10 web snippets
Preloading images with JavaScript
If your web app dynamically displays certain images and you don't want to make sure that the images are downloaded before they are first displayed, you can pre-fetch the images using some simple javascript.
For single-page apps, this should be sufficient:
function preload_images(urls) {
urls.forEach( function(i, url ) {
(new Image()).src = url;
});
}
preload_images( [ 'image1.jpg', 'image2.png', 'image3.tiff' ] );
If you want to add a slight delay (so other web assets can load first) use something like:
setTimeout( function() { preload_images( [ 'image1.jpg', 'image2.png', 'image3.tiff' ] ); }, 500) ;
The single-page-app method above loads each image in the array into memory. However, browsers generally won't cache these images, so if the user navigates to another page without viewing the images, they will be lost.
To make the images cachable, it helps to add the image that is created into the actual DOM tree for the page. Here's one way:
function preload_images(urls) {
var newdiv = document.createElement("div")
if(newdiv.setAttribute) {
newdiv.setAttribute("style","display:none;")
} else if(newdiv.style && newdiv.style.setAttribute) {
newdiv.style.setAttribute("cssText","display:none;")
} else if(newdiv.style) {
newdiv.style.cssText = "display:none;";
} else {
newdiv.style = "display:none;"
}
urls.forEach( function(i, url ) {
var newimg = new Image();
newimg.src = url
newdiv.appendChild(newimg)
});
document.body.appendChild(newdiv)
}
A General Purpose SQL-to-HTML Routine for CoffeeScript/JavaScript/Node.js
Using node-mysql or similar, the following CoffeeScript routine will generate an HTML table containing the data in a SQL result set, including column headings:
sql_to_html:(connection,query,bindvars,callback)->
connection.query query, bindvars, (err,rows,fields)->
if err?
callback(err)
else
buffer = '<table border=1><tr>'
for field in fields
buffer += "<th>#{field.name}</th>"
buffer += '</tr>'
for row in rows
buffer += '<tr>'
for field in fields
buffer += "<td>#{row[field.name]}</td>"
buffer += '</tr>'
buffer += '</table>'
callback(null,buffer)
Backup or mirror a website using wget
To create a local mirror or backup of a website with wget
, run:
wget -r -l 5 -k -w 1 --random-wait <URL>
Where:
-r
(or--recursive
) will causewget
to recursively download files-l N
(or--level=N
) will limit recursion to at most N levels below the root document (defaults to 5, useinf
for infinite recursion)-k
(or--convert-links
) will causewget
to convert links in the downloaded documents so that the files can be viewed locally-w
(or--wait=N
) will causewget
to wait N seconds between requests--random-wait
will causewget
to randomly vary the wait time to0.5x
to1.5x
the value specified by--wait
Some additional notes:
--mirror
(or-m
) can be used as a shortcut for-r -N -l inf --no-remove-listing
which enables infinite recursion and preserves both the server timestamps and FTP directory listings.-np
(--no-parent
) can be used to limitwget
to files below a specific "directory" (path).
Pre-generate pages or load a web cache using wget
Many web frameworks and template engines will defer the generation the HTML version of a document the first time it is accessed. This can make the first hit on a given page significantly slower than subsequent hits.
You can use wget
to pre-cache web pages using a command such as:
wget -r -l 3 -nd --delete-after <URL>
Where:
-r
(or--recursive
) will causewget
to recursively download files-l N
(or--level=N
) will limit recursion to at most N levels below the root document (defaults to 5, useinf
for infinite recursion)-nd
(or--no-directories
) will preventwget
from creating local directories to match the server-side paths--delete-after
will causewget
to delete each file as soon as it is downloaded (so the command leaves no traces behind.)
Uploading a file with curl
To submit the file at foo
to a web service as multi-part form data using curl
:
curl -X POST -F "file=@\"foo\"" 'https://127.0.0.1/example'
The file
part is the name of the corresponding form field.
