[labs/ssr-client] Add digest cache #5074
Conversation
🦋 Changeset detectedLatest commit: 6a1a879 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
| const digestSize = 2; | ||
|
|
||
| // Digest cache | ||
| const digestCache = new WeakMap<TemplateStringsArray, string>(); |
There was a problem hiding this comment.
Let's make sure that we're not exacerbating the static template memory leak with this...
I think we're not because once the static template cache frees a template, the forged TemplateStringsArray will be able to be freed and this cache will release the entry.
There was a problem hiding this comment.
I believe you're correct based on my understanding of how WeakMaps work in JS
There was a problem hiding this comment.
Once an object used as a key has been collected, its corresponding values in any WeakMap become candidates for garbage collection as well — as long as they aren't strongly referred to elsewhere.
-MDN
So as long as the template is only associated with WeakMaps (or I'm assuming other "weak" data structures), we should be good.
Co-authored-by: Justin Fagnani <[email protected]>
| // for `Buffer` from Node's built-in `buffer` module in our Rollup config (see | ||
| // note at the top of this file), and use that. | ||
| return NODE_MODE ? Buffer.from(str, 'binary').toString('base64') : btoa(str); | ||
| digest = NODE_MODE |
There was a problem hiding this comment.
btw, we should start using btoa in Node soon...
This PR adds a cache for previously computed digest results. This should improve throughput when rendering the same
TemplateResultmultiple times.