summaryrefslogtreecommitdiffstats
path: root/xpcom/ds/IncrementalTokenizer.h
diff options
context:
space:
mode:
authorMatt A. Tobin <mattatobin@localhost.localdomain>2018-02-02 04:16:08 -0500
committerMatt A. Tobin <mattatobin@localhost.localdomain>2018-02-02 04:16:08 -0500
commit5f8de423f190bbb79a62f804151bc24824fa32d8 (patch)
tree10027f336435511475e392454359edea8e25895d /xpcom/ds/IncrementalTokenizer.h
parent49ee0794b5d912db1f95dce6eb52d781dc210db5 (diff)
downloadUXP-5f8de423f190bbb79a62f804151bc24824fa32d8.tar
UXP-5f8de423f190bbb79a62f804151bc24824fa32d8.tar.gz
UXP-5f8de423f190bbb79a62f804151bc24824fa32d8.tar.lz
UXP-5f8de423f190bbb79a62f804151bc24824fa32d8.tar.xz
UXP-5f8de423f190bbb79a62f804151bc24824fa32d8.zip
Add m-esr52 at 52.6.0
Diffstat (limited to 'xpcom/ds/IncrementalTokenizer.h')
-rw-r--r--xpcom/ds/IncrementalTokenizer.h122
1 files changed, 122 insertions, 0 deletions
diff --git a/xpcom/ds/IncrementalTokenizer.h b/xpcom/ds/IncrementalTokenizer.h
new file mode 100644
index 000000000..f93668e63
--- /dev/null
+++ b/xpcom/ds/IncrementalTokenizer.h
@@ -0,0 +1,122 @@
+/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/* vim: set ts=8 sts=2 et sw=2 tw=80: */
+/* This Source Code Form is subject to the terms of the Mozilla Public
+* License, v. 2.0. If a copy of the MPL was not distributed with this
+* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
+
+#ifndef INCREMENTAL_TOKENIZER_H__
+#define INCREMENTAL_TOKENIZER_H__
+
+#include "mozilla/Tokenizer.h"
+
+#include "nsError.h"
+#include <functional>
+
+class nsIInputStream;
+
+namespace mozilla {
+
+class IncrementalTokenizer : public TokenizerBase
+{
+public:
+ /**
+ * The consumer callback. The function is called for every single token
+ * as found in the input. Failure result returned by this callback stops
+ * the tokenization immediately and bubbles to result of Feed/FinishInput.
+ *
+ * Fragment()s of consumed tokens are ensured to remain valid until next call to
+ * Feed/FinishInput and are pointing to a single linear buffer. Hence, those can
+ * be safely used to accumulate the data for processing after Feed/FinishInput
+ * returned.
+ */
+ typedef std::function<nsresult(Token const&, IncrementalTokenizer& i)> Consumer;
+
+ /**
+ * For aWhitespaces and aAdditionalWordChars arguments see TokenizerBase.
+ *
+ * @param aConsumer
+ * A mandatory non-null argument, a function that consumes the tokens as they
+ * come when the tokenizer is fed.
+ * @param aRawMinBuffered
+ * When we have buffered at least aRawMinBuffered data, but there was no custom
+ * token found so far because of too small incremental feed chunks, deliver
+ * the raw data to preserve streaming and to save memory. This only has effect
+ * in OnlyCustomTokenizing mode.
+ */
+ explicit IncrementalTokenizer(Consumer aConsumer,
+ const char* aWhitespaces = nullptr,
+ const char* aAdditionalWordChars = nullptr,
+ uint32_t aRawMinBuffered = 1024);
+
+ /**
+ * Pushes the input to be tokenized. These directly call the Consumer callback
+ * on every found token. Result of the Consumer callback is returned here.
+ *
+ * The tokenizer must be initialized with a valid consumer prior call to these
+ * methods. It's not allowed to call Feed/FinishInput from inside the Consumer
+ * callback.
+ */
+ nsresult FeedInput(const nsACString& aInput);
+ nsresult FeedInput(nsIInputStream* aInput, uint32_t aCount);
+ nsresult FinishInput();
+
+ /**
+ * Can only be called from inside the consumer callback.
+ *
+ * When there is still anything to read from the input, tokenize it, store
+ * the token type and value to aToken result and shift the cursor past this
+ * just parsed token. Each call to Next() reads another token from
+ * the input and shifts the cursor.
+ *
+ * Returns false if there is not enough data to deterministically recognize
+ * tokens or when the last returned token was EOF.
+ */
+ MOZ_MUST_USE
+ bool Next(Token& aToken);
+
+ /**
+ * Can only be called from inside the consumer callback.
+ *
+ * Tells the tokenizer to revert the cursor and stop the async parsing until
+ * next feed of the input. This is useful when more than one token is needed
+ * to decide on the syntax but there is not enough input to get a next token
+ * (Next() returned false.)
+ */
+ void NeedMoreInput();
+
+ /**
+ * Can only be called from inside the consumer callback.
+ *
+ * This makes the consumer callback be called again while parsing
+ * the input at the previous cursor position again. This is useful when
+ * the tokenizer state (custom tokens, tokenization mode) has changed and
+ * we want to re-parse the input again.
+ */
+ void Rollback();
+
+private:
+ // Loops over the input with TokenizerBase::Parse and calls the Consumer callback.
+ nsresult Process();
+
+#ifdef DEBUG
+ // True when inside the consumer callback, used only for assertions.
+ bool mConsuming;
+#endif // DEBUG
+ // Modifyable only from the Consumer callback, tells the parser to break, rollback
+ // and wait for more input.
+ bool mNeedMoreInput;
+ // Modifyable only from the Consumer callback, tells the parser to rollback and
+ // parse the input again, with (if modified) new settings of the tokenizer.
+ bool mRollback;
+ // The input buffer. Updated with each call to Feed/FinishInput.
+ nsCString mInput;
+ // Numerical index pointing at the current cursor position. We don't keep direct
+ // reference to the string buffer since the buffer gets often reallocated.
+ nsCString::index_type mInputCursor;
+ // Refernce to the consumer function.
+ Consumer mConsumer;
+};
+
+} // mozilla
+
+#endif