• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Protocol Buffers - Google's data interchange format
2 // Copyright 2008 Google Inc.  All rights reserved.
3 // https://developers.google.com/protocol-buffers/
4 //
5 // Redistribution and use in source and binary forms, with or without
6 // modification, are permitted provided that the following conditions are
7 // met:
8 //
9 //     * Redistributions of source code must retain the above copyright
10 // notice, this list of conditions and the following disclaimer.
11 //     * Redistributions in binary form must reproduce the above
12 // copyright notice, this list of conditions and the following disclaimer
13 // in the documentation and/or other materials provided with the
14 // distribution.
15 //     * Neither the name of Google Inc. nor the names of its
16 // contributors may be used to endorse or promote products derived from
17 // this software without specific prior written permission.
18 //
19 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30 
31 // Author: kenton@google.com (Kenton Varda)
32 //  Based on original Protocol Buffers design by
33 //  Sanjay Ghemawat, Jeff Dean, and others.
34 //
35 // Utility class for writing text to a ZeroCopyOutputStream.
36 
37 #ifndef GOOGLE_PROTOBUF_IO_PRINTER_H__
38 #define GOOGLE_PROTOBUF_IO_PRINTER_H__
39 
40 #include <map>
41 #include <string>
42 #include <vector>
43 #include <google/protobuf/stubs/common.h>
44 
45 #include <google/protobuf/port_def.inc>
46 
47 namespace google {
48 namespace protobuf {
49 namespace io {
50 
51 class ZeroCopyOutputStream;  // zero_copy_stream.h
52 
53 // Records annotations about a Printer's output.
54 class PROTOBUF_EXPORT AnnotationCollector {
55  public:
56   // Annotation is a ofset range and a payload pair.
57   typedef std::pair<std::pair<size_t, size_t>, std::string> Annotation;
58 
59   // Records that the bytes in file_path beginning with begin_offset and ending
60   // before end_offset are associated with the SourceCodeInfo-style path.
61   virtual void AddAnnotation(size_t begin_offset, size_t end_offset,
62                              const std::string& file_path,
63                              const std::vector<int>& path) = 0;
64 
65   // TODO(gerbens) I don't see why we need virtuals here. Just a vector of
66   // range, payload pairs stored in a context should suffice.
AddAnnotationNew(Annotation & a)67   virtual void AddAnnotationNew(Annotation& a) {}
68 
~AnnotationCollector()69   virtual ~AnnotationCollector() {}
70 };
71 
72 // Records annotations about a Printer's output to the given protocol buffer,
73 // assuming that the buffer has an ::Annotation message exposing path,
74 // source_file, begin and end fields.
75 template <typename AnnotationProto>
76 class AnnotationProtoCollector : public AnnotationCollector {
77  public:
78   // annotation_proto is the protocol buffer to which new Annotations should be
79   // added. It is not owned by the AnnotationProtoCollector.
AnnotationProtoCollector(AnnotationProto * annotation_proto)80   explicit AnnotationProtoCollector(AnnotationProto* annotation_proto)
81       : annotation_proto_(annotation_proto) {}
82 
83   // Override for AnnotationCollector::AddAnnotation.
AddAnnotation(size_t begin_offset,size_t end_offset,const std::string & file_path,const std::vector<int> & path)84   virtual void AddAnnotation(size_t begin_offset, size_t end_offset,
85                              const std::string& file_path,
86                              const std::vector<int>& path) {
87     typename AnnotationProto::Annotation* annotation =
88         annotation_proto_->add_annotation();
89     for (int i = 0; i < path.size(); ++i) {
90       annotation->add_path(path[i]);
91     }
92     annotation->set_source_file(file_path);
93     annotation->set_begin(begin_offset);
94     annotation->set_end(end_offset);
95   }
96   // Override for AnnotationCollector::AddAnnotation.
