Clean up text you copied from somewhere else

Copied text carries passengers. Word replaces your quotes with curly ones and your hyphens with dashes. Web pages contribute non-breaking spaces and zero width characters that are invisible until they break something. Copied links arrive with a tail of tracking parameters describing where you found them.

Most of the time this is cosmetic. Occasionally it is not. A zero width character pasted into a password field, a filename or a configuration value produces a failure with no visible cause, and people lose hours to it because the thing on screen looks exactly right. Smart quotes pasted into code break it outright, which at least fails honestly.

The link tracking is the part worth more attention than it gets. Those parameters on a copied URL describe where you found it and often which campaign or which person's share it came from. Passing that on in an email or a document forwards information about your own browsing to whoever receives it, and appends noise to a link that would work perfectly without any of it.

There is a house reason this applet exists at all. Pivotal Path does not use em-dashes, anywhere: not in copy, not in commit messages, not in code comments. It is a standing rule, it is the most frequently broken one, and the usual culprit is text that came from somewhere else carrying a dash that nobody typed. Running pasted text through something that normalises punctuation catches it before it reaches a published page.

The general principle underneath is worth stating: text copied from a rich environment into a plain one should be normalised deliberately rather than hopefully. Paste as plain text is the keyboard shortcut version, and does most of the same job, though it will not touch tracking parameters or tell you what it removed.

Everything happens in your browser. Whatever you are pasting stays with you, which given that people paste passwords and half written emails into anything with a text box seems worth being firm about.