home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #19 / NN_1992_19.iso / spool / comp / os / os2 / programm / 4555 < prev    next >
Encoding:
Internet Message Format  |  1992-08-29  |  9.9 KB
  1. Path: sparky!uunet!vnet.ibm.com
  2. From: dmm@vnet.ibm.com (dave)
  3. Message-ID: <19920829.092516.435@almaden.ibm.com>
  4. Date: Sat, 29 Aug 92 12:21:33 EDT
  5. Newsgroups: comp.os.os2.programmer
  6. Subject: Re: New CSD for C Set/2
  7. Organization: IBM Canada Lab
  8. News-Software: UReply 3.0
  9. X-X-From: dmm@vnet.ibm.com (Dave Mooney)
  10. References: <21269@optima.cs.arizona.edu>
  11. Lines: 133
  12.  
  13.             <1992Aug28.214305.11133@midway.uchicago.edu>
  14.  
  15. In <1992Aug28.214305.11133@midway.uchicago.edu> Timothy F. Sipples writes:
  16. > In article <21269@optima.cs.arizona.edu> merrill@cs.arizona.edu (Darren Merrill) writes:
  17. >> Well, I've been told that a new CSD (CS00013 ?) for C Set/2 is out.
  18. >> This one includes fixes for two of 'my' bugs.  Would some kind
  19. >> sole grab it from CompuServe (when it comes across)
  20. >> and send it on to Hobbes?  Thanks.
  21. >
  22. > It will be there in short order.
  23.  
  24. Thanks, Timothy.  And for anyone who is interested, here's a list of
  25. fixes which CS000013 contains.  dave
  26.  
  27. --- >8 --- Cut here ---
  28.  
  29.   C SET/2 RELEASE 1.0 FIXES
  30.   _________________________
  31.  
  32.   +-----------+---------------------------------------------------------------+
  33.   | APAR      | PROBLEM DESCRIPTION                                           |
  34.   | NUMBER    |                                                               |
  35.   +-----------+---------------------------------------------------------------+
  36.   | PJ03523   | Optimization problem: bad register selection causes general   |
  37.   |           | protection fault at runtime.                                  |
  38.   +-----------+---------------------------------------------------------------+
  39.   | PJ03662   | Optimization problem: bad register selection around a call    |
  40.   |           | causes incorrect output at runtime.                           |
  41.   +-----------+---------------------------------------------------------------+
  42.   | PJ03682   | Optimization problem: tail call optimization was altering a   |
  43.   |           | register erroneously.                                         |
  44.   +-----------+---------------------------------------------------------------+
  45.   | PJ03767   | The compiler could not open files with names containing DBCS  |
  46.   |           | characters where the second byte of the characters was 0x5C.  |
  47.   +-----------+---------------------------------------------------------------+
  48.   | PJ03923   | Optimization problem: generated floating-point code was       |
  49.   |           | causing the floating point register stack to overflow or      |
  50.   |           | underflow.                                                    |
  51.   +-----------+---------------------------------------------------------------+
  52.   | PJ03948   | Optimization problem: multiplying by constant negative        |
  53.   |           | numbers which are powers of 2 was causing incorrect code to   |
  54.   |           | be generated.                                                 |
  55.   +-----------+---------------------------------------------------------------+
  56.   | PJ03952   | Optimization problem: multiplying by constants that are near  |
  57.   |           | powers of 2 was causing incorrect code to be generated.       |
  58.   +-----------+---------------------------------------------------------------+
  59.   | PJ03957   | Compiler was emitting object records longer than the linker's |
  60.   |           | 1024 byte limit.  This causes the linker to produce the       |
  61.   |           | message "L1057: data record too large".  NOTE: There is more  |
  62.   |           | than one cause for this symptom and there is another fix      |
  63.   |           | coming soon.                                                  |
  64.   +-----------+---------------------------------------------------------------+
  65.   | NOAPAR    | Compile time performance was improved.                        |
  66.   +-----------+---------------------------------------------------------------+
  67.   | NOAPAR    | The compiler was unable to find files in a subdirectory on a  |
  68.   |           | LAN drive.                                                    |
  69.   +-----------+---------------------------------------------------------------+
  70.   | NOTE:  The fixes listed above were included in CSD0008.                   |
  71.   +-----------+---------------------------------------------------------------+
  72.   | PJ04295   | The kbhit library function was not working.                   |
  73.   +-----------+---------------------------------------------------------------+
  74.   | NOTE:  The fixes listed above were included in CSD0009.                   |
  75.   +---------------------------------------------------------------------------+
  76.   | PJ04465   | The cprintf, fprintf, printf, and sprintf functions were not  |
  77.   |           | printing floating-point values between 0.5 and 1.0 correctly  |
  78.   |           | when zero precision was specified.                            |
  79.   +-----------+---------------------------------------------------------------+
  80.   | NOTE:  The fixes listed above were included in CSD0010.                   |