AddAnnotationNew(Annotation & a)97   virtual void AddAnnotationNew(Annotation& a) {
98     auto* annotation = annotation_proto_->add_annotation();
99     annotation->ParseFromString(a.second);
100     annotation->set_begin(a.first.first);
101     annotation->set_end(a.first.second);
102   }
103 
104  private:
105   // The protocol buffer to which new annotations should be added.
106   AnnotationProto* const annotation_proto_;
107 };
108 
109 // This simple utility class assists in code generation.  It basically
110 // allows the caller to define a set of variables and then output some
111 // text with variable substitutions.  Example usage:
112 //
113 //   Printer printer(output, '$');
114 //   map<string, string> vars;
115 //   vars["name"] = "Bob";
116 //   printer.Print(vars, "My name is $name$.");
117 //
118 // The above writes "My name is Bob." to the output stream.
119 //
120 // Printer aggressively enforces correct usage, crashing (with assert failures)
121 // in the case of undefined variables in debug builds. This helps greatly in
122 // debugging code which uses it.
123 //
124 // If a Printer is constructed with an AnnotationCollector, it will provide it
125 // with annotations that connect the Printer's output to paths that can identify
126 // various descriptors.  In the above example, if person_ is a descriptor that
127 // identifies Bob, we can associate the output string "My name is Bob." with
128 // a source path pointing to that descriptor with:
129 //
130 //   printer.Annotate("name", person_);
131 //
132 // The AnnotationCollector will be sent an annotation linking the output range
133 // covering "Bob" to the logical path provided by person_.  Tools may use
134 // this association to (for example) link "Bob" in the output back to the
135 // source file that defined the person_ descriptor identifying Bob.
136 //
137 // Annotate can only examine variables substituted during the last call to
138 // Print.  It is invalid to refer to a variable that was used multiple times
139 // in a single Print call.
140 //
141 // In full generality, one may specify a range of output text using a beginning
142 // substitution variable and an ending variable.  The resulting annotation will
143 // span from the first character of the substituted value for the beginning
144 // variable to the last character of the substituted value for the ending
145 // variable.  For example, the Annotate call above is equivalent to this one:
146 //
147 //   printer.Annotate("name", "name", person_);
148 //
149 // This is useful if multiple variables combine to form a single span of output
150 // that should be annotated with the same source path.  For example:
151 //
152 //   Printer printer(output, '$');
153 //   map<string, string> vars;
154 //   vars["first"] = "Alice";
155 //   vars["last"] = "Smith";
156 //   printer.Print(vars, "My name is $first$ $last$.");
157 //   printer.Annotate("first", "last", person_);
158 //
159 // This code would associate the span covering "Alice Smith" in the output with
160 // the person_ descriptor.
161 //
162 // Note that the beginning variable must come before (or overlap with, in the
163 // case of zero-sized substitution values) the ending variable.
164 //
165 // It is also sometimes useful to use variables with zero-sized values as
166 // markers.  This avoids issues with multiple references to the same variable
167 // and also allows annotation ranges to span literal text from the Print
168 // templates:
169 //
170 //   Printer printer(output, '$');
171 //   map<string, string> vars;
172 //   vars["foo"] = "bar";
173 //   vars["function"] = "call";
174 //   vars["mark"] = "";
175 //   printer.Print(vars, "$function$($foo$,$foo$)$mark$");
176 //   printer.Annotate("function", "mark", call_);
177 //
178 // This code associates the span covering "call(bar,bar)" in the output with the
179 // call_ descriptor.
180 
181 class PROTOBUF_EXPORT Printer {
182  public:
183   // Create a printer that writes text to the given output stream.  Use the
184   // given character as the delimiter for variables.
185   Printer(ZeroCopyOutputStream* output, char variable_delimiter);
186 
187   // Create a printer that writes text to the given output stream.  Use the
188   // given character as the delimiter for variables.  If annotation_collector
189   // is not null, Printer will provide it with annotations about code written
190   // to the stream.  annotation_collector is not owned by Printer.
