history.hss 46 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940
  1. [Main]
  2. Title=History of A68k
  3. [Top]
  4. <B>Note:</B> This section has been added to the documentation by the TIGCC Team. It was in <I>history.txt</I> before.
  5. <BR><BR>
  6. <B>Version 2.71.F3w (Kevin Kofler, July 27, 2006)</B>
  7. <UL><LI>Bugfix: Fixed compilation on *BSD and recent versions of OS X.</LI></UL>
  8. <B>Version 2.71.F3v (Kevin Kofler, July 31, 2005)</B>
  9. <UL><LI>Bugfix: Don't crash on EQUs doing arithmetic with other EQUs which do
  10. arithmetic with undefined symbols, print an error instead.</LI></UL>
  11. <B>Version 2.71.F3u (Kevin Kofler, February 2, 2005)</B>
  12. <UL><LI>Bugfix: Fixed compilation with GCC 3.4.</LI></UL>
  13. <B>Version 2.71.F3t (Kevin Kofler, January 31, 2005)</B>
  14. <UL><LI>Added: NO_UNOPTIMIZABLE_RELOCS compile-time define (disabled by default)
  15. to allow generating AmigaOS-compatible object files again.</LI>
  16. <LI>Improved: Branches explicitly coded as .w are now marked unoptimizable.</LI>
  17. <LI>Improved: Label references explicitly coded as .l are now marked
  18. unoptimizable.</LI>
  19. <LI>Bugfix: Emitting relocations for undefined symbols which are already in
  20. the symbol table due to the forward branch optimization code now
  21. works properly.</LI></UL>
  22. <B>Version 2.71.F3s (Kevin Kofler, September 21, 2004)</B>
  23. <UL><LI>Added: Support for the "unoptimizable reloc" flag, keeping the linker
  24. from corrupting some instructions when optimizing.</LI>
  25. <LI>Improved: All platforms now use the same maximum line length and chunk
  26. sizes.</LI>
  27. <LI>Improved: The line count limit has been raised from 32767 to INT_MAX.</LI></UL>
  28. <B>Version 2.71.F3r (Kevin Kofler, July 19, 2004)</B>
  29. <UL><LI>Bugfix: Fixed label(PC,Rn) addressing when "label" was in a different
  30. section (again) (reported by Julien Moutinho).</LI></UL>
  31. <B>Version 2.71.F3q (Kevin Kofler, December 30, 2003)</B>
  32. <UL><LI>Bugfix: Fixed label1-label2(an) in all-relocs mode (reported by
  33. hwti).</LI>
  34. <LI>Added: (In all-relocs mode,) don't emit an address difference when the 2
  35. labels are actually at the same address (or simply the same). This allows to
  36. keep optimizing label1-label1(an) into (an).</LI>
  37. <LI>Bugfix: Fixed a bug when range-checking branches in all-relocs mode
  38. (reported by hwti).</LI>
  39. <LI>Bugfix: Fixed a bug when range-checking short branches in all-relocs mode or
  40. to a different section (reported by hwti).</LI></UL>
  41. <B>Version 2.71.F3p (Kevin Kofler, December 1, 2003)</B>
  42. <UL><LI>Bugfix: Fixed label(PC,Rn) addressing in all-relocs mode or when "label" was
  43. in a different section (reported by Matthieu Gallet).</LI></UL>
  44. <B>Version 2.71.F3o (Kevin Kofler, October 5, 2003)</B>
  45. <UL><LI>Bugfix: Range checking now works correctly in all-relocs mode. (No more
  46. spurious out-of-range errors.)</LI></UL>
  47. <B>Version 2.71.F3n (Kevin Kofler, September 28, 2003)</B>
  48. <UL><LI>Bugfix: Local labels now work in all-relocs mode.</LI></UL>
  49. <B>Version 2.71.F3m (Kevin Kofler, September 2, 2003)</B>
  50. <UL><LI>Bugfix: PC-relative or indexed references to other sections didn't work.</LI>
  51. <LI>Bugfix: x(PC,Dn) or x(PC,An) didn't work in all-relocs mode.</LI></UL>
  52. <B>Version 2.71.F3l (Kevin Kofler, July 15, 2003)</B>
  53. <UL><LI>Added: The -u switch now also allows unaligned code, which can be useful
  54. when the code will be copied somewhere else anyway.</LI>
  55. <LI>Added: Short branches across hunks or object files are now allowed.</LI>
  56. <LI>Added: -a (All relocs) switch, which tells A68k to emit all relocs, even
  57. PC-relative relocs within a section. It also emits address
  58. differences in a special TIGCC-specific format. This will allow
  59. more aggressive linker-side optimization.</LI></UL>
  60. <B>Version 2.71.F3k (Kevin Kofler, May 17, 2003)</B>
  61. <UL><LI>Bugfix: Increased the forward reference buffer in order to allow
  62. multiple labels at the same location, as used in PedroM.</LI>
  63. <LI>Added: -u (Unaligned) switch, which disables automatic alignment of
  64. DC.W, DC.L, DCB.W, DCB.L, DS.W and DS.L.</LI></UL><BR>
  65. <B>Version 2.71.F3i (Kevin Kofler, February 1, 2002)</B>
  66. <UL><LI>Bugfix: Local labels caused a &quot;Pass 1 / Pass 2 phase error&quot;
  67. when preceded by an include file which
  68. contained labels, but was skipped in pass 2 (such as OS.h).</LI></UL><BR>
  69. <B>Version 2.71.F3h (Kevin Kofler, January 23, 2002)</B>
  70. <UL><LI>Bugfix: The TITLE (or TTL) directive did not work at all.
