More comparisons #55
More comparisons #55
Conversation
This is roughly equivalent to the `is` operator in Python in that it checks normal objects for pointer equality, but numbers are checked for value equality. Note that SOM (or, at least, java-som) doesn't define/implement the latter cases properly yet (see SOM-st/SOM#23) so this commit makes us diverge slightly.
|
|
||
| double10 = ( | ||
| run = ( | ||
| (1.0 = 0) println. |
There was a problem hiding this comment.
SOM is a little unusual (to our eyes) in this regadr. = means "check the values for equality". == means (roughly) "check the data pointer for equality" (though for integers this is overridden -- Python effectively does something similar).
There was a problem hiding this comment.
I see. Thanks for the explanation.
| (1.0 > 2) println. | ||
| (2.0 > 0) println. | ||
|
|
||
| (1.0 >= 1) println. |
There was a problem hiding this comment.
Wait, so in SOM, 1.0 == 1 is false, 1.0 > 1 is false, but 1.0 >= 1 is true?
There was a problem hiding this comment.
== is very different from the other two, and unrelated.
1.0 > 1 and 1 > 1 should yield both false, no? (minus issues with number representations for floating points, which hopefully do not matter for 1-like values, but will crop up with int-like values > 53-or-something bits).
And 1.0 >= 1 and 1 >= 1 should both be true in an ideal world (minus floating point issues)
There was a problem hiding this comment.
Ah thanks! With this comment and @ltratt's comment above, this makes a lot more sense now.
|
@smarr Just to confirm, do all the checks make sense to you (at least in yksom)? |
|
@ltratt sorry, I didn't check in detail the true/false values. The test structure makes it hard for me to match up tested expression and expected value. |
|
I compare Java SOM and yksom's output if that's any help. There are differences in |
|
|
|
@smarr OK, thanks! yksom, of course, aims to be an ideal world, a veritable nirvana, a stronghold of God's will on Earth. Except bugs, of course ;) |
|
Okay, I guess this is ready to merge then and I can tell bors the good news? |
|
Yes please! |
|
bors r+ |
55: More comparisons r=ptersilie a=ltratt This PR adds more comparisons to yksom: reference equality (1dca2bc) and various comparisons for doubles (82d35c5). Notice that the former commit means that we diverge slightly from SOM's current semantics (see SOM-st/SOM#23) although I think it's fair to say that @smarr thinks the approach we're taking in this PR is the right one (and if I've got that wrong, hopefully he'll shout!). Co-authored-by: Laurence Tratt <[email protected]>
Timed out |
|
Sorry, my CI changes yesterday broke this. I've applied the fix, but I just need a: bors r+ |
55: More comparisons r=vext01 a=ltratt This PR adds more comparisons to yksom: reference equality (1dca2bc) and various comparisons for doubles (82d35c5). Notice that the former commit means that we diverge slightly from SOM's current semantics (see SOM-st/SOM#23) although I think it's fair to say that @smarr thinks the approach we're taking in this PR is the right one (and if I've got that wrong, hopefully he'll shout!). Co-authored-by: Laurence Tratt <[email protected]>
|
bors r+ |
|
Already running a review |
Build succeeded |
This PR adds more comparisons to yksom: reference equality (1dca2bc) and various comparisons for doubles (82d35c5). Notice that the former commit means that we diverge slightly from SOM's current semantics (see SOM-st/SOM#23) although I think it's fair to say that @smarr thinks the approach we're taking in this PR is the right one (and if I've got that wrong, hopefully he'll shout!).