alexcorvi / anchorme.js Goto Github PK

Now that the library have exceeded 100 stars I think it should be hosted on cloudflare.com CDN.

When using th eplugin with twitter content and youtube URLs, it got broken when id are all transformed to lowercase.

Issue:
When text is run through the anchorme method multiple times it will nest links within the existing links.

Expected result:
Parser recognizes that the link is already within an anchor tag and does not generate a nested link.

Steps to recreate:
Run this text through anchorme:
Initial: http://google.com
1st render: <a href="http://google.com" >http://google.com</a>
2nd render: <a href="http://google.com" ><a href="http://google.com" >http://google.com</a></a>

There seems to be an issue by having an anchor tag with the URL as its content. We're using anchorme with a rich text editing library, which insert an anchor html element with the content being a URL string:
<a href="http://www.google.com">http://www.google.com</a>

Anchorme will place another tag inside the existing tag. This will recurse each time the string is parsed.

I realise that the library's core function is to find URL strings and wrap them, so this might not be the best use for it or even an issue really, but is there a (quick) workaround for this issue?
Maybe an option like ignoreUrlsInTags could be added?

Add the i flag to the regular expression detecting IP patterns here

I am using anchrome.js and trying to open link in a new tab.
While trying using attribute feature, it's not working:

anchorme("screencast.com", {
attribute: {
"target": "_blank"
}
})
above code producing following: "http://screencast.com"

For reference:
http://screencast.com/t/206XmVpSd9LN

Does the library sanitize HTML to protect from cross-site scripting (XSS) attacks??

Is there any way to use this without having to convert the original variable to a anchorme.js variable? Say for example when using angularJS and jg-repeat, is there any way to use it on text as it's being displayed?

If you input this:
(text text..) text text text text text,
you get this
(text text..)texttext text text text

This is coming from a tweet. Shouldn't this be linked?

What's the best way to clean your smartphone? 📱🚿https://t.co/cxjsA6j60J

Should be able to truncate links like this "http://www.google.com/abcedfghijklmnopqstuvwxyz" to this: "http://www.google.com/abcedf..."

E.g. www.🍕💩.ws - GitHub does it correctly!

Currently there is no config which allows one to customise created a tags. It would be great if we have support for adding class and other attributes like target which are supported on a tags.

As per subject. Thoughts?

Currently the library doesn't support decimal IPs. such as 2130706433 (which is basically 127.0.0.1).

I can't really imagine a real world use case for this library where the decimal IP detection would be needed. So if implemented, I might disable it by default, and/or enable it only when port is defined.

The implementation however is rather simple:

function decimalToIP(decimal:number):string{
     return ((decimal>>24)&0xff) + '.' +
     ((decimal>>16)&0xff) + '.' + 
     ((decimal>>8)&0xff) + '.' +
     (decimal&0xff);
}

Hi,

Thanks for writing this script. I'm noticing that you are converting all urls to lowercase. This causes problems with links that have case sensitive servers or base64 hashes such as: http://codepen.io/sdthornton/pen/wBZdXq

Tested against Autolinker.js and linkifyjs on FireFox. After 10000 iterations anchorme.js scored 300ms while others where around 1200

Would love to see the tests, and what the performance differences are in Chrome.

(I've added this library to my StackOverflow answer on replacing plain URLs with strings).

The tests should extensions, with few builtin extensions like url, email, and ip.

This will allow projects to use anchorme for all their anchor-linking tasks such as hashtags and mentions ... etc

According to source, option should be plural (emails). Please consider updating the documentation (readme file).

Also, there's an option "TLDs" which is not explained (should it be?)

As per subject. Thoughts?

After refractoring to typescript lot of things changed on how the library splits up text, identifies link and skips HTML.

I forgot to remove some lines that don't do anything anymore:

• skipping html
• use strict

Hi,

I have a use case with the following string:

http://tfs:8080/_ab/

It is not highlighting without a . in the hostname, but I think it should highlight if it detects the http:// protocol part

the following urls should be valid, but are not:

http://localhost
http://chewbacca
http://chewbacca;80
http://chewbacca:80/test

I used the embedded anchorme on the anchorme.js documentation site to test this.

Many people add url at the end of their sentence and then the url ends in a period, such as http://codepen.io/sdthornton/pen/wBZdXq.

The period then becomes part of the link and the link won't work. While the '.' is a valid character in URLs, it is not part of the allowed base64 set and I suspect that very few actual urls will end in a period.

Hi,

It would be nice if the library interpreted "localhost" as a valid URL: http://localhost I am asking because we use localhost to test our web app, and it would be nice to have the links functional.

Is it possible to do it right now, with an option?

Merc.

