You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@logging.apache.org by "jmestwa-coder (via GitHub)" <gi...@apache.org> on 2026/05/23 10:21:08 UTC
[PR] SONLayout escaping for malformed UTF-8 input [logging-log4cxx]
jmestwa-coder opened a new pull request, #686:
URL: https://github.com/apache/logging-log4cxx/pull/686
## Summary
Fix JSONLayout escaping for malformed UTF-8 input.
Previously, malformed decoded input could bypass replacement escaping and preserve the original malformed bytes in emitted JSON output. This happened because the replacement character (`U+FFFD`) incorrectly followed the printable fast path.
This change ensures malformed decoded input and invalid Unicode scalar values are consistently emitted through the JSON escaping path as `\ufffd`.
## Changes
- Add `escapeReplacement` handling in `JSONLayout::appendItem()`
- Ensure replacement characters bypass the printable fast path
- Preserve existing behavior for valid UTF-8 input
- Add regression coverage for malformed UTF-8 escaping behavior
## Regression Test
Added a focused regression test for malformed UTF-8 input to verify malformed bytes no longer survive into JSON output and are emitted as escaped replacement characters instead.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@logging.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] JSONLayout escaping for malformed UTF-8 input [logging-log4cxx]
Posted by "swebb2066 (via GitHub)" <gi...@apache.org>.
swebb2066 closed pull request #686: JSONLayout escaping for malformed UTF-8 input
URL: https://github.com/apache/logging-log4cxx/pull/686
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@logging.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] JSONLayout escaping for malformed UTF-8 input [logging-log4cxx]
Posted by "swebb2066 (via GitHub)" <gi...@apache.org>.
swebb2066 commented on code in PR #686:
URL: https://github.com/apache/logging-log4cxx/pull/686#discussion_r3295753892
##########
src/main/cpp/jsonlayout.cpp:
##########
@@ -223,18 +223,21 @@ void JSONLayout::appendItem(const LogString& input, LogString& buf)
{
auto lastCodePoint = nextCodePoint;
auto ch = Transcoder::decode(input, nextCodePoint);
+ bool escapeReplacement = false;
if (nextCodePoint == lastCodePoint) // failed to decode input?
{
nextCodePoint = input.end();
ch = 0xFFFD; // The Unicode replacement character
+ escapeReplacement = true;
}
else if ((0xD800 <= ch && ch <= 0xDFFF) || 0x10FFFF < ch)
{
ch = 0xFFFD; // The Unicode replacement character
+ escapeReplacement = true;
}
else if (0x22 == ch || 0x5c == ch) // double quote or backslash?
;
- else if (0x20 <= ch) // not a control character?
+ else if (!escapeReplacement && 0x20 <= ch) // not a control character?
Review Comment:
This change does nothing as `escapeReplacement` is always `false` in this conditional expression.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@logging.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] JSONLayout escaping for malformed UTF-8 input [logging-log4cxx]
Posted by "jmestwa-coder (via GitHub)" <gi...@apache.org>.
jmestwa-coder commented on code in PR #686:
URL: https://github.com/apache/logging-log4cxx/pull/686#discussion_r3297047255
##########
src/main/cpp/jsonlayout.cpp:
##########
@@ -223,18 +223,21 @@ void JSONLayout::appendItem(const LogString& input, LogString& buf)
{
auto lastCodePoint = nextCodePoint;
auto ch = Transcoder::decode(input, nextCodePoint);
+ bool escapeReplacement = false;
if (nextCodePoint == lastCodePoint) // failed to decode input?
{
nextCodePoint = input.end();
ch = 0xFFFD; // The Unicode replacement character
+ escapeReplacement = true;
}
else if ((0xD800 <= ch && ch <= 0xDFFF) || 0x10FFFF < ch)
{
ch = 0xFFFD; // The Unicode replacement character
+ escapeReplacement = true;
}
else if (0x22 == ch || 0x5c == ch) // double quote or backslash?
;
- else if (0x20 <= ch) // not a control character?
+ else if (!escapeReplacement && 0x20 <= ch) // not a control character?
Review Comment:
Thanks for the review and for pointing this out.
I re-validated the runtime behavior against the malformed UTF-8 testcase and confirmed the change does not actually alter the emitted output as intended. The new condition is effectively a no-op in the current control flow.
Given that, I’m going to withdraw this patch rather than push a speculative follow-up change without a demonstrated failing case.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@logging.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org