  71. It now works as it is supposed to.</LI>
  72. <LI>Bugfix: No longer crashes when trying to generate a
  73. listing file with cross-references (using the -x
  74. switch).</LI></UL><BR>
  75. <B>Version 2.71.F3g (Kevin Kofler, January 22, 2002)</B>
  76. <UL><LI>Bugfix: The RORG and PUBLIC directives are now recognized
  77. correctly even if the first letter is not
  78. capitalized. Thanks to Brandon Sterner for
  79. reporting the problem with the RORG directive.</LI></UL><BR>
  80. <B>Version 2.71.F3f (Kevin Kofler, January 8, 2002)</B>
  81. <UL><LI>Now also accepts &quot;;&quot;, and for the -v switch also &quot;=&quot;, as
  82. separator in command line arguments (wherever &quot;,&quot; is
  83. allowed), for compatibility with the &quot;-WA,&quot; switch in
  84. <CODE>tigcc</CODE>.</LI></UL><BR>
  85. <B>Version 2.71.F3e (Kevin Kofler, August 3, 2001)</B>
  86. <UL><LI>Bugfix: The SUB-&gt;LEA optimisation did not change the
  87. sign of the immediate value.</LI></UL><BR>
  88. <B>Version 2.71.F3d (Kevin Kofler, July 28, 2001)</B>
  89. <UL><LI>Bugfix: The ADD(A)/SUB(A)-&gt;LEA and LEA-&gt;ADDQ/SUBQ
  90. optimizations caused &quot;Pass 1 / Pass 2 phase
  91. errors&quot; if an optimization was known to be
  92. possible only in pass 2 (i.e. if using forward
  93. references to labels in the instruction). A68k
  94. will not try to optimize those anymore. Maybe a
  95. future version will provide support for this kind
  96. of optimization, but it is very difficult to
  97. implement.</LI>
  98. <LI>Bugfix: A68k did not compile correctly on Visual C++
  99. because I had done a define in an incorrect way.</LI></UL><BR>
  100. <B>Version 2.71.F3c (Kevin Kofler, July 20, 2001)</B>
  101. <UL><LI>Applied a few patches submitted by Paul Froissart:
  102. <UL><LI>A few more adaptations for compiling with Microsoft
  103. Visual C++. (Note that I still recommend using GCC for
  104. Mingw32.)</LI>
  105. <LI>A new command line switch:
  106. <DL><DT>-v<DD>Allows to set a variable in the command line (like
  107. -d with GCC). The variable will be a SET, not EQU
  108. variable. Syntax: &quot;-v&lt;name&gt;[,&lt;value&gt;]&quot; (without
  109. spaces, and without the quotes). Note that &lt;value&gt;
  110. can only be a NUMBER at the moment. (We might
  111. support symbols as values in a future version.)
  112. The default value of &lt;value&gt; is 1.</DL></LI>
  113. <LI>Improved the optimization:
  114. <UL><LI>made optional the optimization which changes MOVEM
  115. with a single register to MOVE</LI>
  116. <LI>added an optional optimization which changes ADD(A)
  117. and SUB(A) to LEA when applicable</LI>
  118. <LI>added an optional optimization which changes LEA to
  119. ADDQ or SUBQ when applicable</LI>
  120. <LI>added a command line switch which allows you to
  121. selectively disable some optimizations:
  122. <DL><DT>-r<DD>Allows to disable specific optimizations:
  123. <TABLE CLASS="NOBORDER"><TR><TD CLASS="NOBORDER">-rm</TD><TD CLASS="NOBORDER">Disable the MOVEM -&gt; MOVE optimization</TD></TR>
  124. <TR><TD CLASS="NOBORDER">-ra</TD><TD CLASS="NOBORDER">Disable the ADD(A)/SUB(A) -&gt; LEA optimization</TD></TR>
  125. <TR><TD CLASS="NOBORDER">-rl</TD><TD CLASS="NOBORDER">Disable the LEA -&gt; ADDQ/SUBQ optimization</TD></TR></TABLE>
  126. You may use more than one -r switch (as in
  127. &quot;-rm -ra&quot;) or combine them into a single switch
  128. (as in &quot;-rma&quot;).
  129. You can still use -n to disable ALL optimizations
  130. (even those not covered by -r).</DL></LI></UL></LI>
  131. <LI>Added 2 more synonyms:
  132. <UL><LI>SLO = SCS</LI>
  133. <LI>SHS = SCC</LI></UL></LI>
  134. <LI>The END directive is now optional. If you do not use
  135. it, the assembly will automatically end at the end of
  136. your source file.</LI></UL></LI>
  137. <LI>Also fixed a bug found by Paul Froissart involving the
  138. reporting of invalid flags with the SECTION directive.
  139. (Not that anyone would use those flags anyway.)</LI>
  140. <LI>Adapted and applied a very old patch by John Antonishek,
  141. which allows you to define a symbol as XDEF even if it
  142. has been declared as XREF before. This is especially
  143. useful when using a common include file for globals.</LI>
  144. <LI>Moreover, END is not allowed in macros or include files
  145. anymore since that was causing a few bugs because A68k
  146. simply did not expect anyone to use it there. The
  147. quickest remedy is to simply remove the END instruction,
  148. since it is now optional anyway.</LI></UL><BR>
  149. <B>Version 2.71.F3b (Kevin Kofler, July 12, 2001)</B>
  150. <UL><LI>Bugfix: Now raises an error if an undefined local label is
  151. used.</LI>
  152. <LI>Bugfix: Branches coded with an explicit length (.W or .L)
  153. are not optimized (or flagged as optimizable)
  154. anymore.<BR>
  155. (In previous versions, .B or .S was not optimized,
  156. but .W or .L could be optimized to a short branch.
  157. Now, only branches without an explicit length will
  158. be optimized.)</LI>
  159. <LI>Bugfix: Quoted strings can now be passed as macro
  160. arguments without the need for '&lt;' and '&gt;', even
  161. if they contain spaces or separators (',' or ';').
  162. You can still use '&lt;' and '&gt;', for example in
  163. order to pass strings like &lt;129,' version'&gt; (which
  164. will appear as &quot;ß version&quot; on a TI-89/92/92+).</LI></UL><BR>
  165. <B>Version 2.71.F3a (Kevin Kofler, July 9, 2001)</B>
  166. <UL><LI>Win32 port for the TIGCC project, compiled with GCC for
  167. Mingw32.<BR>
  168. Even though we recommend Mingw32, A68k should also compile
  169. with Microsoft Visual C++, thanks to Paul Froissart.