191   Printer(ZeroCopyOutputStream* output, char variable_delimiter,
192           AnnotationCollector* annotation_collector);
193 
194   ~Printer();
195 
196   // Link a subsitution variable emitted by the last call to Print to the object
197   // described by descriptor.
198   template <typename SomeDescriptor>
Annotate(const char * varname,const SomeDescriptor * descriptor)199   void Annotate(const char* varname, const SomeDescriptor* descriptor) {
200     Annotate(varname, varname, descriptor);
201   }
202 
203   // Link the output range defined by the substitution variables as emitted by
204   // the last call to Print to the object described by descriptor. The range
205   // begins at begin_varname's value and ends after the last character of the
206   // value substituted for end_varname.
207   template <typename SomeDescriptor>
Annotate(const char * begin_varname,const char * end_varname,const SomeDescriptor * descriptor)208   void Annotate(const char* begin_varname, const char* end_varname,
209                 const SomeDescriptor* descriptor) {
210     if (annotation_collector_ == NULL) {
211       // Annotations aren't turned on for this Printer, so don't pay the cost
212       // of building the location path.
213       return;
214     }
215     std::vector<int> path;
216     descriptor->GetLocationPath(&path);
217     Annotate(begin_varname, end_varname, descriptor->file()->name(), path);
218   }
219 
220   // Link a subsitution variable emitted by the last call to Print to the file
221   // with path file_name.
Annotate(const char * varname,const std::string & file_name)222   void Annotate(const char* varname, const std::string& file_name) {
223     Annotate(varname, varname, file_name);
224   }
225 
226   // Link the output range defined by the substitution variables as emitted by
227   // the last call to Print to the file with path file_name. The range begins
228   // at begin_varname's value and ends after the last character of the value
229   // substituted for end_varname.
Annotate(const char * begin_varname,const char * end_varname,const std::string & file_name)230   void Annotate(const char* begin_varname, const char* end_varname,
231                 const std::string& file_name) {
232     if (annotation_collector_ == NULL) {
233       // Annotations aren't turned on for this Printer.
234       return;
235     }
236     std::vector<int> empty_path;
237     Annotate(begin_varname, end_varname, file_name, empty_path);
238   }
239 
240   // Print some text after applying variable substitutions.  If a particular
241   // variable in the text is not defined, this will crash.  Variables to be
242   // substituted are identified by their names surrounded by delimiter
243   // characters (as given to the constructor).  The variable bindings are
244   // defined by the given map.
245   void Print(const std::map<std::string, std::string>& variables,
246              const char* text);
247 
248   // Like the first Print(), except the substitutions are given as parameters.
249   template <typename... Args>
Print(const char * text,const Args &...args)250   void Print(const char* text, const Args&... args) {
251     std::map<std::string, std::string> vars;
252     PrintInternal(&vars, text, args...);
253   }
254 
255   // Indent text by two spaces.  After calling Indent(), two spaces will be
256   // inserted at the beginning of each line of text.  Indent() may be called
257   // multiple times to produce deeper indents.
258   void Indent();
259 
260   // Reduces the current indent level by two spaces, or crashes if the indent
261   // level is zero.
262   void Outdent();
263 
264   // Write a string to the output buffer.
265   // This method does not look for newlines to add indentation.
266   void PrintRaw(const std::string& data);
267 
268   // Write a zero-delimited string to output buffer.
269   // This method does not look for newlines to add indentation.
270   void PrintRaw(const char* data);
271 
272   // Write some bytes to the output buffer.
273   // This method does not look for newlines to add indentation.
274   void WriteRaw(const char* data, int size);
275 
276   // FormatInternal is a helper function not meant to use directly, use
277   // compiler::cpp::Formatter instead. This function is meant to support
278   // formatting text using named variables (eq. "$foo$) from a lookup map (vars)
279   // and variables directly supplied by arguments (eq "$1$" meaning first
280   // argument which is the zero index element of args).
