c-sharp 8.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209
  1. * MCS: The Ximian C# compiler
  2. The Mono C# compiler is considered feature complete at this
  3. point and relatively mature. MCS is able to compile itself
  4. and many more C# programs (there is a test suite included that
  5. you can use). It is routinely used to compile Mono, roughly
  6. half a million lines of C# code.
  7. * Pending tasks
  8. There are a few known bugs in the Mono C# compiler, but
  9. they are very very few at this point,
  10. you can also browse the MCS <a href="http://bugzilla.ximian.com/buglist.cgi?product=Mono%2FMCS&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&email1=&emailtype1=substring&emailassigned_to1=1&email2=&emailtype2=substring&emailreporter2=1&changedin=&chfieldfrom=&chfieldto=Now&chfieldvalue=&short_desc=&short_desc_type=substring&long_desc=&long_desc_type=substring&bug_file_loc=&bug_file_loc_type=substring&keywords=&keywords_type=anywords&op_sys_details=&op_sys_details_type=substring&version_details=&version_details_type=substring&cmdtype=doit&newqueryname=&order=Reuse+same+sort+as+last+time&form_name=query">bugs</a> from Bugzilla.
  11. A test suite is maintained to track the progress of
  12. the compiler and various programs are routinely compiled and
  13. ran.
  14. ** Slides
  15. Slides for the Mono C# Compiler presentation at .NET ONE are
  16. available <a
  17. href="http://primates.ximian.com/~miguel/slides-europe-nov-2002/Mono_C_Sharp_Overview_1007.sxi">here</a>
  18. in StarOffice format.
  19. ** Obtaining MCS
  20. The Mono C# compiler is part of the `mcs' module in the Mono CVS
  21. you can get it from our <a href="anoncvs.html">Anonymous CVS</a> server,
  22. or you can get nightly <a href="download.html">download page</a>.
  23. ** Running MCS
  24. MCS is written in C# and uses heavily the .NET APIs. MCS runs
  25. on Linux with the Mono runtime and on Windows with both the
  26. .NET runtime and the Mono runtime.
  27. ** Reporting Bugs in MCS
  28. When you report a bug, try to provide a small test case that would
  29. show the error so we can include this as part of the Mono C# regression
  30. test suite.
  31. If the bug is an error or a warning that we do not flag, write
  32. a sample program called `csXXXX.cs' where XXXX is the code number
  33. that is used by the Microsoft C# compiler that illustrates the
  34. problem. That way we can also do regression tests on the invalid
  35. input.
  36. ** Phases of the compiler
  37. The compiler has a number of phases:
  38. <ul>
  39. * Lexical analyzer: hand-coded lexical analyzer that
  40. provides tokens to the parser.
  41. * The Parser: the parser is implemented using Jay (A
  42. Berkeley Yacc port to Java, that I ported to C#).
  43. The parser does minimal work and syntax checking,
  44. and only constructs a parsed tree.
  45. Each language element gets its own class. The code
  46. convention is to use an uppercase name for the
  47. language element. So a C# class and its associated
  48. information is kept in a "Class" class, a "struct"
  49. in a "Struct" class and so on. Statements derive
  50. from the "Statement" class, and Expressions from the
  51. Expr class.
  52. * Parent class resolution: before the actual code
  53. generation, we need to resolve the parents and
  54. interfaces for interface, classe and struct
  55. definitions.
  56. * Semantic analysis: since C# can not resolve in a
  57. top-down pass what identifiers actually mean, we
  58. have to postpone this decision until the above steps
  59. are finished.
  60. * Code generation: The code generation is done through
  61. the System.Reflection.Emit API.
  62. </ul>
  63. ** CIL Optimizations.
  64. The compiler performs a number of simple optimizations on its input:
  65. constant folding (this is required by the C# language spec) and
  66. can perform dead code elimination.
  67. Other more interesting optimizations like hoisting are not possible
  68. at this point since the compiler output at this point does not
  69. generate an intermediate representation that is suitable to
  70. perform basic block computation.