  170. (Please do NOT ask me questions about compiling with
  171. Visual C++. I have just put in Paul Froissart's ifdefs.)
  172. It also compiles with Cygwin, but since the result is
  173. larger than the Mingw32 version, I do not see any benefit
  174. in using it.</LI>
  175. <LI>Bugfix: My Win32 port of v.2.71.F3 did not support labels
  176. or variable names which contain accented
  177. characters.</LI>
  178. <LI>Bugfix: Now detects all errors in include files. All
  179. include files will be treated in pass 2 if an
  180. error is detected.</LI>
  181. <LI>Bugfix by Paul Froissart: fixed a random crash bug due to
  182. a missing initialization</LI>
  183. <LI>Bugfix by Julien Muchembled: allow odd displacements for
  184. words and longwords if the
  185. resulting address might be
  186. even</LI>
  187. <LI>The Win32 version is now less greedy with the memory, so
  188. its limitations are less annoying:
  189. <UL><LI>lines can now be 256 bytes long (Paul Froissart)</LI>
  190. <LI>filenames (with path) can now be 259 bytes long (the
  191. Win32 maximum)</LI>
  192. <LI>the default heap sizes have also been increased (Paul
  193. Froissart)</LI></UL></LI>
  194. <LI>Added the following synonyms:
  195. <UL><LI>BLO = BCS</LI>
  196. <LI>BHS = BCC</LI>
  197. <LI>DBLO = DBCS</LI>
  198. <LI>DBHS = DBCC</LI>
  199. <LI>ROLX = ROXL</LI>
  200. <LI>RORX = ROXR</LI></UL></LI></UL><BR>
  201. <B>Version 2.71.F3 (David Ellsworth, July 11, 2000)</B>
  202. <UL><LI>Fixed a bug introduced by my modifications that prevented
  203. branches from being optimized in Global XRef mode.</LI></UL><BR>
  204. <B>Version 2.71.F2 (David Ellsworth, January 22, 1998)</B>
  205. <UL><LI>Made A68k properly differentiate between PC-relative and
  206. absolute references to external symbols. Allowed branches
  207. to reference external labels in Global XRef (-g parameter)
  208. mode.</LI></UL><BR>
  209. <B>Version 2.71.F1 (David Ellsworth, January 7, 1998)</B>
  210. <UL><LI>Made changes so that combinations of EQU, XREF, and XDEF
  211. directives work better together. These changes were made in
  212. order for Fargo II library linking to be easier to do.</LI>
  213. <LI>The F1 stands for Fargo patchlevel 1.</LI></UL><BR>
  214. <B>Version 2.71 (Charlie Gibbs, April 16, 1991)</B>
  215. <BR><BR>
  216. The following bugs in version 2.62 have been corrected:
  217. <UL><LI>Enforcer checks were being generated if command-line
  218. parameters that took a numeric value had no value or
  219. the value was invalid. CalcValue was attempting to
  220. look up a label in the symbol table, which having not
  221. been allocated yet was causing a null pointer to be
  222. dereferenced. (Patrick Quaid)</LI></UL>
  223. The following enhancements have been added:
  224. <UL><LI>Function prototyping has been added. All function
  225. prototypes are in a new include file, protos.h, which
  226. is included by A68kdef.h. Prototyping can be disabled
  227. (for compilers which do not support it) by defining
  228. the symbol __NOPROTO; in this case old-style function
  229. declarations are generated instead.</LI></UL><BR>
  230. <B>Version 2.70 (Charlie Gibbs, February 25, 1991)</B>
  231. <BR><BR>
  232. The following bugs in version 2.62 have been corrected:
  233. <UL><LI>The definition of tempstr in WriteSymTab (24 bytes)
  234. was being overrun in some cases, causing A68k to hang.
  235. Its length has been increased to MAXLINE.
  236. (Paul Gittings, John Antonishek)</LI>
  237. <LI>If A7 was in a list of registers in the source operand
  238. of a MOVEM instruction, all registers would be moved
  239. (i.e. the mask was set to 0xFFFF). (Risto Kaivola)</LI>
  240. <LI>Octal or binary values that set bit 31 were being
  241. flagged as overflow errors. (Harvey Taylor)</LI>
  242. <LI>ORGs were unnecessarily restricted when the -s flag was
  243. specified. Everything should be absolute when S-records
  244. are generated, and any absolute ORG should be allowed.</LI></UL>
  245. The following enhancements have been added:
  246. <UL><LI>A new INCBIN directive has been added. It takes a
  247. single operand, a file name whose contents are
  248. included as is at the current position in the
  249. object code file. (Julian Gold, Colin Fox)</LI>
  250. <LI>The opcode TTL is now accepted as a synonym for TITLE.</LI>
  251. <LI>The file mode in the creat() call in xopen() has been
  252. changed from 1 to 0644; this provides a reasonable file
  253. mode when compiled on a Unix system. (Paul Gittings)</LI>
  254. <LI>A new command-line keyword (-g) has been added. It
  255. causes all undefined symbols to be treated as XREF.
  256. (Paul Gittings, Steve Hawtin (who provided the code))</LI>
  257. <LI>All initialized fields in opcodes.c have been made global
  258. for compatibilty with more compilers. (Steve Hawtin)</LI>
  259. <LI>The register list in a MOVEM instruction can now be an
  260. immediate operand which specifies the actual mask bits.
  261. (Paul Gittings, who provided the code)</LI>
  262. <LI>The default value for the -q option has been changed
  263. from 10 to 100.</LI></UL><BR>
  264. <B>Version 2.62 (Charlie Gibbs, March 19, 1990)</B>
  265. <BR><BR>
  266. The following enhancements have been added:
  267. <UL><LI>A new command-line keyword (-m) has been added.
  268. It must be immediately followed by a number which
  269. specifies the offset from the beginning of a small
  270. data section to the base register specified in the
  271. NEAR directive (defaulting to A4). If this parameter
  272. is omitted, the offset defaults to 32768 bytes. This
  273. parameter is meaningful only if the NEAR directive
  274. is used. (Colin Fox)</LI></UL><BR>
  275. <B>Version 2.61 (Charlie Gibbs, January 11, 1990)</B>
  276. <BR><BR>
  277. The following bugs in version 2.6 have been corrected:
  278. <UL><LI>ORG and RORG at the beginning of the program were
  279. being processed incorrectly. (Jukka Jarvinen)</LI>
  280. <LI>A branch instruction to its own label, e.g.
  281. <PRE>lab bra lab</PRE>
  282. would cause phase errors; the instruction was
  283. being shortened on pass 2 but not on pass 1.
  284. (Kevin Hoare)</LI></UL>
  285. The following enhancements have been added:
  286. <UL><LI>Labels may begin with '@' if the next character
  287. is not numeric (to avoid confusion with octal
  288. constants). (Colin Fox)</LI>
  289. <LI>Write errors now cause A68k to terminate gracefully
  290. with an appropriate error message.</LI></UL><BR>
  291. <B>Version 2.6 (Charlie Gibbs, November 2, 1989)</B>
  292. <BR><BR>
  293. The following bugs in version 2.5 have been corrected:
  294. <UL><LI>If a space is left between a file keyword and the file name
  295. (e.g. A68k -l foo.lst foo.asm) the source file was scratched.
  296. (E. Lenz)</LI>
  297. <LI>All code using post-increment addressing in references to
  298. toupper() has been reworked to avoid post-increment. Such
  299. code does not work correctly if toupper() is a macro.
  300. (John K. Antonishek)</LI>
  301. <LI>The spelling of the include files A68kdef.h and A68kglb.h
  302. has been made correct as to case. This simplifies porting
  303. to case-sensitive file systems. (John K. Antonishek)</LI>
  304. <LI>If comments immediately follow the operands of an XDEF,
  305. XREF, or PUBLIC statement with no intervening white space
  306. (as in any of the following statements), A68k would hang:
  307. <PRE> XDEF foo;comments
  308. XREF bar;comments
  309. PUBLIC blah;comments</PRE>
  310. (Bruce Dawson)</LI>
  311. <LI>Numeric values (any radix) which do not fit into 32 bits
  312. were not being flagged. (E. Lenz)</LI>
  313. <LI>DC statements with no operands were not being flagged.