  81.   +-----------+---------------------------------------------------------------+
  82.   | PJ04602   | Optimization problem: simple loops with induction variable    |
  83.   |           | increments greater than one were being incorrectly turned     |
  84.   |           | into memset or memcpy builtin function calls.                 |
  85.   +-----------+---------------------------------------------------------------+
  86.   | PJ04604   | When the compiler turned switch statements into jump tables,  |
  87.   |           | the default code segment was always being referenced even if  |
  88.   |           | the function containing the switch statement was specified in |
  89.   |           | an alloc_text pragma.                                         |
  90.   +-----------+---------------------------------------------------------------+
  91.   | PJ04652   | Optimization problem: the stack pointer was not being decre-  |
  92.   |           | mented before an indirect _Optlink call where the return      |
  93.   |           | value was being used as a parameter to a tail call.           |
  94.   +-----------+---------------------------------------------------------------+
  95.   | PJ04674   | When #pragma handler was specified for a 32-bit callback      |
  96.   |           | function the compiler generated bad code for copying the      |
  97.   |           | parameters from the 16-bit stack to the 32-bit stack.         |
  98.   +-----------+---------------------------------------------------------------+
  99.   | NOTE:  The fixes listed above were included in CSD0011.                   |
  100.   +-----------+---------------------------------------------------------------+
  101.   | PJ04811   | Incorrect code was being generated for a complicated          |
  102.   |           | expression where a short variable was being added to the      |
  103.   |           | address of an element of an array of structures.              |
  104.   +-----------+---------------------------------------------------------------+
  105.   | PJ05088   | Optimization problem: the compiler was trapping when opti-    |
  106.   |           | mizing a loop containing a divide and modulo operation with   |
  107.   |           | the same operands.                                            |
  108.   +-----------+---------------------------------------------------------------+
  109.   | PJ05100   | Optimization problem: incorrect code was being generated for  |
  110.   |           | a switch statement containing a case zero where the switch    |
  111.   |           | variable had to be converted to an int from an unsigned char. |
  112.   +-----------+---------------------------------------------------------------+
  113.   | PJ05124   | Incorrect code was being generated when the result of com-    |
  114.   |           | paring two shorts with less than or greater than operators    |
  115.   |           | was stored into an integer variable.                          |
  116.   +-----------+---------------------------------------------------------------+
  117.   | PJ05140   | Incorrect code was being generated for _Optlink functions     |
  118.   |           | that have a number of floating-point or aggregate parameters  |
  119.   |           | followed by integer parameters.                               |
  120.   +-----------+---------------------------------------------------------------+
  121.   | NOTE:  The fixes listed above were included in CSD0012.                   |
  122.   +---------------------------------------------------------------------------+
  123.   | PJ05013   | The compiler was trapping when trying to compile a compli-    |
  124.   |           | cated expression where the address of an element of an array  |
  125.   |           | of structures was indexed and then a member of the structure  |
  126.   |           | was accessed.                                                 |
  127.   +-----------+---------------------------------------------------------------+
  128.   | PJ04674   | The problem with #pragma handler on callback functions was    |
  129.   |           | really only a problem with _Far16 _Cdecl callback functions.  |
  130.   |           | So the fix in CSD0011 broke the case where #pragma handler    |
  131.   |           | was used on a _Far16 _Pascal callback function.  Now both     |
  132.   |           | work.                                                         |
  133.   +-----------+---------------------------------------------------------------+
  134.   | PJ04390   | The C runtime termination routine was not freeing all storage |
  135.   |           | allocated during initialization.                              |
  136.   +-----------+---------------------------------------------------------------+
  137.   | PJ05087   | Partial static initialization of a structure containing a     |
  138.   |           | union containing a structure containing bitfields was not     |
  139.   |           | working.                                                      |
  140.   +-----------+---------------------------------------------------------------+
  141.   | NOAPAR    | Some enhancements were made to the compiler to reduce memory  |
  142.   |           | usage during compilation and therefore reduce compile times.  |
  143.   +-----------+---------------------------------------------------------------+
  144.   | NOTE:  The fixes listed above were included in CSD0013.                   |
  145.   +---------------------------------------------------------------------------+
  146.