In my project, user writes text in a box, but behind the scenes,

tags are inserted (it applies to

anchorme.js('www.google.com')
"www.google.com"
anchorme.js('
www.google.com
')
"
www.google.com
"
anchorme.js('
 www.google.com 
')
"
 www.google.com 
"
anchorme.js('
google.com
')
"
google.com
"
anchorme.js('
 google.com 
')
"
 google.com 
"

[email protected],

the comma at the end gets included in the result

After updating, the node distribution does not servers a one-file package thus I need to move edit JSFiddle snippets to have a rawgit link instead of unkpg.

The IP detection pattern ends with (\/.*)?.

It's better to add allowed charecters like the URL

When I enter "file:///folder/test.txt" in the demo the generated href is "file:///file:///folder/test.txt"

Eg: yojna...oracle, care...to, core...so

Above provided examples are getting converted into link.

Hi,

We are facing issues with files converted to links. For example:
download.zip is getting converted to: <a href="http://download.zip" >download.zip</a>

We came across with property, files:false but it seems doesn't work:
jsFiddle

Please help.

The email pattern regex and the URL pattern regex share some allowed characters. Maybe it's best to put those characters in the lists file.

running 'en.wiktionary.org/wiki/銀' through anchorme works fine, however if you go to that link in a browser and copy and paste it the actual string is already encoded:
'https://en.wiktionary.org/wiki/%E9%8A%80'
If you run the latter through anchorme, the link ends up as
'https://en.wiktionary.org/wiki/%25E9%258A%2580'

Try entering the following value: 12.12.12.
You'll get the following error:

Hi,

Is it possible to convert the url with space in the parameter?
eg: www.google.com?key=add value&key2=value

I tried to replace space with %20 www.google.com?key=add%20value&key2=value
the outcome for the url become
www.google.com?key=add%2520value&key2=value

Hi,

"http://www.google.com/?aQueryParam=blah" works fine but removing the last backslash makes it not converted to url.

Thanks,
Elaine

Default settings appear to output a console.log upon every parse. I have edited my local version but it would be nice if it didn't output these logs or if there was an option for them.

Hi Alex,

I'm using emojis which are included with an <object> tag. For some reason anchorme does not detect this as html and does mistakenly process the url within the data attribute.

Lorem ipsum <object class="emoji" data="https://www.mydomain.com/svg/emoji/123.svg" type="image/svg+xml" standby=":)">:)</object>

I suppose it has something to do with data="https... because when I change it to src="https..., the <object> tag will be recognized and not break any more - but of course then the emoji will not show up.

Thanks in advance for any help :)

Is it no longer an option to assign a target attribute to the anchor? It no longer appears in the source or in the example options.

For example, it identifies (www.google.com) and [[www.google.com]] no problem but doesn't seem to parse "www.google.com" or 'www.google.com' correctly. Not sure if I am doing something wrong.

Update IANA list. I might also setup an automated script that retrieves the list every week and updates the package automatically.

Is custom link text supported?

Hi,

How to use it with Node/npm? require('anchorme.js') always throws Error: Cannot find module 'anchorme.js'. I assume, this is because the name ends with .js.

Node version: v6.4.0

Some of the documents on our file server have spaces in their URL, e.g.

https://mydomain.com/My%20document.pdf

An URL with a percent sign is not recognized by anchorme.

e.g.

> anchorme.js("https://mydomain.com/My%20document.pdf");
< "https://mydomain.com/My%20document.pdf"

> anchorme.js("https://mydomain.com/My20document.pdf");
< "<a href='https://mydomain.com/My20document.pdf'>https://mydomain.com/My20document.pdf</a>"

Hi there,

Love the truncation option, but would it be possible even to have the ellipsis in the middle of the cut down URL (e.g. http://www.go...bla=blub&nice=1) instead of at the end, and maybe even drop the protocol from it too? All in order to make the shortened version even more useful and indicative of the actual link...

Cheers!

Not converted:
"Do a search (go here: google.com)"

Converted (added an extra space after google.com):
"Do a search (go here: google.com )"

Wouldn't work on IP addresses with ports defined.

e.g:
127.0.0.1:3000

unless a protocol defined:
ftp://127.0.0.1:3000

Hi,

If we apply target= "_blank" in the attributes, we are getting target attribute added even for text converted to mailto:// .

for example:

anchorme("drop an email to [email protected]", {
                    attributes: {
                        "target": "_blank"
                    }
                }) 

Output:"drop an email to <a href='mailto:[email protected]' target='_blank' >mailto:[email protected]</a>"

Problem: It is unnecessarily opening a blank window on click.

Should there be someway, we can ignore the adding target attribute for email?

-Sumeet

It's not clear in the documents, but if I pass a string, can I get an array of valid URLs returned (much like a str.match(...regex...)?

Thanks!