  314. (E. Lenz)</LI>
  315. <LI>The -f option was suggesting short branches where the
  316. displacement would be zero, which is illegal. (E. Lenz)</LI>
  317. <LI>Branches outside the current section had the offset
  318. set to zero. (Matt Dillon, who provided a fix)</LI>
  319. <LI>ADDI, ANDI, CMPI, EORI, ORI, and SUBI instructions whose
  320. source operand was not immediate were not being flagged.
  321. (E. Lenz)</LI>
  322. <LI>Unary NOT of a byte or word immediate operand whose value
  323. was negative was being flagged as a size error.
  324. (John Aycock)</LI>
  325. <LI>All forward branches were rejected when the NEAR
  326. directive was active. (Colin Fox)</LI></UL>
  327. The following enhancements have been added:
  328. <UL><LI>Single-byte immediate operands (e.g. MOVE.B #-1,(a0))
  329. are now padded with a high-order byte of zero, rather
  330. than being sign-extended. (E. Lenz)</LI>
  331. <LI>The message &quot;Error in operand format.&quot; has been changed
  332. to &quot;Addressing mode not allowed here.&quot; in places where
  333. the latter message is more appropriate.</LI>
  334. <LI>If the -q option species a value of zero (or no value
  335. is given, defaulting to zero), all console output will
  336. be suppressed except for error messages, if any.
  337. (Matt Dillon)</LI>
  338. <LI>The 128-byte restriction on constant length no longer
  339. applies to the entire code generated by a single DCB
  340. statement; statements such as
  341. <PRE> DCB.L 64,0</PRE>
  342. can now be handled. (Colin Fox)</LI>
  343. <LI>Forward branches are now optimized. The occasional
  344. instruction may be missed due to ripple effects, but
  345. this shouldn't happen frequently. The -f switch
  346. will flag any such instructions.</LI>
  347. <LI>The -d switch can now be followed by an optional
  348. prefix string (with or without a leading !) which
  349. specifies which symbols should or should not be
  350. included in the symbol table dump. (Lionel Hummel)</LI></UL><BR>
  351. <B>Version 2.5 (Charlie Gibbs, June 18, 1989)</B>
  352. <BR><BR>
  353. The following bugs in version 2.42 have been corrected:
  354. <UL><LI>Upon normal termination, A68k occasionally crashed in
  355. quit_cleanup by trying to free the relocation table twice.
  356. (Jeff Lydiatt and D. McClelland, who worked out a fix)</LI>
  357. <LI>MEMF_CHIP and MEMF_FAST bits were being set in the
  358. hunk length, rather than in the hunk type. (Richard Man)</LI>
  359. <LI>BCHG.L, BCLR.L, BSET.L, and BTST.L were causing
  360. phase errors. The test to ignore the .L specification
  361. (added in version 1.21) was being skipped in pass 1
  362. by an optimization added in version 2.4. (David Hankins)</LI>
  363. <LI>PC-relative offset to a label was calculated as two bytes
  364. too great for MOVEM instructions. (Tony Parkhurst)</LI></UL>
  365. The following enhancements have been added:
  366. <UL><LI>If the length code on an opcode is not .B, .W, .L, .S,
  367. or omitted, it will be flagged as an error.</LI>
  368. <LI>JMP.S and JSR.S are flagged as errors. (Jim Butterfield)</LI>
  369. <LI>Operands of the form (xxxx).W and (xxxx).L are now
  370. supported. This enables absolute short or absolute
  371. long addressing to be explicitly specified.</LI>
  372. <LI>All optimization can be disabled by the new -n switch.
  373. (David Hankins)</LI>
  374. <LI>The NEAR directive can now take a single operand, which
  375. can be any address register (or equated symbol) except
  376. A7. If omitted, the register defaults to A4.</LI>
  377. <LI>Instructions of the form BTST Dn,#nn are no longer
  378. flagged. This obscure variant is nonetheless legal.</LI></UL><BR>
  379. <B>Version 2.42 (Charlie Gibbs, January 10, 1989)</B>
  380. <BR><BR>
  381. The following bugs in version 2.41 have been corrected:
  382. <UL><LI>Small code/data conversion was sometimes taking place
  383. when no NEAR directive was active. (Jeff Lydiatt)</LI></UL><BR>
  384. <B>Version 2.41 (Charlie Gibbs, January 6, 1989)</B>
  385. <BR><BR>
  386. The following bugs in version 2.4 have been corrected:
  387. <UL><LI>The second operand of LINK instructions was
  388. being erroneously flagged.</LI>
  389. <LI>If a macro was used before it was defined, it
  390. was being expanded during pass 2 but not during
  391. pass 1, causing severe phase errors. Attempts
  392. to use a macro before it is defined will now
  393. be flagged as invalid opcodes. (Colin Fox)</LI></UL><BR>
  394. <B>Version 2.4 (Charlie Gibbs, January 4, 1989)</B>
  395. <BR><BR>
  396. The following bugs in version 2.31 have been corrected:
  397. <UL><LI>If comments immediately followed the operands of
  398. a DC statement with no intervening white space,
  399. A68k would hang. (Ulf Nordquist)</LI>
  400. <LI>In the following command:
  401. <PRE>a68k -w 15000 myprog.asm</PRE>
  402. the space between the -w and 15000 would cause A68k
  403. to look for a source file called &quot;15000&quot;, and to think
  404. that the object file is to be called &quot;myprog.asm&quot;.
  405. When it can't find &quot;15000&quot; it would display an error
  406. message and scratch &quot;myprog.asm&quot;. (Jeff Lydiatt)</LI>
  407. <LI>If an INCLUDE file that is skipped on pass 2 contains
  408. a macro call, subsequent uses of \@ (macro sequence
  409. number are subsequently flagged. The macro counter
  410. must be bumped along with the line number when
  411. skipping an INCLUDE. (Colin Fox, Harvey Taylor)</LI></UL>
  412. The following enhancements have been added:
  413. <UL><LI>ORG and RORG are now fully implemented.</LI>
  414. <LI>The SET symbols A68k, a68K, and a68k are defined in the
  415. same way as A68K, making it effectively case-insensitive.
  416. (Colin Fox)</LI>
  417. <LI>MOVEM and REG now accept equated register names (EQUR)
  418. in register lists. (Bruce Dawson)</LI>
  419. <LI>INCLUDE files will now be skipped on pass 2 even when
  420. a listing file is requested, if the listing has been
  421. turned off by a NOLIST directive before the INCLUDE,
  422. and is not turned on until after the end of the
  423. INCLUDE file has been reached. (Colin Fox)</LI>
  424. <LI>A new switch (-f) causes forward branches (Bcc, BRA, BSR)
  425. that could be coded as short branches (Bcc.S etc.) to be
  426. flagged. This flag is not considered to be an error.</LI>
  427. <LI>A limited small code / small data model has been provided.
  428. It is activated by a NEAR directive in the source code, and
  429. is de-activated by a FAR directive. External variables
  430. must be declared at the beginning of the program, which
  431. must consist of only two sections (CODE and DATA or BSS).
