gimme Work Item Rss Feed

CLOSED ISSUE: Gimme.query is broken

sstchur — Wed, 05 Dec 2007 21:15:19 GMT

Most recent checkin to fix an issue with Safari2 broken the CSS Querying Engine for other browsers
Comments: Fixed with changset #14359

CREATED ISSUE: Gimme.query is broken

sstchur — Wed, 05 Dec 2007 21:03:24 GMT

Most recent checkin to fix an issue with Safari2 broken the CSS Querying Engine for other browsers

CLOSED ISSUE: Gimme.query is slow and should be optimized

sstchur — Tue, 16 Oct 2007 20:54:30 GMT

Gimme.query is often slow when retrieving elements on a page with a large and complicated DOM structure. This is because Gimme blindly queries backwards and doesn't bother to trim the tree at all.

Ex: #a #b .dog

If there is only one .dog element, Gimme will still check all elements on the page, because it works backwards: first, retrieving all elements and weeding out those that do not have the class .dog; then walking the ancestry chain to ensure that those that do have the class .dog have ancestors #b and #a.

This could be made much more efficient by first ensuring that #b and #a exist in the DOM. If not, no need to progress further. If so, we can trim the tree by only retrieving the descendants of #b and making sure than they have the class .dog (as opposed to checking all elements).
Comments: Resolved

COMMENTED ISSUE: Gimme.query is slow and should be optimized

sstchur — Tue, 16 Oct 2007 20:54:22 GMT

Resolved in Caspian (v1.5)

CREATED ISSUE: Gimme.query is slow and should be optimized

sstchur — Thu, 06 Sep 2007 18:03:30 GMT

CREATED ISSUE: .getScreenPosition and .getPagePosition could be better

sstchur — Thu, 06 Sep 2007 17:58:42 GMT

The two functions: .getScreenPosition and .getPagePosition might be able to be improved by using some proprietary methods from IE and Mozilla.

Currently both implementations rely on the rather buggy properties of .offsetLeft and .offsetTop (browsers have sharp disagreements over what these properties should return and often have bugs regarding padding, borders, and overflow).

IE offers .getBoundingClientRect, which appears to be very accurate, though it seems to return the screen position, not the page position, so extra will be needed to ensure the proper x/y values are returned when the page is scrolled.

Mozilla offsets document.getBoxObjectFor, which also appears to be more accurate than .offsetTop/Left. However, it is not a publicly documented method. It is apparently meant for Add-on developers of Firefox, and the Mozilla devs discourage that use of this in public web sites -- special care will be needed to determine if this is a safe method to use.

CLOSED TASK: Separate "glitz" from core Gimme

sstchur — Wed, 20 Jun 2007 04:46:04 GMT

Gimme has an extension model, now that Gimme has been updated to use .prototype. This should be leveraged and all "glitzy", fx-related functions should be moved to gimmefx.js, while the core functionality of gimme will remain in gimme.js

COMMENTED TASK: Separate "glitz" from core Gimme

sstchur — Wed, 20 Jun 2007 04:45:59 GMT

This is now complete. Two files: gimme.animation.js and gimme.effects.js work together to offer "glitzy" effects to the library

CREATED TASK: Separate "glitz" from core Gimme

sstchur — Thu, 26 Apr 2007 00:29:04 GMT

COMMENTED TASK: Implement ParseComplexSelector(..)

sstchur — Thu, 26 Apr 2007 00:26:26 GMT

The regular expression used to split up complex selectors is not robust enough. Looks like this is going to need to be parsed manually. Comments: ** Comment from web user: sstchur **

Implemented basic ParseComplexSelector. It seems to work for my simple samples, but I suspect as the tests get more rigorous, this function may need additional tweaking. Marking as fixed, but not closed -- further testing is needed.

CREATED TASK: Implement ParseComplexSelector(..)

sstchur — Mon, 23 Apr 2007 21:47:04 GMT

The regular expression used to split up complex selectors is not robust enough. Looks like this is going to need to be parsed manually.

CREATED ISSUE: g('E[attr=val') broken

sstchur — Sat, 21 Apr 2007 00:29:39 GMT

Desc: Gimme by attribute selector broke during the implementation of ComplexGimme(..) Reason: ComplexGimme(..) splits the selector by \s, ~, + and |. But when spaces exist in an attribute value, ComplexGimme(..) thinks the space is a descendant combinator -- not good. Fix: Not sure yet, what the best fix for this is.