Make [emoji] implementation deterministic #103
Conversation
|
|
There was a problem hiding this comment.
Actually I just caught something with the help of @sokra "over my shoulder".
Instead of the .9999 approach, it needs to be more like https://github.com/webpack/loader-utils/blob/master/lib/getHashDigest.js where emoji is another base in addition to existing ones there.
Let me know if that makes sense @kohlikohl
|
@TheLarkInn I think that makes sense. I'll take another stab at it. |
kohlikohl
force-pushed
the
deterministic-emojis
branch
2 times, most recently
from
dcb56b2
to
43f0889
Compare
|
@TheLarkInn @sokra Pushed an update. Is this what you had in mind? |
kohlikohl
force-pushed
the
deterministic-emojis
branch
2 times, most recently
from
7ee5269
to
88497a3
Compare
lib/getHashDigest.js
Outdated
| return emojis.splice(0, maxLength).join(""); | ||
| } | ||
|
|
||
| function sliceOutput(output, digestType, maxLength) { |
There was a problem hiding this comment.
To simpily this could you:
- make
encodeBufferToBasereturn an array. - Use
Array.prototype.sliceinstead ofString.prototype.substringto ensuremaxLength - Convert array to string with join after that.
- Remove sliceEmojis, sliceOutput, emoji-regex
There was a problem hiding this comment.
As for the emoji-regex package. I'd encourage to keep it as it is better in detecting the different code points of emojis. Here is an example https://runkit.com/kohlikohl/emoji-regex.
The simple regex seems to only detect emojis that have two code points. Flags for example have four so it would miss identify and split them.
Maybe it is not relevant in this case, if so let me know and I'll make the final change.
lib/getHashDigest.js
Outdated
| @@ -36,15 +43,35 @@ function encodeBufferToBase(buffer, base) { | |||
| return output; | |||
| } | |||
|
|
|||
| function getBase(digestType) { | |||
There was a problem hiding this comment.
To simpily this could you:
- Add a
encodeTableargument toencodeBufferToBase - Add a
encodeBufferToDigestTypewhich callsencodeBufferToBasewith the correct arguments according todigestType. - Remove emojis from the
baseEncodeTables(looks weird)
lib/getHashDigest.js
Outdated
| digestType === "base49" || digestType === "base52" || digestType === "base58" || | ||
| digestType === "base62" || digestType === "base64") { | ||
| return encodeBufferToBase(hash.digest(), digestType.substr(4)).substr(0, maxLength); | ||
| if(digestType && !!baseEncodeTables[getBase(digestType)]) { |
There was a problem hiding this comment.
Maybe add a const array of valid digestTypes?
There was a problem hiding this comment.
This was the idea with usingbaseEncodeTables as a const to not have to repeat the digestTypes. Lets see 😄
lib/getHashDigest.js
Outdated
| const emojisList = require("emojis-list"); | ||
| const emojiRegex = require("emoji-regex")(); | ||
|
|
||
| const BASE_EMOJI = "baseEmoji"; |
|
Requested changes pending, otherwise this is looking great so far. Excellent work @kohlikohl |
kohlikohl
force-pushed
the
deterministic-emojis
branch
from
88497a3
to
5be0725
Compare
…a deterministic way
kohlikohl
force-pushed
the
deterministic-emojis
branch
from
5be0725
to
c60ec70
Compare
|
Pushed the changes based on the comments. Please have another look @TheLarkInn @sokra As mentioned in the response to the comments, I'd encourage to keep the |
|
When this PR will be merged? :) |
|
@TheLarkInn @sokra Are you guys happy with this PR? Is it ready to be merged in? |
|
Would love to see this merged. @kohlikohl can i suggest fixing the coverage regression, as maybe the lack of a ✅ on the PR checks is halting the approvals? |
|
/cc @kohlikohl sorry for big delay, can you rebase feature? |
|
No valid anymore |
PR for issue #102.
Make use of hashing function to choose an emoji instead of randomly selecting one.
Fun fact: Not that this would be needed much but it supports 1k+ emojis per hash. 🚀