  432. All forward data references are assumed to be PC-relative
  433. if in the CODE section, A4-relative if in the DATA/BSS
  434. section, and absolute word if absolute values. Any
  435. forward references which cannot be resolved to one of
  436. these three in pass 2 will be flagged as errors, as will
  437. any attempt to define more than two sections. A4 is
  438. assumed to point to the start of the DATA/BSS section
  439. plus 32768 bytes, and must be loaded by a MOVE.L
  440. instruction using immediate mode unless this instruction
  441. is not enclosed within NEAR and FAR directives.</LI>
  442. <LI>Miscellaneous optimizations, for speed, including:
  443. Most of the object code generator in pass 1 is bypassed.
  444. If GetValue gets a single term it takes a short cut.
  445. IsOperator now uses a table look-up.
  446. Instructions now only searches that portion of the
  447. opcode table whose opcodes start with the same letter
  448. as the OpCode being searched for.</LI></UL><BR>
  449. <B>Version 2.31 (Charlie Gibbs, November 30, 1988)</B>
  450. <BR><BR>
  451. The following bugs in version 2.3 have been corrected:
  452. <UL><LI>Even though a macro definition was being skipped
  453. by IFxx/ENDC, its ENDM directive was still being
  454. detected, causing spurious diagnostics. (Harvey Taylor)</LI>
  455. <LI>NOP was not being recognized. When moving all
  456. directives into the opcode table, NOL and NOLIST
  457. were placed after NOP, rather than before. (Colin Fox)</LI>
  458. <LI>Symbols defined in the current module and declared
  459. as PUBLIC were not being written to the object code
  460. file when -d was specified. (Colin Fox)</LI>
  461. <LI>Conversion of 0(An) to (An) (implemented in version
  462. 1.2) was causing errors in the MOVEP instruction,
  463. which requires a displacement even if it is zero.
  464. This conversion is now disabled for MOVEP instructions.</LI>
  465. <LI>User macros containing invalid opcodes caused A68k
  466. to get lost when returning to the outer source file.
  467. (Colin Fox)</LI>
  468. <LI>Large values of -w (over 6000 or so) would cause
  469. a visit from the Guru. The work field in HashIt
  470. was overflowing and going negative. Changing it
  471. to unsigned corrected the problem. (Colin Fox)</LI>
  472. <LI>Although user macros are no longer displayed when
  473. -q is a negative number, the calling file's name
  474. was still being displayed at the end of the macro.</LI></UL><BR>
  475. <B>Version 2.3 (Charlie Gibbs, November 21, 1988)</B>
  476. <BR><BR>
  477. The following enhancements have been added:
  478. <UL><LI>All file I/O has been rewritten to use level 1 I/O
  479. (open, creat, close, read, write, and lseek) instead
  480. of level 2 I/O. A68k now does its own buffering and
  481. unbuffering to reduce system overhead and increase speed.
  482. (Bruce Dawson)</LI>
  483. <LI>All assembler directives have been incorporated into
  484. the opcode table. Since the opcode search now looks
  485. up directives as well, speed is increased.</LI>
  486. <LI>Miscellaneous code optimization for additional speed.</LI></UL><BR>
  487. <B>Version 2.2 (Charlie Gibbs, November 4, 1988)</B>
  488. <BR><BR>
  489. The following bugs in version 2.1 have been corrected:
  490. <UL><LI>Macro definitions within an INCLUDE file were
  491. disabling the test for skipping the file on pass 2.</LI>
  492. <LI>Errors encountered in an INCLUDE file on pass 1
  493. were not disabling the skip of the file on pass 2 -
  494. the pertinent error messages could not appear.</LI>
  495. <LI>XDEF information and optional symbol table dumps were
  496. not being written to the object code file for any
  497. hunks that did not contain relocatable code or data.
  498. (Colin Fox)</LI></UL>
  499. The following enhancements have been added:
  500. <UL><LI>If the -q option is specified as a negative value,
  501. user macros are no longer included in line number
  502. displays, reducing clutter.</LI>
  503. <LI>Some source code has been re-arranged to reduce size.</LI></UL><BR>
  504. <B>Version 2.1 (Charlie Gibbs, November 1, 1988)</B>
  505. <BR><BR>
  506. The following bugs in version 2.00 have been corrected:
  507. <UL><LI>Macro definitions that span two chunks of memory
  508. were causing garbage and probably a crash when
  509. the macro was being expanded. Pointers were not
  510. being handled properly when linking the two chunks.</LI>
  511. <LI>Statements such as EQU and SET were not being flagged
  512. as illegal forward references if referencing a label
  513. defined on the same line, e.g.
  514. <PRE>LABEL SET LABEL+1</PRE></LI>
  515. <LI>The position within macros and INCLUDE files was
  516. sometimes out by one line when reported in error
  517. messages (and the new feature of the -q switch).</LI></UL>
  518. The following enhancements have been added:
  519. <UL><LI>If the -q option is specified as a negative value,
  520. line numbers will be displayed as positions within
  521. the current module (whose name is also displayed),
  522. rather than a total statement count. (Bruce Dawson)</LI>
  523. <LI>INCLUDE files can be skipped on pass 2 even if they
  524. contain SET statements - the values of all symbols
  525. SET in the INCLUDE file are stored (as at the end
  526. of the file) in a separate table and are patched
  527. when the INCLUDE file is skipped. (Bruce Dawson)</LI></UL><BR>
  528. <B>Version 2.00 (Charlie Gibbs, October 26, 1988)</B>
  529. <BR><BR>
  530. The following bugs in version 1.24 have been corrected:
  531. <UL><LI>The last digit of the statement number display
  532. (lengthened in version 1.24) was not being erased
  533. before displaying error messages.</LI>
  534. <LI>A68k would go into a loop if a user macro was
  535. missing an ENDC directive. This error is now
  536. flagged (see below).</LI></UL>
  537. The following enhancements have been added:
  538. <UL><LI>The highest statement number displayed at the end of
  539. each pass is now left on the screen. This means that,
  540. at the end of pass 1, you can always see how many lines
  541. A68k will have to process in pass 2, giving an idea of
  542. how how much longer you have to wait. (Colin Fox)</LI>
  543. <LI>The symbol table is now built using a hashing algorithm.
  544. This eliminates the slowdown that occurs in pass 1 as
  545. the symbol table grows, due to the old insertion process.
  546. (Bruce Dawson)</LI>
  547. <LI>If A68k terminates abnormally for any reason (such as
  548. insufficient memory) the object file is scratched
  549. (unless the -k option is set). (Bruce Dawson)</LI>
  550. <LI>Any INCLUDE files which cannot be found are flagged
  551. as errors in pass 1, and the assembly is aborted
  552. at the end of pass 1. (Bruce Dawson)</LI>
  553. <LI>Missing ENDC directives are flagged in macro expansions.
