single float

[rurban]

Attached patch adds experimental support for 4-byte single float. It also changes and simplifies the numeric converters by eliminating all special endianizers.

There are only numeric converters, the endianizers are added automatically.

Missing:

• jit is turned off
• reading 16-byte and 12-byte long double on 4-byte float platforms not yet implemented. Not needed most likely.

Those tests had to be rewritten for less precision.

Test Summary Report
-------------------
t/steps/auto_jit-01                     (Wstat: 256 Tests: 50 Failed: 1)
  Failed test:  12
  Non-zero exit status: 1
t/compilers/imcc/imcpasm/opt1           (Wstat: 256 Tests: 78 Failed: 1)
  Failed test:  75
  Non-zero exit status: 1
t/op/arithmetics                        (Wstat: 1024 Tests: 30 Failed: 4)
  Failed tests:  8-9, 18-19
  Non-zero exit status: 4
t/op/integer                            (Wstat: 256 Tests: 56 Failed: 1)
  Failed test:  33
  Non-zero exit status: 1
t/op/number                             (Wstat: 1280 Tests: 56 Failed: 5)
  Failed tests:  1, 27, 32-33, 40
  Non-zero exit status: 5
t/op/sprintf                            (Wstat: 0 Tests: 308 Failed: 17)
  Failed tests:  11-12, 14-15, 102, 104, 122-123, 162-165
                167, 178-179, 185-186
t/op/sysinfo                            (Wstat: 512 Tests: 14 Failed: 2)
  Failed tests:  3-4
  Non-zero exit status: 2
t/op/time                               (Wstat: 256 Tests: 7 Failed: 1)
  Failed test:  3
  Non-zero exit status: 1
t/pmc/bigint                            (Wstat: 256 Tests: 44 Failed: 1)
  Failed test:  4
  Non-zero exit status: 1
t/pmc/complex                           (Wstat: 139 Tests: 0 Failed: 0)
  Parse errors: Bad plan.  You planned 467 tests but ran 0.
t/pmc/complex_nan                       (Wstat: 256 Tests: 0 Failed: 0)
  Non-zero exit status: 1
  Parse errors: Bad plan.  You planned 16 tests but ran 0.
t/pmc/float                             (Wstat: 4864 Tests: 61 Failed: 19)
  Failed tests:  4, 14, 44-53, 55-61
  Non-zero exit status: 19
t/pmc/nci                               (Wstat: 512 Tests: 69 Failed: 2)
  Failed tests:  5-6
  Non-zero exit status: 2
t/pmc/string                            (Wstat: 0 Tests: 164 Failed: 1)
  Failed test:  5
t/pmc/sub                               (Wstat: 256 Tests: 64 Failed: 1)
  Failed test:  40
  Non-zero exit status: 1
t/pmc/timer                             (Wstat: 512 Tests: 6 Failed: 2)
  Failed tests:  4-5
  Non-zero exit status: 2
t/dynoplibs/myops                       (Wstat: 256 Tests: 10 Failed: 1)
  Failed test:  5
  Non-zero exit status: 1
t/library/dumper                        (Wstat: 256 Tests: 26 Failed: 1)
  Failed test:  17
  Non-zero exit status: 1
t/library/mt19937ar                     (Wstat: 256 Tests: 10 Failed: 1)
  Failed test:  10
  Non-zero exit status: 1
t/library/pcre                          (Wstat: 256 Tests: 1 Failed: 1)
  Failed test:  1
  Non-zero exit status: 1
t/library/yaml_dumper                   (Wstat: 256 Tests: 26 Failed: 1)
  Failed test:  17
  Non-zero exit status: 1
t/tools/pbc_merge                       (Wstat: 256 Tests: 4 Failed: 1)
  Failed test:  2
  Non-zero exit status: 1
Files=397, Tests=11203, 1677 wallclock secs ( 7.78 usr  3.80 sys + 499.64 cusr 495.17 csys = 1006.38 CPU)
Result: FAIL
make: *** [test] Error 1 

[whiteknight]

Packfile format is deprecated, and future implemenations will be very different indeed. It's not worth the effort right now to change the current system to accept different types of floats. I am going to reject this ticket now. We will keep the ideas of single-precision floats in packfiles in mind when redesigning the packfile format.