General update to drop getTable() in favor of blob()
#47
Conversation
oyiptong
changed the title
getTable() in favor of blob()
|
This doesn't build for me: |
index.bs
Outdated
| 1. Let |blob| be a new {{Blob}} whose contents are |table|'s [=font table/data bytes=] and {{Blob/type}} attribute is \``application/octet-stream`\`. | ||
| 1. Append the [=ecma-record|Record=] { \[[Key]]: |tag|, \[[Value]]: |blob| } to |backing|.\[[MapData]]. | ||
| 1. [=/Resolve=] |promise| with |map|. | ||
| 1. Let |blob| be a new {{Blob}} whose contents are |metadata|'s [=font representation/data bytes=] and {{Blob/type}} attribute is \``application/octet-stream`\`. |
There was a problem hiding this comment.
metadata is referenced for the first time here. I think it's support to be [=this=]? But that also suggests that [=font representation/data bytes=] is in the wrong place, and instead it should be a child of FontMetadata.
There was a problem hiding this comment.
I thought the explanation of [=font representation/data bytes=] as being concept/context. And the link between Font Metadata and font representation is made with this sentence:
Each FontMetadata has an associated font representation.
font representation, I thought was important to stand on its own because it delved into the concept of tables and specifically, the name table, to explain where postscriptName, fullName, and family come from.
Do you have thoughts on how this could flow better?
I'm working on making a change to specref to include SFNT. |
index.bs
Outdated
| * Be available from Workers | ||
| * Allow multiple levels of privacy preservation; e.g., full access for "trusted" sites and degraded access for untrusted scenarios | ||
| * Reflect local font access state in the Permissions API | ||
| * Restrict access to local font data to Secure Contexts | ||
| * Provide unique identification of families and instances (variants like "bold" and "italic"), including PostScript names | ||
| * Provide access to all browser-allowed font tables (may vary per browser) | ||
| * Shield applications from unnecessary complexity by requiring that browser implementations produce valid OpenType data in the returned data |
There was a problem hiding this comment.
We're not actually requiring/implementing this, are we? i.e. in Chrome we're not relying on OTS?
There was a problem hiding this comment.
Correct, we're not relying on OTS. That said I want to express that the byte data should be valid SFNT.
WDYT of switchingOpenType for SFNT in this sentence?
specref change here: tobie/specref#629 |
This PR updates the spec and explainer with information closer to the proposal we want to make.
Namely, the spec drops references to
getTable(), which is no longer part of the spec that we're proposing, in favor ofblob().This also introduces the concepts necessary to explain font data, namely the
SFNTcontainer.The explainer has been updated to contain current information.
Fixes #27
Fixes #37
Fixes #40