  554. Also, missing or unpaired ENDC directives in user macros
  555. are flagged.</LI>
  556. <LI>If an INCLUDE file doesn't generate any code and no
  557. listing file is required, it won't be read again in
  558. pass 2. The statement numbers will be bumped to keep
  559. in proper alignment. This can really speed up
  560. assemblies that INCLUDE lots of equates. (Colin Fox)</LI></UL><BR>
  561. <B>Version 1.24 (Charlie Gibbs, October 11, 1988)</B>
  562. <BR><BR>
  563. The following bugs in version 1.23 have been corrected:
  564. <UL><LI>MOVEA to a data register was not being flagged, even
  565. though all other invalid addressing modes were.</LI>
  566. <LI>Attempts to ORG out of the current hunk (including
  567. to an absolute address) were not being flagged. (E. Lenz)</LI>
  568. <LI>If the size of the bottom of the primary heap (symbols
  569. and macro text) exceeded 32K, any further macro
  570. definitions would expand as endless garbage. (Colin Fox)</LI>
  571. <LI>If the size of the bottom of the primary heap (symbols
  572. and macro text) exceeded 64K, any further external
  573. symbols (XDEF) would be flagged as relocatability
  574. errors upon each reference. (Colin Fox)</LI></UL>
  575. The following enhancements have been added:
  576. <UL><LI>Where statement numbers are displayed as fixed-length
  577. fields, their maximum length has been increased
  578. from 4 digits to 5. (Colin Fox)</LI>
  579. <LI>The PUBLIC directive has been implemented.
  580. As with the Aztec assembler, any labels defined as
  581. PUBLIC will be treated as XDEF if defined within
  582. the current module, and XREF otherwise. (Jeff Lydiatt)</LI></UL><BR>
  583. <B>Version 1.23 (Charlie Gibbs, September 20, 1988)</B>
  584. <BR><BR>
  585. The following bugs in version 1.22 have been corrected:
  586. <UL><LI>The test for a third operand was producing erroneous
  587. error messages on instructions whose second operand
  588. was in immediate mode. The '#' was not being taken
  589. into account, since it is not copied to DestOp.</LI></UL><BR>
  590. <B>Version 1.22 (Charlie Gibbs, August 31, 1988)</B>
  591. <BR><BR>
  592. The following bugs in version 1.21 have been corrected:
  593. <UL><LI>Expressions of the form R-A, where R is a relocatable
  594. term or expression and A is an absolute term or
  595. expression, were being flagged as relocation errors.
  596. This was due to a bug in the routine which should
  597. (but did not) flag expressions of the form A-R.
  598. (David Ashley)</LI>
  599. <LI>Instructions with three operands were not being
  600. flagged as errors. This can be caused by an extra
  601. comma being typed in the instruction, as in:
  602. <PRE> BTST #0,state+3,(a5)</PRE>
  603. The second comma should not be present. (David Ashley)</LI></UL>
  604. The following enhancements have been added:
  605. <UL><LI>Excess spacing has been removed from the listing file.
  606. These changes are similar to those already made to the
  607. console output (probably at about version 1.05).</LI>
  608. <LI>If the first statement in the source file is TTL or
  609. PAGE, an empty page is no longer produced at the
  610. start of the listing.</LI></UL><BR>
  611. <B>Version 1.21 (Charlie Gibbs, July 29, 1988)</B>
  612. <BR><BR>
  613. The following bugs in version 1.2 have been corrected:
  614. <UL><LI>The instruction
  615. <PRE> BTST.L #8,D0</PRE>
  616. had a long-word value generated for the bit number.
  617. This bug also applies to BSET, BCLR, and BCHG.
  618. The .L specification is now ignored. (Ulf Nordquist)</LI></UL><BR>
  619. <B>Version 1.2 (Charlie Gibbs, July 19, 1988)</B>
  620. <BR><BR>
  621. The following bugs in version 1.12 have been corrected:
  622. <UL><LI>A reference to the label of the current instruction
  623. was being converted to PC-relative on pass 2 but not
  624. on pass 1. This was causing phase errors. The label
  625. hasn't been added to the symbol table at the time the
  626. instruction is processed. Conversion to PC-relative
  627. addressing will now not be attempted in this case,
  628. although references to * can and will be converted.</LI>
  629. <LI>All string-type DC statements, regardless of length,
  630. were being treated as DC.B. For example, DC.L 'A'
  631. would generate only one byte of object code.
  632. (Gerald Hull)</LI>
  633. <LI>DC.W and DC.B statements were not being checked to
  634. ensure that their values would fit into a word or
  635. a byte respectively.</LI>
  636. <LI>If a comment line had white space preceding the
  637. asterisk, A68k would hang. Actually, it was
  638. interpreting the asterisk as an opcode and trying
  639. to open a macro file called &quot;*&quot;. Since under
  640. AmigaDOS such a file is the console, A68k was
  641. actually waiting for console input.</LI>
  642. <LI>If an instruction with no operands (such as RTS
  643. or NOP) followed MOVE.L #rel,D0 where &quot;rel&quot; was
  644. a relocatable symbol, the RTS (etc.) would have
  645. its nonexistent operands flagged as invalid.</LI>
  646. <LI>SECTION names enclosed in quotes were not being
  647. handled correctly.</LI>
  648. <LI>Source modules that did not generate any code, data,
  649. or BSS areas, but only defined symbols, such as
  650. <PRE>label equ 4
  651. xdef label
  652. end</PRE>
  653. were generating incomplete object modules.</LI></UL>
  654. The following enhancements have been added:
  655. <UL><LI>Jeff's experimental hunk code (prefixing hunk names
  656. with a sequence number before adding to the symbol
  657. table) has been permanently incorporated. It seems
  658. to work better with BLink on programs that have
  659. hunks continued farther on in the source code.
  660. (Jeff Lydiatt)</LI>
  661. <LI>The macro parameter \0, which is replaced by the
  662. size specification in the macro call (B, W, or L,
  663. defaulting to W) is now supported. (Gerald Hull)</LI>
  664. <LI>Operands of the form 0(An) will be treated as (An).
