CHANGES 11 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208
  1. Change Log
  2. ----------
  3. Version 1.39
  4. Drew Yao fixed buffer overflow problems in mk_sharp_const.
  5. Version 1.38
  6. Interim release until the rewrite, mostly incorporating modifications from
  7. Kevin Cozens. Small addition for Cygwin in the makefile, and modifications
  8. by Andrew Guenther for Apple platforms.
  9. Version 1.37
  10. Joe Buehler submitted reserve_cells.
  11. Version 1.36
  12. Joe Buehler fixed a patch in the allocator.
  13. Alexander Shendi moved the comment handling in the scanner, which
  14. fixed an obscure bug for which Mike E had provided a patch as well.
  15. Kevin Cozens has submitted some fixes and modifications which have not
  16. been incorporated yet in their entirety.
  17. Version 1.35
  18. Todd Showalter discovered that the number of free cells reported
  19. after GC was incorrect, which could also cause unnecessary allocations.
  20. Version 1.34
  21. Long missing version. Lots of bugfixes have accumulated in my email, so
  22. I had to start using them. In this version, Keenan Pepper has submitted
  23. a bugfix for the string comparison library procedure, Wouter Boeke
  24. modified some code that was casting to the wrong type and crashed on
  25. some machines, "SheppardCo" submitted a replacement "modulo" code and
  26. Scott Fenton submitted lots of corrections that shut up some compiler
  27. warnings. Brian Maher submitted instructions on how to build on OS-X.
  28. I have to dig deeper into my mailbox and find earlier emails, too.
  29. Version 1.33
  30. Charles Hayden fixed a nasty GC bug of the new stack frame, while in
  31. the process of porting TinyScheme to C++. He also submitted other
  32. changes, and other people also had comments or requests, but the GC
  33. bug was so important that this version is put through the door to
  34. correct it.
  35. Version 1.32
  36. Stephen Gildea put some quality time on TinyScheme again, and made
  37. a whole lot of changes to the interpreter that made it noticeably
  38. faster.
  39. Version 1.31
  40. Patches to the hastily-done version 1.30. Stephen Gildea fixed
  41. some things done wrongly, and Richard Russo fixed the makefile
  42. for building on Windows. Property lists (heritage from MiniScheme)
  43. are now optional and have dissappeared from the interface. They
  44. should be considered as deprecated.
  45. Version 1.30
  46. After many months, I followed Preston Bannister's advice of
  47. using macros and a single source text to keep the enums and the
  48. dispatch table in sync, and I used his contributed "opdefines.h".
  49. Timothy Downs contributed a helpful function, "scheme_call".
  50. Stephen Gildea contributed new versions of the makefile and
  51. practically all other sources. He created a built-in STRING-APPEND,
  52. and fixed a lot of other bugs.
  53. Ruhi Bloodworth reported fixes necessary for OS X and a small
  54. bug in dynload.c.
  55. Version 1.29
  56. The previous version contained a lot of corrections, but there
  57. were a lot more that still wait on a sheet of paper lost in a
  58. carton someplace after my house move... Manuel Heras-Gilsanz
  59. noticed this and resent his own contribution, which relies on
  60. another bugfix that v.1.28 was missing: a problem with string
  61. output, that this version fixes. I hope other people will take
  62. the time to resend their contributions, if they didn't make it
  63. to v.1.28.
  64. Version 1.28
  65. Many people have contacted me with bugfixes or remarks in
  66. the three months I was inactive. A lot of them spotted that
  67. scheme_deinit crashed while reporting gc results. They suggested
  68. that sc->outport be set to NIL in scheme_deinit, which I did.
  69. Dennis Taylor remarked that OP_VALUEPRINT reset sc->value instead
  70. of preserving it. He submitted a modification which I adopted
  71. partially. David Hovemeyer sent me many little changes, that you
  72. will find in version 1.28, and Partice Stoessel modified the
  73. float reader to conform to R5RS.
  74. Version 1.27
  75. Version 1.27 is the successor of 1.25. Bug fixes only, but I had to
  76. release them so that everybody can profit. 'Backchar' tried to write
  77. back to the string, which obviously didn't work for const strings.
  78. 'Substring' didn't check for crossed start and end indices. Defines
  79. changed to restore the ability to compile under MSVC.
  80. Version 1.26
  81. Version 1.26 was never released. I changed a lot of things, in fact
  82. too much, even the garbage collector, and hell broke loose. I'll
  83. try a more gradual approach next time.
  84. Version 1.25
  85. Types have been homogenized to be able to accomodate a different
  86. representation. Plus, promises are no longer closures. Unfortunately,
  87. I discovered that continuations and force/delay do not pass the SCM
  88. test (and never did)... However, on the bright side, what little
  89. modifications I did had a large impact on the footprint:
  90. USE_NO_FEATURES now produces an object file of 63960 bytes on Linux!
  91. Version 1.24
  92. SCM tests now pass again after change in atom2str.
  93. Version 1.23
  94. Finally I managed to mess it up with my version control. Version
  95. 1.22 actually lacked some of the things I have been fixing in the
  96. meantime. This should be considered as a complete replacement for
  97. 1.22.
  98. Version 1.22
  99. The new ports had a bug in LOAD. MK_CLOSURE is introduced.
  100. Shawn Wagner inquired about string->number and number->string.
  101. I added string->atom and atom->string and defined the number
  102. functions from them. Doing that, I fixed WRITE applied to symbols
  103. (it didn't quote them). Unfortunately, minimum build is now
  104. slightly larger than 64k... I postpone action because Jason's idea
  105. might solve it elegantly.