281   void FormatInternal(const std::vector<std::string>& args,
282                       const std::map<std::string, std::string>& vars,
283                       const char* format);
284 
285   // True if any write to the underlying stream failed.  (We don't just
286   // crash in this case because this is an I/O failure, not a programming
287   // error.)
failed()288   bool failed() const { return failed_; }
289 
290  private:
291   // Link the output range defined by the substitution variables as emitted by
292   // the last call to Print to the object found at the SourceCodeInfo-style path
293   // in a file with path file_path. The range begins at the start of
294   // begin_varname's value and ends after the last character of the value
295   // substituted for end_varname. Note that begin_varname and end_varname
296   // may refer to the same variable.
297   void Annotate(const char* begin_varname, const char* end_varname,
298                 const std::string& file_path, const std::vector<int>& path);
299 
300   // Base case
PrintInternal(std::map<std::string,std::string> * vars,const char * text)301   void PrintInternal(std::map<std::string, std::string>* vars,
302                      const char* text) {
303     Print(*vars, text);
304   }
305 
306   template <typename... Args>
PrintInternal(std::map<std::string,std::string> * vars,const char * text,const char * key,const std::string & value,const Args &...args)307   void PrintInternal(std::map<std::string, std::string>* vars, const char* text,
308                      const char* key, const std::string& value,
309                      const Args&... args) {
310     (*vars)[key] = value;
311     PrintInternal(vars, text, args...);
312   }
313 
314   // Copy size worth of bytes from data to buffer_.
315   void CopyToBuffer(const char* data, int size);
316 
push_back(char c)317   void push_back(char c) {
318     if (failed_) return;
319     if (buffer_size_ == 0) {
320       if (!Next()) return;
321     }
322     *buffer_++ = c;
323     buffer_size_--;
324     offset_++;
325   }
326 
327   bool Next();
328 
329   inline void IndentIfAtStart();
330   const char* WriteVariable(
331       const std::vector<std::string>& args,
332       const std::map<std::string, std::string>& vars, const char* format,
333       int* arg_index,
334       std::vector<AnnotationCollector::Annotation>* annotations);
335 
336   const char variable_delimiter_;
337 
338   ZeroCopyOutputStream* const output_;
339   char* buffer_;
340   int buffer_size_;
341   // The current position, in bytes, in the output stream.  This is equivalent
342   // to the total number of bytes that have been written so far.  This value is
343   // used to calculate annotation ranges in the substitutions_ map below.
344   size_t offset_;
345 
346   std::string indent_;
347   bool at_start_of_line_;
348   bool failed_;
349 
350   // A map from variable name to [start, end) offsets in the output buffer.
351   // These refer to the offsets used for a variable after the last call to
352   // Print.  If a variable was used more than once, the entry used in
353   // this map is set to a negative-length span.  For singly-used variables, the
354   // start offset is the beginning of the substitution; the end offset is the
355   // last byte of the substitution plus one (such that (end - start) is the
356   // length of the substituted string).
357   std::map<std::string, std::pair<size_t, size_t> > substitutions_;
358 
359   // Keeps track of the keys in substitutions_ that need to be updated when
360   // indents are inserted. These are keys that refer to the beginning of the
361   // current line.
362   std::vector<std::string> line_start_variables_;
363 
364   // Returns true and sets range to the substitution range in the output for
365   // varname if varname was used once in the last call to Print. If varname
366   // was not used, or if it was used multiple times, returns false (and
367   // fails a debug assertion).
368   bool GetSubstitutionRange(const char* varname,
369                             std::pair<size_t, size_t>* range);
370 
371   // If non-null, annotation_collector_ is used to store annotations about
372   // generated code.
373   AnnotationCollector* const annotation_collector_;
374 
375   GOOGLE_DISALLOW_EVIL_CONSTRUCTORS(Printer);
376 };
377 
378 }  // namespace io
379 }  // namespace protobuf
380 }  // namespace google
381 
382 #include <google/protobuf/port_undef.inc>
383 
384 #endif  // GOOGLE_PROTOBUF_IO_PRINTER_H__
385