  665. (Bruce Dawson)</LI></UL><BR>
  666. <B>Version 1.12 (Charlie Gibbs, May 25, 1988)</B>
  667. <BR><BR>
  668. The following bugs in version 1.11 have been corrected:
  669. <UL><LI>If an instruction with no operands (e.g. RTS)
  670. followed a MOVE.L #label,D0 the RTS would be
  671. flagged with a relocatability error. Src.Mode
  672. and Dest.Mode were not being cleared. (Colin Fox)</LI></UL><BR>
  673. <B>Version 1.11 (Charlie Gibbs, April 6, 1988)</B>
  674. <BR><BR>
  675. The following bugs in version 1.10 have been corrected:
  676. <UL><LI>A68k would go into a loop while processing the
  677. arguments of a macro call, if these arguments are
  678. followed by comments separated from the arguments
  679. by one or more tab characters, and the -t switch
  680. is specified on the command line. All tests for
  681. blanks have been replaced by calls to isspace().</LI>
  682. <LI>The operand alignment checks added in version 1.06
  683. were erroneously testing the following instructions:
  684. <UL><LI>BCHG</LI>
  685. <LI>BCLR</LI>
  686. <LI>BSET</LI>
  687. <LI>BTST</LI>
  688. <LI>NBCD</LI>
  689. <LI>Scc</LI>
  690. <LI>TAS</LI></UL>
  691. These instructions are now exempt from alignment checking.</LI></UL>
  692. The following enhancements have been added:
  693. <UL><LI>A listing file name can now be specified with the
  694. -x switch; it is no longer necessary to specify
  695. both the -l and -x switches to produce a cross-
  696. reference listing with a name other than the default.</LI>
  697. <LI>DS statements with more than one operand are
  698. flagged and ignored (in case they should be DC).</LI>
  699. <LI>A character string used as a numeric value is
  700. flagged and set to zero if it is more than four
  701. characters long.</LI></UL><BR>
  702. <B>Version 1.10 (Charlie Gibbs, March 20, 1988)</B>
  703. <BR><BR>
  704. The following bugs in version 1.07 have been corrected:
  705. <UL><LI>BSS sections were not being written to the object
  706. code file except for a BSS section at the end of
  707. a program. This is due to a bug in the code added
  708. in version 1.05 to overwrite null sections.</LI>
  709. <LI>If a source module contained a mixture of lengths
  710. (8, 16, or 32 bits) in external references (XREF)
  711. to the same label, all references were being treated
  712. as if they has the length of the first reference.</LI></UL>
  713. The following enhancements have been added:
  714. <UL><LI>DS operands that are either a forward references
  715. or relocatable are now flagged.</LI>
  716. <LI>Short branches (Bcc.S, including BRA and BSR) to
  717. the next instruction (i.e. a displacement of zero)
  718. are illegal - the processor takes the displacement
  719. from the next word. Attempts to generate a short
  720. displacement of zero are now flagged.</LI></UL><BR>
  721. <B>Version 1.07 (Charlie Gibbs, March 11, 1988)</B>
  722. <BR><BR>
  723. The following bugs in version 1.06 have been corrected:
  724. <UL><LI>Instructions that take no operands (such as RTS)
  725. were being flagged if they had comments that were
  726. not preceded by a semicolon.</LI></UL>
  727. The following enhancements have been added:
  728. <UL><LI>The following synonyms have been added:
  729. <PRE>CSEG for CODE (Aztec compatibility)
  730. DSEG for DATA &quot; &quot;
  731. ENDIF for ENDC (Assempro compatibility)
  732. = for EQU &quot; &quot;
  733. | for ! &quot; &quot;</PRE></LI>
  734. <LI>Strings and character values may be delimited by
  735. either apostrophes (') or quotation marks (&quot;).
  736. The character not used as a delimiter can be used
  737. within the string without doubling it. For example,
  738. <PRE> DC.B &quot;This is Charlie's assembler&quot;</PRE>
  739. produces the same code as
  740. <PRE> DC.B 'This is Charlie''s assembler';</PRE></LI>
  741. <LI>The object code file will be scratched if any errors
  742. were found, unless the -k (keep) flag is set.
  743. (Bruce Dawson)</LI>
  744. <LI>The symbol .A68K is automatically defined at the
  745. beginning of each assembly as a SET symbol with an
  746. absolute value of 1. This enables programs to check
  747. whether they're being assembled by this assembler.
  748. (Jeff Lydiatt)</LI>
  749. <LI>The symbol table insertion routine has been
  750. greatly speeded up.</LI></UL><BR>
  751. <B>Version 1.06 (Charlie Gibbs, March 6, 1988)</B>
  752. <BR><BR>
  753. The following bugs in version 1.05 have been corrected:
  754. <UL><LI>Lines skipped by IFxx/ENDC were not being counted
  755. in the line number given in error messages.</LI>
  756. <LI>DATA and BSS sections may be unnamed, or have names
  757. the same as CODE sections. Honest, I thought section
  758. names had to be unique even across types.</LI>
  759. <LI>CHIP and FAST options on the CODE, DATA, and BSS
  760. synonyms for the SECTION directive were not being
  761. handled correctly.</LI>
  762. <LI>XDEF records and symbol table records (if desired)
  763. were not being produced for symbols defined ahead
  764. of the first object-code producing instruction.</LI></UL>
  765. The following enhancements have been added:
  766. <UL><LI>The CNOP instruction can now force alignment
  767. relative to any boundary up to 128 bytes.
  768. The second operand must still be a power of 2.</LI>
  769. <LI>The -q switch has been added to change the frequency
  770. with which progress reports (current line number) are
  771. displayed on the console. The default remains at
  772. every 10 lines (-q10). If you specify -q (no interval)
  773. or -q0 the line number displays will be suppressed.
  774. This will make assemblies run slightly faster due to
  775. reduced console I/O. (Bill Henning)</LI>
  776. <LI>The -t switch has been added to keep any tabs in the
  777. source file when producing the listing file, as well as
  778. generating tabs elsewhere whenever possible. This
  779. speeds up assemblies and gives smaller listing files,
  780. but such listing files cannot be displayed on devices
  781. that do not assume a tab stop in every 8th position.
  782. (Bruce Dawson)</LI>
  783. <LI>Any single-operand instruction with two operands,
  784. and any no-operand instruction with any operands,
  785. will be flagged.</LI>
  786. <LI>Relocatable 8- or 16-bit immediate operands
  787. will be flagged. They blow up BLink.</LI>
  788. <LI>Named local labels are now supported. Their names
  789. are formed in the same way as normal labels, but are
  790. then preceded by a backslash. Their scope is the
  791. same as normal local labels (nnn$). (Colin Fox)</LI>
  792. <LI>An alignment error will be flagged in the following cases:
  793. <UL><LI>Odd displacement on a LINK instruction</LI>
  794. <LI>Bcc or DBcc to an odd address</LI>
  795. <LI>In any word or long-word instruction, any operand
  796. using the following addressing modes:
  797. <UL><LI>Address register indirect with displacement</LI>
  798. <LI>Address register indirect with index and displacement</LI>
  799. <LI>Absolute short</LI>
  800. <LI>Absolute long</LI>
  801. <LI>Program counter indirect with displacement</LI>
  802. <LI>Program counter indirect with index and displacement</LI></UL></LI>
  803. <LI>LEA and PEA instructions are exempt from these tests.</LI></UL></LI>
  804. <LI>If a section is found to contain no data, A68k will
  805. back up to its beginning and overwrite it with the
  806. next section. The result is that null sections
  807. will no longer appear in the object file.</LI></UL><BR>
  808. <B>Version 1.05 (Charlie Gibbs, October 30, 1987)</B>
  809. <BR><BR>
  810. The following bugs in version 1.04 have been corrected:
  811. <UL><LI>If a section was continued later in the program, e.g.
