Announcing Scala.js 1.3.1. #523
Conversation
|
|
||
| Scala.js code has been known for bad performance on Firefox with the default output mode of ECMAScript 2015. | ||
|
|
||
| Long story short: this release fixes this issue. |
There was a problem hiding this comment.
I think the long story short should at least mention that this may cause a size penalty. Especially because it is the default (so somebody who didn't fiddle with their linker options at all will see this).
There was a problem hiding this comment.
I've put the paragraph about the code size before the long story. Is it better?
There was a problem hiding this comment.
So so TBH. I would stop reading at this line if I didn't care about Firefox performance and assume I don't need to care. So I think it should contain enough for somebody to realize that they need to read on, if they see size regressions in their output.
This is IMO the strict minimum, since it assumes a user noticed the size regressions and is looking for the cause. If they do not notice, we might simply boost their JS size without them noticing (immediately); potentially leading to a long, painful debugging story that could have been avoided.
So thinking about this some more, this probably even warrants a sentence in the introductory text.
|
|
||
| Scala.js code has been known for bad performance on Firefox with the default output mode of ECMAScript 2015. | ||
|
|
||
| Long story short: this release fixes this issue. |
There was a problem hiding this comment.
So so TBH. I would stop reading at this line if I didn't care about Firefox performance and assume I don't need to care. So I think it should contain enough for somebody to realize that they need to read on, if they see size regressions in their output.
This is IMO the strict minimum, since it assumes a user noticed the size regressions and is looking for the cause. If they do not notice, we might simply boost their JS size without them noticing (immediately); potentially leading to a long, painful debugging story that could have been avoided.
So thinking about this some more, this probably even warrants a sentence in the introductory text.
| This is primarily a bugfix release, including fixes for three regressions in 1.3.0 ([#4252](https://github.com/scala-js/scala-js/issues/4252), [#4268](https://github.com/scala-js/scala-js/issues/4268) and [#4271](https://github.com/scala-js/scala-js/issues/4271)). | ||
| It also contains some improvements to the JDK support. | ||
|
|
||
| More importantly, this release fixes the performance issues of the ECMAScript 2015 mode (the default) on Firefox. |
There was a problem hiding this comment.
Re-reading this, it feels contradictory with the previous paragraph: If this is primarily a bugfix release, then why is it more important that it fixes a performance issue on FF?
IMO the performance fix should go first, as it is likely to affect the largest fraction of users unexpectedly (folks that care about regressions or new features are likely actively looking for them in the release notes).
There was a problem hiding this comment.
(re-re-reading this, I realize that a performance fix can also be considered a bugfix... 🤷 )
|
I've reworked the introduction and the section on performance/code size to make it much more prominent that code size is affected. |
|
Thank you for bringing attention to the important issues ;) |
No description provided.