Announcing Scala.js 1.14.0. #617
Conversation
|
Needs to wait for:
|
| The linker will now avoid generating a default bridge in a subclass if an existing bridge in the superclass can be used. | ||
| This can lead to minor code-size improvements. | ||
|
|
||
| ## Call-site inline |
There was a problem hiding this comment.
Put this title and "new ECMAScript core types" under a ## Improvements?
We should also mention scala-js/scala-js#4904 for Scala 3 users.
There was a problem hiding this comment.
I've opted to put everything under a ## Improvements. This also made it possible to move default bridge generation lower (IMO less relevant for most users).
|
|
||
| ### Reachability Analysis is parallel by default | ||
|
|
||
| We have observed up to 2x speedup of reachability analysis on the Scala.js test suite with this change. |
There was a problem hiding this comment.
I'm wondering if the 2x is too marketing-y here. It is true, but most people will not care about the reachability analysis in isolation but about the whole linking process.
There was a problem hiding this comment.
Maybe mention how much of the whole linker is taken up by the analysis?
There was a problem hiding this comment.
Yes :-/ Do we have these numbers?
There was a problem hiding this comment.
Seems we do: scala-js/scala-js#4897 (comment) :) I'll run some analysis.
There was a problem hiding this comment.
> d %>%
filter(!grepl(':', op)) %>%
group_by(variant, op) %>%
summarise(t_ns = median(t_ns)) %>%
group_by(variant) %>%
summarise(t_ns = sum(t_ns))
`summarise()` has grouped output by 'variant'. You can override using the `.groups` argument.
# A tibble: 3 × 2
variant t_ns
<fct> <dbl>
1 main 1090421195
2 parallel 835999739
3 pr 866628092
So overall a 20% speedup.
There was a problem hiding this comment.
| We have observed up to 2x speedup of reachability analysis on the Scala.js test suite with this change. | |
| We have observed up to 2x speedup of reachability analysis on the Scala.js test suite with this change, which corresponds to a 20% speedup to the overall linker. |
|
This is complete now.
|
| We are pleased to announce the release of Scala.js 1.14.0! | ||
|
|
||
| This release mainly brings speed improvements in the reachability analysis by parallizing it. | ||
| TODO Mention 2.13.12 upgrade. |
|
|
||
| ### Reachability Analysis is parallel by default | ||
|
|
||
| We have observed up to 2x speedup of reachability analysis on the Scala.js test suite with this change. |
There was a problem hiding this comment.
| We have observed up to 2x speedup of reachability analysis on the Scala.js test suite with this change. | |
| We have observed up to 2x speedup of reachability analysis on the Scala.js test suite with this change, which corresponds to a 20% speedup to the overall linker. |
|
|
||
| ### Implicit conversions to js.Any for js.UndefOr in Scala 3 | ||
|
|
||
| Because `js.UndefOr` is implemented differently in Scala 3, implicit conversion to `js.Any` was not available. |
There was a problem hiding this comment.
| Because `js.UndefOr` is implemented differently in Scala 3, implicit conversion to `js.Any` was not available. | |
| Because `js.UndefOr` is implemented differently in Scala 3, the implicit conversion to `js.Any` was not available. |
gzm0
force-pushed
the
version-1.14.0
branch
from
79c5cd6 to
6c89bdd
Compare
gzm0
force-pushed
the
version-1.14.0
branch
from
6c89bdd to
7996d4e
Compare
|
No description provided.