msiebuhr / charcod.es Goto Github PK

We're beginning to have some quite complex code in the client. Quite a lot of it doesn't touch the DOM, so we should be able to unit-test it before we build/deploy changes.

Stuff I'd like tested:

• Core search engine (http-pub/js/search.js)
• Bits in http-pub/js/app.js: isCharCode(), Bram's unicode-surrogate/BMP-stuff, data parsing/loading/indexing, location.hash-handling, etc.

I imagine splitting the code into separate modules that works both in browsers and Node.js, and then do the testing there.

The application halts slow machines/devices quite a lot, so perhaps we should split the index into smaller parts - possibly with various-size chunks, so we can adapt to faster/slower machines and network connections. Eg. naming data-FROM%-TO%.json:

#5% chunks
data-0-5.json
data-5-10.json
…

#10 % chunks
data-0-10.json
data-10-20.json
…

#25 % chunks
data-0-25.json
data-25-50.json
…

#50 % chunks
data-0-50.json
data-50-100.json

Then the client could start out downloading data-0-10.json and parse it. If that takes to long, degrade to 5% chunks, and if It's fast, upgrade to 25%-chunks.

We'd have to have some more data lying around (about 2MB per size), and - more difficult - figure out a dynamic download client.

What about a command-line client in plain-ish node.js:

unicodefinder :)
☺
unicodefinder -- -->
→ ↔

Hi,

We do like arrows but .. :

Are the two links wanted ?

http://twitter.github.com/bootstrap/javascript.html#modals

The defer(url, callback) doesn't insert the DOM-element for fetching the script in the DOM. Specifically, the following line seem to be broken:

firstScript.parentNode.insertBefore(inc, firstScript);

Can you look into it, @Munter?

Try wrapping them in zero width non joiner:

http://en.wikipedia.org/wiki/Zero-width_non-joiner

This made them show up in most cases.

There's really no reason to load the data as JSON, as it's quite static. Wrapping it into some JS should be fairly simple:

"var rawData = " + JSON data + ";"

And then include it in the HTML.

I'm thinking something with a few clear (but muted/saturated) colours, to go along with a new logo per #47.

The simplicity of what the site does, I think something material-like would be nice. And implementable without pulling in a ton of unused CSS scaffolding.

Searches for numbers, html entities, \u-thingies and so forth should give results...

Someone at Penn State University has written up a nice list of names/unicode-points, which we should blatantly steal.

See http://cssarrowplease.com/ for example.

https://charcod.es/ works, but complains about the certificate being invalid. It's issued to *.netlify.com -- I thought the site was hosted on github pages?

I ran into this because I'd like to replace the appcache manifest with a generated service worker, but that'll only work with a secure origin.

https://en.wikipedia.org/wiki/Regional_Indicator_Symbol

Would break a whole lot of assumptions in the code, but might be worth it.

Should make it modular, enable sane testing (see #30 and #37).

I have the initial work laying around on a branch somewhere, but it needs to be merged (and the core app should support AMD as well; see #40 )

We wouldn't want to look like a http://charcod.es/#128053

For those of us no longer spending time in Unixes w. mark-to-copy....

Show the name of the page as a split flap display, randomly changing a letter in the title for a unicode look-alike. (Or, perhaps, a similar visual effect.)

The letters should probably be pre-picked so hey look alike and work on most platforms.

I'm thinking along the lines of https://www.iconfinder.com/icons/174048/mark_question_icon or https://www.iconfinder.com/icons/221269/dictionary_icon, but I'm up for whatever.

Twitter has a nice collection of emojis that seem to be updated very quickly after each release. Perhaps we could show those when available?

See https://github.com/twitter/twemoji

It's fun to find & share new characters - let's add some kind of permalinks so it's easy to share!

(Also, I should probably buy a dedicated domain for it!)

We can link from codes to Wolfram Alpha to let users' explore more.

The query-url looks like this:

http://www.wolframalpha.com/input/?i=unicode+DECIMAL_CODEPOINT

Request

There is a project similar in nature to this here:

http://unicod.es/

Other visitors need to know about it, so perhaps a link in the footer will do?

Someone from the Unicode consortium has looked into what's available as equivalent unicode chars: http://unicode.org/~asmus/web-wing-ding-ext.pdf.

There's also quite a few pages there and there, that has crafted similar lists. We could possibly scrape one of those to get an alias-map.

When searching for, example foot, some chars flow to the next line, putting some chars between the "tab" and the box showing the details.

Broken with Chrome 22 on Ubuntu 12.04 when the window is less than 970 px wide (per window.outerWidth).

When a user searches and - predictably - presses the enter-key - the page navigates to itself, clearing the users' query.

Thanks, @bramstein.

Try searching for AKTIESELSKAB vs. aktieselskab.

This seems to be a problem with how the URL fragment is formed. I'm not sure there's a good way of fixing it, but perhaps a / could be encoded in the fragment in some way in order to avoid this issue?

Emoji-list published 2018-02-17...

Netlify is a freemium static site hosting service, that can take care of building, serving (including HTTPS) for us.

Reasons to switch:

• Get HTTPS. GratisDNS + Github Pages doesn't do that, and we'll need it to get some of the new web features to work
• The current build-pipeline is quite fragile and I keep forgetting how it works.

Reasons for Netlify:

• Cloudflare + gh-pages would solve the HTTP-problem, but leave the build-pipeline. Also, CF requires that their DNS servers take care of charcod.es, which GratisDNS doesn't allow us to do without paying some rather hefty fees.
• Netlify runs a build-pipeline that magically does the right thing for us

TODO

• Switch to Netlify
• Get HTTPS working
• Clean up old build-pipeline
• Do a release-branch for Netlify to track.

Add an OpenSearch description, so browsers (/others?) can easier use our site.

Things should be usable on a mobile phone of 320px width as well.

I'll get on that right away

When ex. searching for ->, it shows up URI-encoded in the search-bar.

Tried copying http://charcod.es/#hand/128406 into the search field and go no results

Should use assetgraph-builder.

Soft hyphen should be code point U+00AD in the Latin-1 Supplement, but I am unable to find it on charcod.es

'nuff said.

The unicode specification is available as a small tonne of PDF-files at http://unicode.org/charts/.

There seem to be some naming-scheme, but I can't quite figure out the offsets - but if we could do this automatically, it should be quite easy to add.

Unicode now allows adopting characters. Perhaps we should show them as well?

https://unicode.org/consortium/adopted-characters.html

Unfortunately, there does not seem to be a machine-readable list anywhere...

Currently there is an inconsistent rendering of charcodes across platforms.

If there is a webfont available with a wide range of supported unicode chars, then that might be a better choice to give a more pleasant experience for all users.

We might even be able to switch fonts for certain blocks where there is missing support in our preferred font.

W3C Has an XML file with HTML/XHTML/LaTeX/... aliases for what looks like most of unicode: http://www.w3.org/2003/entities/2007xml/unicode.xml

Don't import non-printable chars from Unicode.

Ex. the new power-symbol (see http://unicodepowersymbol.com/we-did-it-how-a-comment-on-hackernews-lead-to-4-%C2%BD-new-unicode-characters/) has aliased some old codes to have power-related meanings - but charcod.es doesn't capture that...