  812. <PRE> SECTION prog,CODE
  813. &lt;code&gt;
  814. SECTION variables,BSS
  815. &lt;DS statements&gt;
  816. SECTION prog,CODE
  817. &lt;more code&gt;</PRE>
  818. bad relocation information was being generated for
  819. the continuation of the SECTION. This bug was left
  820. over from version 1.03.</LI></UL>
  821. The following enhancements have been added:
  822. <UL><LI>All console output except for error messages is now
  823. sent to stderr - this enables stdout to be redirected,
  824. producing an error file.</LI>
  825. <LI>Console (stderr) output has been modified to require
  826. fewer lines on the screen.</LI>
  827. <LI>If an error occurs while expanding a macro or INCLUDE
  828. file, the position of the call in each outer file is
  829. given along with the position in the current (innermost)
  830. file. Tracing continues until the outermost file (i.e.
  831. the original source file) is reached.</LI></UL><BR>
  832. <B>Version 1.04 (Charlie Gibbs, October 21, 1987)</B>
  833. <BR><BR>
  834. The following bugs in version 1.03 have been corrected:
  835. <UL><LI>MOVE was being converted to MOVEQ regardless of
  836. operand size - this conversion is legal only
  837. for longword MOVEs.</LI>
  838. <LI>Modifications to version 1.03 caused bad relocatable
  839. entries to be generated.</LI></UL><BR>
  840. <B>Version 1.03 (Charlie Gibbs, October 14, 1987)</B>
  841. <BR><BR>
  842. The following bugs in version 1.02 have been corrected:
  843. <UL><LI>The following situation was causing phase errors:
  844. <PRE> xdef label
  845. bra label
  846. .
  847. &lt;at least 128 bytes of object code&gt;
  848. .
  849. label:</PRE>
  850. (The XDEF was fooling A68k into thinking that &quot;label&quot;
  851. was defined within 128 bytes of the BRA instruction
  852. on pass 1, although on pass 2 it knew better.</LI>
  853. <LI>If the first operand of an two-operand executable
  854. instruction contained a character term containing a
  855. left or right parenthesis, it would generate error
  856. messages and be incorrectly evaluated.</LI>
  857. <LI>Labels that don't begin in column 1 (denoted by a
  858. trailing colon) caused a Guru Meditation.</LI>
  859. <LI>Certain ADD and SUB instructions using PC-relative
  860. addressing may cause phase errors. If the displacement
  861. is in the range 1 to 8 inclusive, the instruction was
  862. erroneously converted to ADDQ or SUBQ during pass 2.</LI></UL>
  863. The following enhancements have been added:
  864. <UL><LI>The -z option has been added to display the
  865. current source program line on stdout as it
  866. is read, optionally over a given range.
  867. This feature is provided for debugging purposes.</LI>
  868. <LI>Bcc, BSR, and DBcc to labels in other than the current
  869. section is now supported. A 16-bit relocation entry
  870. will be generated for each such reference.</LI>
  871. <LI>PC relative mode will be generated for backward
  872. references to labels within the current CODE section
  873. if legal for the current instruction. Forward
  874. references will not be converted, since there is
  875. no way of telling which section the label is in
  876. during pass 1.</LI>
  877. <LI>The cumulative sizes of all sections by type (i.e.
  878. CODE, DATA, and BSS) will be displayed at the end
  879. of the listing file and the console display.
  880. (Bruce Dawson)</LI>
  881. <LI>In the symbol table dump, section names will no
  882. longer be indicated just as SECTION, but rather
  883. as CODE, DATA, or BSS, depending on type.</LI></UL><BR>
  884. <B>Version 1.02 (Charlie Gibbs, September 9, 1987)</B>
  885. <BR><BR>
  886. The following bugs in version 1.01 have been corrected:
  887. <UL><LI>Duplicate labels were not being flagged.</LI>
  888. <LI>XDEF symbols were not being dumped to the
  889. object code file when the -d option was set.</LI></UL>
  890. The following enhancements have been added:
  891. <UL><LI>A header file is now supported. If the parameter
  892. -h&lt;filespec&gt; is included on the command line, the
  893. specified file will be included as if the source
  894. file's first line was &quot; include &lt;filespec&gt;&quot;.
  895. The file specification may include a path name,
  896. although the include path names given by the
  897. -i parameter (if any) will also be searched.</LI>
  898. <LI>An equate file can now be produced. If the parameter
  899. -e&lt;filespec&gt; is included on the command line, a file
  900. will be written containing EQU statements for any
  901. symbol whose value is absolute. If -e is specified
  902. without &lt;filespec&gt;, the name of the file will be
  903. formed in the same way as the list file, except with
  904. an extension of &quot;.equ&quot;. (Bruce Dawson)</LI></UL>
  905. The following changes have been made to existing logic:
  906. <UL><LI>No symbol table dump will be produced unless the
  907. -x (cross-reference) switch is set. Formerly a
  908. symbol table dump was always produced, with only
  909. the cross-reference portion optional.</LI></UL><BR>
  910. <B>Version 1.01 (Charlie Gibbs, August 20, 1987)</B>
  911. <BR><BR>
  912. The following bugs in version 1.00 have been corrected:
  913. <UL><LI>Long-word constants and storage areas were being
  914. aligned on a double-word boundary. The only place
  915. where double-word alignment is now forced is at a
  916. break between SECTIONs, since the length of an
  917. AmigaDOS hunk must be a multiple of 4 bytes.
  918. (CNOP 0,4 can still be used if double-word
  919. alignment is desired by the programmer.)</LI>
  920. <LI>If a label on an END statement or the first statement
  921. of a SECTION was named in an XDEF statement, it would
  922. not be written to the object code file. The latter
  923. case includes both the label of a SECTION directive
  924. and the label of the first executable instruction in
  925. the absence of any SECTION directives (defaulting to
  926. an unnamed CODE section). In the final case (default
  927. unnamed CODE section), references to XREF symbols
  928. in the first statement would also not be written
  929. to the object code file.</LI>
  930. <LI>If the last statement in the source file was not
  931. terminated with a newline character (premature EOF),
  932. it was being ignored altogether.</LI>
  933. <LI>A register list as the source operand of a MOVE
  934. instruction was not being flagged as an error.
  935. (MOVE to a register list was being flagged, however.)</LI>
  936. <LI>MOVE from USP was generating incorrect code. Also,
  937. MOVE from SR or CCR to an address register was
  938. generating incorrect code rather than being flagged.</LI></UL><BR>
  939. <B>Version 1.00 (Charlie Gibbs, June 18, 1987) - initial release</B>