Note that you can submit multiple files:
curl -X POST -F "f1=@\"foo\"&f2=@\"bar\"" 'https://127.0.0.1/example'
Or add additional body or query string parameters:
curl -X POST -F "f1=@\"foo\"&x=y" 'https://127.0.0.1/example?a=b'
Cross-browser CSS transitions
.foo {
transition: opacity .25s ease-in-out;
-webkit-transition: opacity .25s ease-in-out;
-moz-transition: opacity .25s ease-in-out;
-o-transition: opacity .25s ease-in-out;
}
Where the right-hand side has the general form:
PROPERTY DURATION [TIMING-FUNCTION] [TRANSITION-DELAY]
Transitions can chained:
.foo { transition: opacity .25s ease-in-out, height 1s linear }
Transition timing functions
linear
- constant rate of change between states.cubic-bezier(0.0, 0.0, 1.0, 1.0)
ease
- (the default) slow acceleration, then faster, before rapidly slowing again at the end.cubic-bezier(0.25, 0.1, 0.25, 1.0)
ease-in-out
- like ease but accelerating/decelerating more rapidly (with a shorter transition between acceleration and deceleration).cubic-bezier(0.42, 0, 0.58, 1.0)
ease-in
- equivalent to the first half of ease-in-out; rapid accelerating then transitioning to a constant rate of change at the end.cubic-bezier(0.42, 0, 1, 1.0)
ease-out
- equivalent to the second half of ease-in-out; a constant rate of change transitioning rapid deceleration at the end.cubic-bezier(0.42, 0, 0.58, 1.0)
cubic-bezier(x1,y1,x2,y2)
- follows cubic Bézier curve using the control points (0,0), (x1,y1), (x2,y2) and (1,1).steps( n, [start|end] )
- stepwise function with n steps
How to read cubic-bezier
functions
Imagine the transition as a stepwise function starting at (0,0) and ending at (1,1) and with two steps in between. The cubic-bezier(x1,y1,x2,y2)
function specifies the points in the middle. In other words, x1 and x2 are the times at which the second and third steps happen, respectively (expressed as a fraction of the total transition duration) and y1 and y2 are how far along the transition is at x1 and x2, respectively (expressed as a fraction of the total transition "distance"). Very loosely, the cubic-bezier function "smooths" those steps.
These are nicely demonstrated here.
Presentational CSS Classes
These are the opposite of semantic markup, but I find them useful:
.align-both { text-align: justify; }
.align-center { text-align: center; }
.align-left { text-align: left; }
.align-right { text-align: right; }
.float-left { float: left; }
.float-right { float: right; }
.nobr { white-space: nowrap; }
.small { font-size: 80%; } /* should really sync with the <small> tag rules */
Fixing CSS Resets
CSS reset frameworks often strip out all formatting. These CSS rules contain some re-resets:
/* I understand the theory behind replacing <i> and <b> with <em> and <strong>, but c'mon, really? */
i { font-style: italic; }
b { font-weight: bold; }
small { font-size: 80%; }
/* Make superscripts and subscripts actually do something: */
sup, sub { height: 0; line-height: 1; vertical-align: baseline; _vertical-align: bottom; position: relative; }
sup { bottom: 1ex; }
sub { top: .5ex; }
/* Mono-spaced types. */
pre, code, kbd, samp, tt { font-family: 'droid sans mono slashed', 'droid sans mono dotted', 'droid sans mono', monospace, monospace; }
Vertically centering block elements with CSS.
Via phrogz.net.
<style type="text/css">
#wrapper { position:relative; } /* position:absolute is also ok */
#wrapped { position:absolute; top:50%; height:10em; margin-top:-5em; } /* margin-top = -1 * height/2 */
</style>
...
<div id="wrapper">
<div id="wrapped">
<p>Block content.</p>
<p>But still vertically centered.</p>
</div>
</div>