  71. Adding an intermediate layer to enable the basic block
  72. computation to the compiler should be a simple task, but we
  73. are considering having a generic CIL optimizer. Since all the
  74. information that is required to perform basic block-based
  75. optimizations is available at the CIL level, we might just skip
  76. this step altogether and have just a generic IL optimizer that
  77. would perform hoisting on arbitrary CIL programs, not only
  78. those produced by MCS.
  79. If this tool is further expanded to perform constant folding
  80. (not needed for our C# compiler, as it is already in there)
  81. and dead code elimination, other compiler authors might be
  82. able to use this generic CIL optimizer in their projects
  83. reducing their time to develop a production compiler.
  84. ** History
  85. MCS was able to parse itself on April 2001, MCS compiled itself
  86. for the first time on December 28 2001. MCS became self hosting
  87. on January 3rd, 2002.
  88. The Mono Runtime and the Mono execution engine were able to make
  89. our compiler self hosting on March 12, 2002.
  90. ** Questions and Answers
  91. Q: Why not write a C# front-end for GCC?
  92. A: I wanted to learn about C#, and this was an exercise in this
  93. task. The resulting compiler is highly object-oriented, which has
  94. lead to a very nice, easy to follow and simple implementation of
  95. the compiler.
  96. I found that the design of this compiler is very similar to
  97. Guavac's implementation.
  98. Targeting the CIL/MSIL byte codes would require to re-architecting
  99. GCC, as GCC is mostly designed to be used for register machines.
  100. The GCC Java engine that generates Java byte codes cheats: it does
  101. not use the GCC backend; it has a special backend just for Java, so
  102. you can not really generate Java bytecodes from the other languages
  103. supported by GCC.
  104. Q: If your C# compiler is written in C#, how do you plan on getting
  105. this working on a non-Microsoft environment.
  106. We will do this through an implementation of the CLI Virtual
  107. Execution System for Unix (our JIT engine).
  108. Our JIT engine is working for the purposes of using the compiler.
  109. The supporting class libraries are being worked on to fully support
  110. the compiler.
  111. Q: Do you use Bison?
  112. A: No, currently I am using Jay which is a port of Berkeley Yacc to
  113. Java that I later ported to C#. This means that error recovery is
  114. not as nice as I would like to, and for some reason error
  115. productions are not being caught.
  116. In the future I want to port one of the Bison/Java ports to C# for
  117. the parser.
  118. Q: Should someone work on a GCC front-end to C#?
  119. A: I would love if someone does, and we would love to help anyone that
  120. takes on that task, but we do not have the time or expertise to
  121. build a C# compiler with the GCC engine. I find it a lot more fun
  122. personally to work on C# on a C# compiler, which has an intrinsic
  123. beauty.
  124. We can provide help and assistance to anyone who would like to work
  125. on this task.
  126. Q: Should someone make a GCC backend that will generate CIL images?
  127. A: I would love to see a backend to GCC that generates CIL images. It
  128. would provide a ton of free compilers that would generate CIL
  129. code. This is something that people would want to look into
  130. anyways for Windows interoperation in the future.
  131. Again, we would love to provide help and assistance to anyone
  132. interested in working in such a project.
  133. Q: What about making a front-end to GCC that takes CIL images and
  134. generates native code?
  135. A: I would love to see this, specially since GCC supports this same
  136. feature for Java Byte Codes. You could use the metadata library
  137. from Mono to read the byte codes (ie, this would be your
  138. "front-end") and generate the trees that get passed to the
  139. optimizer.
  140. Ideally our implementation of the CLI will be available as a shared
  141. library that could be linked with your application as its runtime
  142. support.
  143. Again, we would love to provide help and assistance to anyone
  144. interested in working in such a project.
  145. Q: But would this work around the GPL in the GCC compiler and allow
  146. people to work on non-free front-ends?
  147. A: People can already do this by targeting the JVM byte codes (there
  148. are about 130 compilers for various languages that target the JVM).
  149. Q: Why are you writing a JIT engine instead of a front-end to GCC?
  150. A: The JIT engine and runtime engine will be able to execute CIL
  151. executables generated on Windows.
  152. You might also want to look at the <a href="faq.html#gcc">GCC</a>
  153. section on the main FAQ