This is an automated email from the ASF dual-hosted git repository.
jianliangqi pushed a commit to branch clucene
in repository https://gitbox.apache.org/repos/asf/doris-thirdparty.git
The following commit(s) were added to refs/heads/clucene by this push:
new 9e60ec66 [fix](keyword) fix the keyword type index length limit (#114)
9e60ec66 is described below
commit 9e60ec666b3ccf7dd8b7c3e331ac03ccf87d5845
Author: zzzxl <[email protected]>
AuthorDate: Fri Aug 25 17:24:57 2023 +0800
[fix](keyword) fix the keyword type index length limit (#114)
---
src/core/CLucene/debug/error.h | 1 +
src/core/CLucene/index/IndexWriter.cpp | 11 ++++-------
src/core/CLucene/index/SDocumentWriter.cpp | 14 ++++++++++----
3 files changed, 15 insertions(+), 11 deletions(-)
diff --git a/src/core/CLucene/debug/error.h b/src/core/CLucene/debug/error.h
index 821d9e13..0c48db3d 100644
--- a/src/core/CLucene/debug/error.h
+++ b/src/core/CLucene/debug/error.h
@@ -33,6 +33,7 @@
#define CL_ERR_MergeAborted 22
#define CL_ERR_OutOfMemory 23
#define CL_ERR_FieldReader 24
+#define CL_ERR_MaxBytesLength 25
////////////////////////////////////////////////////////
//error try/throw/catch definitions
diff --git a/src/core/CLucene/index/IndexWriter.cpp
b/src/core/CLucene/index/IndexWriter.cpp
index 87cfaa5a..fe269a92 100644
--- a/src/core/CLucene/index/IndexWriter.cpp
+++ b/src/core/CLucene/index/IndexWriter.cpp
@@ -722,13 +722,10 @@ void IndexWriter::addDocument(Document *doc, Analyzer
*an) {
message(string("hit exception adding document"));
{
- SCOPED_LOCK_MUTEX(this->THIS_LOCK)
- // If docWriter has some aborted files that were
- // never incref'd, then we clean them up here
- if (docWriter != NULL) {
- const std::vector<std::string> *files =
docWriter->abortedFiles();
- if (files != NULL)
- deleter->deleteNewFiles(*files);
+ std::vector<std::string> files;
+ directory->list(files);
+ for (auto& file : files) {
+ directory->deleteFile(file.c_str());
}
}
})
diff --git a/src/core/CLucene/index/SDocumentWriter.cpp
b/src/core/CLucene/index/SDocumentWriter.cpp
index e304999b..a91f67d2 100644
--- a/src/core/CLucene/index/SDocumentWriter.cpp
+++ b/src/core/CLucene/index/SDocumentWriter.cpp
@@ -319,7 +319,7 @@ void
SDocumentsWriter<T>::ThreadState::FieldData::processField(Analyzer *sanalyz
// docFieldsFinal.values[j] = NULL;
}
- } catch (exception &ae) {
+ } catch (CLuceneError& ae) {
throw ae;
}
}
@@ -536,6 +536,12 @@ void
SDocumentsWriter<T>::ThreadState::FieldData::addPosition(Token *token) {
const int32_t textLen1 = 1 + tokenTextLen;
if (textLen1 + threadState->scharPool->tUpto > CHAR_BLOCK_SIZE) {
+ if (textLen1 > CHAR_BLOCK_SIZE) {
+ std::string errmsg = "bytes can be at most " +
+ std::to_string(CHAR_BLOCK_SIZE - 1) +
+ " in length; got " +
std::to_string(tokenTextLen);
+ _CLTHROWA(CL_ERR_MaxBytesLength, errmsg.c_str());
+ }
threadState->scharPool->nextBuffer();
}
T *text = threadState->scharPool->buffer;
@@ -835,7 +841,7 @@ bool SDocumentsWriter<T>::updateDocument(Document *doc,
Analyzer *sanalyzer) {
finishDocument(state);)
success = true;
}
- _CLFINALLY(
+ _CLFINALLY(
if (!success) {
// If this thread state had decided to flush, we
// must clear it so another thread can flush
@@ -844,8 +850,8 @@ bool SDocumentsWriter<T>::updateDocument(Document *doc,
Analyzer *sanalyzer) {
flushPending = false;
}
})
- } catch (exception &ae) {
- abort(nullptr);
+ } catch (CLuceneError& ae) {
+ throw ae;
}
return state->doFlushAfter;
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
