Re^2: Overcoming 5.10.0 vs 5.38.2 incompatibilities

I am aware that it is almost hopeless to sift through the perldelta files. I did send some code snippets and hope (really hope) that somebody may spot the issue or ask for the offending code. This is how I go through the expected values:

diag ("\n");
for my $delta ($delta_time, $delta_date){
    for my $field (qw (seconds minutes hours days months years)) {
    for my $type ('','total_') {
        my $func = "${type}${field}";
        for my $prec ('','trunc') {
            diag ("comapring ".$delta->$func($prec)." to ".$test_value
+s[$test_index]." at ".$test_index." func=".$func." prec=".$prec." typ
+e=".$type." field=".$field);
         ok($delta->$func($prec) == $test_values[$test_index++],"$func
+($prec) ($delta_text[$delta_index])")
           #or diag("Expected " . $test_values[$test_index++] . " got 
+" . $delta->$func($prec));
          or diag("Expected " . $test_values[$test_index] . " got " . 
+$delta->$func($prec));
        }
    }
    }
    $delta_index++;
}
[download]

I do not see in the code where Date::Manip methods are used

Comment on Re^2: Overcoming 5.10.0 vs 5.38.2 incompatibilities Download Code

Replies are listed 'Best First'.
Re^3: Overcoming 5.10.0 vs 5.38.2 incompatibilities by bliako (Abbot) on Mar 12, 2024 at 09:15 UTC
`I do not see in the code where Date::Manip methods are used` [download] I have added some comments to your code so that you understand what the script does, especially with regards to the above: for my $delta ($delta_time, $delta_date){ # the above are likely to be Date::Manip[::*] objects - which ones???? for my $field (qw (seconds minutes hours days months years)) { for my $type ('','total_') { my $func = "${type}${field}"; # this is the method name to be called via the $delta objects for my $prec ('','trunc') { # this is the a parameter to above method, one is empty on +e is 'trunc' diag ("comapring ".$delta->$func($prec)." to ".$test_value +s[$test_index]." at ".$test_index." func=".$func." prec=".$prec." typ +e=".$type." field=".$field); ok($delta->$func($prec) == $test_values[$test_index++],"$func +($prec) ($delta_text[$delta_index])") # it calls e.g. this: $delta->total_seconds('trunc') # and it compares the returned result with an item from an ar +ray # of expected results $test_values[$test_index] [download] just do some plain 'ol debugging using common sense and forget about the deltas for the moment. For example, from your failed tests: # comapring 571645.000000 to 571645 at 7 func=total_minutes prec=trunc + type=total_ field=minutes # comapring 9527.417778 to 23.417778 at 8 func=hours prec= type= field +=hours # Failed test 'hours() (with time)' # at t/test.t line 97. # Expected 23 got 9527.417778 # comapring 9527 to 23 at 9 func=hours prec=trunc type= field=hours # Failed test 'hours(trunc) (with time)' # at t/test.t line 97. # Expected 9527.417778 got 9527 # comapring 9527.417778 to 9527.417778 at 10 func=total_hours prec= ty +pe=total_ field=hours [download] it seems that you have 2 issues : the easy one: `9527.417778 got 9527`. The other is that probably your expected results got mixed up at `Expected 23 got 9527.417778` especially since `34298704/60 = 571645 and 571645/60 = 9527.4166` you mixed your minutes somewhere along the road... bw, bliako Edit: actually for the 1st issue (`Expected 9527.417778 got 9527`) the deltas may help you. It is likely that at some stage there was a revision in D::M to return only the integer part. In general, when you compare floating point numbers you should not do an exact test but a range test: `ok(abs($expected-$got)<1E-09, "yep, they are the same up to 1E-09")` (although it would not be helpful in your case).	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re^3: Overcoming 5.10.0 vs 5.38.2 incompatibilities
by bliako (Abbot) on Mar 12, 2024 at 09:15 UTC

I do not see in the code where Date::Manip methods are used
[download]

I have added some comments to your code so that you understand what the script does, especially with regards to the above:

for my $delta ($delta_time, $delta_date){
# the above are likely to be Date::Manip[::*] objects - which ones????
    for my $field (qw (seconds minutes hours days months years)) {
    for my $type ('','total_') {
        my $func = "${type}${field}";
        # this is the method name to be called via the $delta objects
        for my $prec ('','trunc') {
            # this is the a parameter to above method, one is empty on
+e is 'trunc'
            diag ("comapring ".$delta->$func($prec)." to ".$test_value
+s[$test_index]." at ".$test_index." func=".$func." prec=".$prec." typ
+e=".$type." field=".$field);
         ok($delta->$func($prec) == $test_values[$test_index++],"$func
+($prec) ($delta_text[$delta_index])")
         # it calls e.g. this: $delta->total_seconds('trunc')
         # and it compares the returned result with an item from an ar
+ray
         # of expected results $test_values[$test_index]
[download]

just do some plain 'ol debugging using common sense and forget about the deltas for the moment.

For example, from your failed tests:

# comapring 571645.000000 to 571645 at 7 func=total_minutes prec=trunc
+ type=total_ field=minutes
# comapring 9527.417778 to 23.417778 at 8 func=hours prec= type= field
+=hours

#   Failed test 'hours() (with time)'
#   at t/test.t line 97.
# Expected 23 got 9527.417778
# comapring 9527 to 23 at 9 func=hours prec=trunc type= field=hours

#   Failed test 'hours(trunc) (with time)'
#   at t/test.t line 97.
# Expected 9527.417778 got 9527
# comapring 9527.417778 to 9527.417778 at 10 func=total_hours prec= ty
+pe=total_ field=hours
[download]

it seems that you have 2 issues : the easy one: 9527.417778 got 9527. The other is that probably your expected results got mixed up at Expected 23 got 9527.417778 especially since 34298704/60 = 571645 and 571645/60 = 9527.4166 you mixed your minutes somewhere along the road...

bw, bliako

Edit: actually for the 1st issue (Expected 9527.417778 got 9527) the deltas may help you. It is likely that at some stage there was a revision in D::M to return only the integer part. In general, when you compare floating point numbers you should not do an exact test but a range test: ok(abs($expected-$got)<1E-09, "yep, they are the same up to 1E-09") (although it would not be helpful in your case).

[reply]
[d/l]
[select]