01:00
<ptomato>
what does this mean concretely? "if an early-exiting if/else has just two branches (i.e. not an if-else cascade like in some Math functions), and the two branches are substantially similar and short, prefer a single step"
01:00
<ptomato>

does it mean to prefer

  1. If Something(), return thing; else return other.
01:01
<ptomato>

over

  1. If Something(), return thing.
  2. Return other.

?

01:02
<bakkot>
yeah
01:02
<ptomato>
ok
01:02
<ptomato>
thanks
01:02
<ptomato>
but not this?
01:02
<ptomato>
  1. If Something(), throw a TypeError exception; else return thing.
01:02
<ptomato>
because the two branches aren't substantially similar?
01:03
<bakkot>
"substantially similar" is always vibes, but personally I would write that as two steps for that reason, yes
01:03
<ptomato>
all right, got it. thanks
18:07
<ptomato>

existing preference on

the Foo column of Table 777

vs

the "Foo" column of Table 777

18:07
<ptomato>
?
18:09
<ptomato>
both 402 and 262 use both, but the latter is more prevalent in 262 and vice versa
18:16
<Michael Ficarra>
I only see quoted occurrences in 262
18:17
<Michael Ficarra>
I'm searching for column of Table and see 14 occurrences
18:20
<ptomato>
in NumericToRawBytes there's "Let conversionOperation be the abstract operation named in the Conversion Operation column in Table 70 for Element Type type."
18:31
<Michael Ficarra>
ah, we should fix that to use column of Table and quote the column name like the rest
18:43
<ptomato>
I think I also saw one instance of something like "column one of Table 777"
18:47
<Michael Ficarra>
I'd be very surprised if we referred to a column numerically in 262, but if we did, please open an issue.
18:56
<ptomato>
found it again, it's in CreateIntrinsics https://tc39.es/ecma262/#sec-createintrinsics
18:59
<ptomato>
https://github.com/tc39/ecma262/issues/3755
19:01
<Michael Ficarra>
thanks