summaryrefslogtreecommitdiffstats
path: root/src/lib-fts/fts-tokenizer-private.h
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-15 17:36:47 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-15 17:36:47 +0000
commit0441d265f2bb9da249c7abf333f0f771fadb4ab5 (patch)
tree3f3789daa2f6db22da6e55e92bee0062a7d613fe /src/lib-fts/fts-tokenizer-private.h
parentInitial commit. (diff)
downloaddovecot-0441d265f2bb9da249c7abf333f0f771fadb4ab5.tar.xz
dovecot-0441d265f2bb9da249c7abf333f0f771fadb4ab5.zip
Adding upstream version 1:2.3.21+dfsg1.upstream/1%2.3.21+dfsg1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/lib-fts/fts-tokenizer-private.h')
-rw-r--r--src/lib-fts/fts-tokenizer-private.h52
1 files changed, 52 insertions, 0 deletions
diff --git a/src/lib-fts/fts-tokenizer-private.h b/src/lib-fts/fts-tokenizer-private.h
new file mode 100644
index 0000000..b7615b1
--- /dev/null
+++ b/src/lib-fts/fts-tokenizer-private.h
@@ -0,0 +1,52 @@
+#ifndef FTS_TOKENIZER_PRIVATE_H
+#define FTS_TOKENIZER_PRIVATE_H
+
+#include "fts-tokenizer.h"
+
+#define FTS_TOKENIZER_CLASSES_NR 2
+
+struct fts_tokenizer_vfuncs {
+ int (*create)(const char *const *settings,
+ struct fts_tokenizer **tokenizer_r, const char **error_r);
+ void (*destroy)(struct fts_tokenizer *tok);
+
+ void (*reset)(struct fts_tokenizer *tok);
+ int (*next)(struct fts_tokenizer *tok, const unsigned char *data,
+ size_t size, size_t *skip_r, const char **token_r,
+ const char **error_r);
+};
+
+enum fts_tokenizer_parent_state {
+ FTS_TOKENIZER_PARENT_STATE_ADD_DATA = 0,
+ FTS_TOKENIZER_PARENT_STATE_NEXT_OUTPUT,
+ FTS_TOKENIZER_PARENT_STATE_FINALIZE
+};
+
+struct fts_tokenizer {
+ const char *name;
+ const struct fts_tokenizer_vfuncs *v;
+ int refcount;
+
+ struct fts_tokenizer *parent;
+ buffer_t *parent_input;
+ enum fts_tokenizer_parent_state parent_state;
+
+ const unsigned char *prev_data;
+ size_t prev_size;
+ size_t prev_skip;
+ bool prev_reply_finished;
+ bool skip_parents; /* Return token as is, do not hand to parents. */
+ /* Instead of handing child tokens separately to parent tokenizer,
+ treat the returned tokens as a continuous stream. The final token
+ isn't returned until the child tokenizer also sees 0-sized data. */
+ bool stream_to_parents;
+ /* Parent stream still needs to be finalized, so any final pending
+ tokens will be returned. This is used only with
+ stream_to_parents=TRUE. */
+ bool finalize_parent_pending;
+};
+
+void fts_tokenizer_register(const struct fts_tokenizer *tok_class);
+void fts_tokenizer_unregister(const struct fts_tokenizer *tok_class);
+
+#endif