mburkley/tms9900-gcc

Issues

• Does not compile

  #69 opened 3 months ago by nils-eilers
  3

• [feature/question] Compile and link in one command

  #68 opened 4 months ago by Fabrizio-Caruso
  2

• [bug][OK in branch 1.32] A variable is assigned a wrong value by 1 (only under some circumstances)

  #58 opened 4 months ago by Fabrizio-Caruso
  2

• [bug][OK in branch 1.32] Random behavior in previously working code. Hard/impossible to reproduce.

  #59 opened 4 months ago by Fabrizio-Caruso
  1

• [bug][OK in branch 1.32] 19-19=255 with this updated version of GCC for the TI instead of 0

  #60 opened 4 months ago by Fabrizio-Caruso
  1

• Unoptimised 32 bit shifts fail

  #48 opened 4 months ago by mburkley
  2

• [bug][OK in branch 1.32] Wrong value read from vector

  #61 opened 4 months ago by Fabrizio-Caruso
  0

• [bug][OK in branch 1.32] 28 - 0 = 27 with a minimal example with source code and EA5 binary

  #62 opened 4 months ago by Fabrizio-Caruso
  1

• Left or right shift by a char shift-count can fail

  #63 opened 4 months ago by mburkley
  1

• [1.32 branch][bug][OK in latest changes of branch 1.32] Mostly good news except for "internal compiler error: in subreg_highpart_offset" when dividing a constant by an unsigned char

  #64 opened 4 months ago by Fabrizio-Caruso
  4

• [bug][ok now in latest changes in branch 1.32] cprintf from libti99 fails to print if we do 2*i before but works fine with (i<<1)

  #66 opened 4 months ago by Fabrizio-Caruso
  2

• [bug][OK in branch 1.32] Shift operations and addition seem to give wrong result, but this is triggered or affected by PRINT commands

  #65 opened 4 months ago by Fabrizio-Caruso
  1

• I think this is the wrong syntax, which we don't notice as long as everything works.

  #57 opened 6 months ago by jedimatt42
  1

• Access words in a __packed__ struct yields wrong result

  #52 opened a year ago by mburkley
  0

• tms9900_output_ascii should not close a string if it has already been closed due to max length reached

  #53 opened 6 months ago by mburkley
  1

• lib1funcs functions must save temp regs

  #49 opened a year ago by mburkley
  0

• gcc assert in reload

  #45 opened a year ago by mburkley
  0

• libgcc needs 32 bit shift operations

  #46 opened a year ago by mburkley
  0

• R0 needs to marked as fixed

  #47 opened a year ago by mburkley
  0

• Subtract can create negative results

  #42 opened a year ago by mburkley
  0

• Review subreg offset removal in andhi3, addhi3, subhi3

  #43 opened a year ago by mburkley
  0

• Forced constants should not be masked

  #44 opened a year ago by mburkley
  0

• Using literal consts for byte compares and byte math is more efficient

  #41 opened a year ago by mburkley
  0

• sub_const_hi assumes immediate is a const_int but it could be a label

  #38 opened a year ago by mburkley
  0

• movqi can fail to truncate HI to QI

  #32 opened a year ago by mburkley
  1

• Byte subtract can give wrong result

  #37 opened a year ago by mburkley
  1

• Segfault with debug turned on

  #39 opened a year ago by XZCE
  3

• Add TMS9995 support

  #36 opened a year ago by mburkley
  0

• Shouldn't compare to zero by doing MOV in place

  #34 opened a year ago by mburkley
  0

• Masking and checking shift operand is redundant

  #33 opened a year ago by mburkley
  0

• Binary and using insn "andhi3" fails to generate properly aligned code when using -O2

  #29 opened a year ago by mburkley
  0

• long multiplication/division results incorrect

  #31 opened a year ago by mcvde
  12

• Setting floats to literal values doesn't work

  #12 opened a year ago by mburkley
  0

• Macros to query patch level

  #28 opened a year ago by mburkley
  0

• Save of R11 should check for usage in a function

  #26 opened a year ago by mburkley
  0

• Shifting a long by a variable gives wrong result

  #21 opened a year ago by mburkley
  0

• It should be possible to build ROM independent images

  #23 opened a year ago by mburkley
  0

• R15 / SP should be marked as fixed so compiler does not try to save to stack

  #25 opened a year ago by mburkley
  0

• byte compare params in wrong order

  #19 opened a year ago by mburkley
  0

• byte immediate-and clears values

  #20 opened a year ago by mburkley
  0

• Initialising a long with a negative value causes a crash

  #15 opened a year ago by mburkley
  0

• Or operations on bytes clear the byte

  #16 opened a year ago by mburkley
  0

• Or operations can crash if output requires socb

  #17 opened a year ago by mburkley
  0

• Initialising very long strings (2k or more) crashes the compiler

  #14 opened a year ago by mburkley
  0

• output_operand: invalid expression as operand when doing modulo with bytes

  #5 opened a year ago by mburkley
  0

• stack frame needs separate space for saved regs

  #6 opened a year ago by mburkley
  1