  106. Version 1.21
  107. Jason Felice submitted a radically different datatype representation
  108. which he had implemented. While discussing its pros and cons, it
  109. became apparent that the current implementation of ports suffered
  110. from a grave fault: ports were not garbage-collected. I changed the
  111. ports to be heap-allocated, which enabled the use of string ports
  112. for loading. Jason also fixed errors in the garbage collection of
  113. vectors. USE_VERBATIM is gone. "ssp_compiler.c" has a better solution
  114. on HTML generation. A bug involving backslash notation in strings
  115. has been fixed. '-c' flag now executes next argument as a stream of
  116. Scheme commands. Foreign functions are now also heap allocated,
  117. and scheme_define is used to define everything.
  118. Version 1.20
  119. Tracing has been added. The toplevel loop has been slightly
  120. rearranged. Backquote reading for vector templates has been
  121. sanitized. Symbol interning is now correct. Arithmetic functions
  122. have been corrected. APPLY, MAP, FOR-EACH, numeric comparison
  123. functions fixed. String reader/writer understands \xAA notation.
  124. Version 1.19
  125. Carriage Return now delimits identifiers. DOS-formatted Scheme files
  126. can be used by Unix. Random number generator added to library.
  127. Fixed some glitches of the new type-checking scheme. Fixed erroneous
  128. (append '() 'a) behavior. Will continue with r4rstest.scm to
  129. fix errors.
  130. Version 1.18
  131. The FFI has been extended. USE_VERBOSE_GC has gone. Anyone wanting
  132. the same functionality can put (gcverbose #t) in init.scm.
  133. print-width was removed, along with three corresponding op-codes.
  134. Extended character constants with ASCII names were added.
  135. mk_counted_string paves the way for full support of binary strings.
  136. As much as possible of the type-checking chores were delegated
  137. to the inner loop, thus reducing the code size to less than 4200 loc!
  138. Version 1.17
  139. Dynamically-loaded extensions are more fully integrated.
  140. TinyScheme is now distributed under the BSD open-source license.
  141. Version 1.16
  142. Dynamically-loaded extensions introduced (USE_DL).
  143. Santeri Paavolainen found a race condition: When a cons is executed,
  144. and each of the two arguments is a constructing function, GC could
  145. happen before all arguments are evaluated and cons() is called, and
  146. the evaluated arguments would all be reclaimed!
  147. Fortunately, such a case was rare in the code, although it is
  148. a pitfall in new code and code in foreign functions. Currently, only
  149. one such case remains, when COLON_HOOK is defined.
  150. Version 1.15
  151. David Gould also contributed some changes that speed up operation.
  152. Kirk Zurell fixed HASPROP.
  153. The Garbage Collection didn't collect all the garbage...fixed.
  154. Version 1.14
  155. Unfortunately, after Andre fixed the GC it became obvious that the
  156. algorithm was too slow... Fortunately, David Gould found a way to
  157. speed it up.
  158. Version 1.13
  159. Silly bug involving division by zero resolved by Roland Kaufman.
  160. Macintoch support from Shmulik Regev.
  161. Float parser bug fixed by Alexander Shendi.
  162. GC bug from Andru Luvisi.
  163. Version 1.12
  164. Cis* incorrectly called isalpha() instead of isascii()
  165. Added USE_CHAR_CLASSIFIERS, USE_STRING_PORTS.
  166. Version 1.11
  167. BSDI defines isnumber... changed all similar functions to is_*
  168. EXPT now has correct definition. Added FLOOR,CEILING,TRUNCATE
  169. and ROUND, courtesy of Bengt Kleberg. Preprocessor symbols now
  170. have values 1 or 0, and can be set as compiler defines (proposed
  171. by Andy Ganor *months* ago). 'prompt' and 'InitFile' can now be
  172. defined during compilation, too.
  173. Version 1.10
  174. Another bug when file ends with comment!
  175. Added DEFINE-MACRO in init.scm, courtesy of Andy Gaynor.
  176. Version 1.09
  177. Removed bug when READ met EOF. lcm.
  178. Version 1.08
  179. quotient,remainder and modulo. gcd.
  180. Version 1.07
  181. '=>' in cond now exists
  182. list? now checks for circularity
  183. some reader bugs removed
  184. Reader is more consistent wrt vectors
  185. Quote and Quasiquote work with vectors
  186. Version 1.06
  187. #! is now skipped
  188. generic-assoc bug removed
  189. strings are now managed differently, hack.txt is removed
  190. various delicate points fixed
  191. Version 1.05
  192. Support for scripts, *args*, "-1" option.
  193. Various R5RS procedures.
  194. *sharp-hook*
  195. Handles unmatched parentheses.
  196. New architecture for procedures.
  197. Version 1.04
  198. Added missing T_ATOM bits...
  199. Added vectors
  200. Free-list is sorted by address, since vectors need consecutive cells.
  201. (quit <exitcode>) for use with scripts
  202. Version 1.03 (26 Aug 1998):
  203. Extended .h with useful functions for FFI
  204. Library: with-input-* etc.
  205. Finished R5RS I/O, added string ports.
  206. Version 1.02 (25 Aug 1998):
  207. First part of R5RS I/O.