전송하도록 개선했습니다. +- 2026-08-06: `beta.1.6.3`에서 non-ASCII multipart 파일명을 ASCII fallback/ + `filename*` 없이 원본 `filename`의 UTF-8 바이트로 전송하도록 복원했습니다. +- 2026-08-06: `beta.1.6.2`에서 non-ASCII multipart 파일명에 ASCII `filename` + fallback과 UTF-8 `filename*`를 추가했습니다. - 2026-08-04: `kri.auth.apikey.allowed-paths`와 `kri.auth.apikey.excluded-paths`를 추가했습니다. 미설정·`null`·빈 목록은 기존 전역 API Key 인증 동작을 유지하며, 제외 패턴은 허용 패턴보다 우선합니다. diff --git a/build.gradle b/build.gradle index b31dbca..427408f 100644 --- a/build.gradle +++ b/build.gradle @@ -4,7 +4,7 @@ plugins { } // 버전 변경은 여기서 -def commonVer = "beta.1.6.2" +def commonVer = "beta.1.6.3" group = 'com.kri.common' version = commonVer diff --git a/src/main/java/com/kri/common/feature/http/ExternalApiClient.java b/src/main/java/com/kri/common/feature/http/ExternalApiClient.java index 1249762..210a08a 100644 --- a/src/main/java/com/kri/common/feature/http/ExternalApiClient.java +++ b/src/main/java/com/kri/common/feature/http/ExternalApiClient.java @@ -10,8 +10,6 @@ import java.nio.charset.StandardCharsets; import java.time.Duration; import java.util.List; import java.util.Map; -import java.util.Set; -import java.util.TreeSet; import java.util.UUID; /** @@ -133,33 +131,25 @@ public class ExternalApiClient { } ByteArrayOutputStream out = new ByteArrayOutputStream(); - Set usedAsciiFilenames = collectAsciiFilenames(parts); for (ExternalApiMultipartPart part : parts) { if (part == null) { throw new IllegalArgumentException("parts must not contain null"); } - writeAscii(out, "--" + boundary + CRLF); - writeAscii(out, "Content-Disposition: form-data; name=\"" + headerValue(part.getName()) + "\""); + writeUtf8(out, "--" + boundary + CRLF); + writeUtf8(out, "Content-Disposition: form-data; name=\"" + headerValue(part.getName()) + "\""); if (part.getFilename() != null) { - String filename = safeHeader(part.getFilename()); - if (isAscii(filename)) { - writeAscii(out, "; filename=\"" + headerValue(filename) + "\""); - } else { - String fallbackFilename = uniqueAsciiFallback(filename, usedAsciiFilenames); - writeAscii(out, "; filename=\"" + headerValue(fallbackFilename) + "\""); - writeAscii(out, "; filename*=UTF-8''" + percentEncodeUtf8(filename)); - } + writeUtf8(out, "; filename=\"" + headerValue(part.getFilename()) + "\""); } - writeAscii(out, CRLF); + writeUtf8(out, CRLF); if (part.getContentType() != null && !part.getContentType().isBlank()) { - writeAscii(out, "Content-Type: " + safeHeader(part.getContentType()) + CRLF); + writeUtf8(out, "Content-Type: " + safeHeader(part.getContentType()) + CRLF); } - writeAscii(out, CRLF); + writeUtf8(out, CRLF); out.write(part.content()); - writeAscii(out, CRLF); + writeUtf8(out, CRLF); } - writeAscii(out, "--" + boundary + "--" + CRLF); + writeUtf8(out, "--" + boundary + "--" + CRLF); return out.toByteArray(); } @@ -175,85 +165,10 @@ public class ExternalApiClient { }); } - private static void writeAscii(ByteArrayOutputStream out, String value) throws IOException { - out.write(value.getBytes(StandardCharsets.US_ASCII)); - } - - private static Set collectAsciiFilenames(List parts) { - Set filenames = new TreeSet<>(String.CASE_INSENSITIVE_ORDER); - for (ExternalApiMultipartPart part : parts) { - if (part != null && part.getFilename() != null) { - String filename = safeHeader(part.getFilename()); - if (isAscii(filename)) { - filenames.add(filename); - } - } - } - return filenames; - } - - private static String uniqueAsciiFallback(String filename, Set usedFilenames) { - String extension = asciiExtension(filename); - String candidate = "attachment" + extension; - int sequence = 2; - while (!usedFilenames.add(candidate)) { - candidate = "attachment-" + sequence++ + extension; - } - return candidate; - } - - private static String asciiExtension(String filename) { - int lastSeparator = Math.max(filename.lastIndexOf('/'), filename.lastIndexOf('\\')); - int lastDot = filename.lastIndexOf('.'); - if (lastDot <= lastSeparator || lastDot == filename.length() - 1) { - return ""; - } - - String extension = filename.substring(lastDot + 1); - if (extension.length() > 16 || !extension.chars().allMatch(Character::isLetterOrDigit) - || !isAscii(extension)) { - return ""; - } - return "." + extension; - } - - private static boolean isAscii(String value) { - return value.chars().allMatch(ch -> ch <= 0x7F); - } - - private static String percentEncodeUtf8(String value) { - byte[] bytes = value.getBytes(StandardCharsets.UTF_8); - StringBuilder encoded = new StringBuilder(bytes.length * 3); - char[] hex = "0123456789ABCDEF".toCharArray(); - for (byte current : bytes) { - int unsigned = current & 0xFF; - if (isRfc5987AttrChar(unsigned)) { - encoded.append((char) unsigned); - } else { - encoded.append('%') - .append(hex[unsigned >>> 4]) - .append(hex[unsigned & 0x0F]); - } - } - return encoded.toString(); - } - - private static boolean isRfc5987AttrChar(int value) { - return value >= 'A' && value <= 'Z' - || value >= 'a' && value <= 'z' - || value >= '0' && value <= '9' - || value == '!' - || value == '#' - || value == '$' - || value == '&' - || value == '+' - || value == '-' - || value == '.' - || value == '^' - || value == '_' - || value == '`' - || value == '|' - || value == '~'; + private static void writeUtf8(ByteArrayOutputStream out, String value) throws IOException { + // Multipart delimiters remain byte-identical because UTF-8 is ASCII-compatible, while + // user-visible field names and filenames retain non-ASCII characters. + out.write(value.getBytes(StandardCharsets.UTF_8)); } private static String headerValue(String value) {