Bug Tracker

Custom Query (13852 matches)

Filters
 
Or
 
  
 
Columns

Show under each result:


Results (25 - 27 of 13852)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Ticket Resolution Summary Owner Reporter
#29 fixed TortoiseSVN breaks on jQuery SVN john
Description

Mike, I did some more digging and I think I've resolved the problem (yet again). It had to do with a 'mesg y' being in the global bashrc file. Since removing it, it no longer throws that error. Try again and let me know what you find.

Alas, no luck yet, it's working the same as before. I can check out code using an SSH tunnel through SecureCRT, but when I try to commit a change, Tortoise prompts me for a username and password, and no variation of my jquery.com username and password seems to work.

If I don't use the SecureCRT SSH tunnel but use the svn+ssh: URL instead, I can't even check out code. TortoisePLINK prompts me for a password but doesn't accept it.

We can take this offlist unless anybody else is interested...

Thanks for giving it a shot anyway! :-)

-Mike

#30 fixed .pushStack() leaks memory [email protected]
Description

When someone uses the remove(), find(), parent(), parents(), siblings(), not(), filter(), or add() core methods, those methods push the previous node list onto the .stack property. Chaining several of these methods will push several node lists onto the stack. The user can retrieve those nodes using the end() method, but has no simple way to retain the .cur nodes but dump the stack. This may cause memory leaks in general usage, because many users will not be aware of the stacking. For example, this sets a variable in global scope:

exts = $("a").css(fontWeight: "bold").find("[rel='external']"].css(color, "red");

The jQuery object exts ends up with a pushed nodelist of all anchor objects that the user probably doesn't want, but they did want to keep the rel=external links.

I like the stack functionality, but not when it's inside core methods. I would suggest taking out the "hidden" pushStack() calls in the core, renaming pushStack/end to something symmetrical like jpush/jpop, and letting the user explicitly decide when to push or pop the stack in situations where they know the want to keep the nodes before doing further filtering or unhooking them from the DOM tree.

A small performance issue as well, I know that shift/unshift is much more expensive than push/pop for maintaining an array stack in IE Javascript. Given that the stack depth will be shallow it doesn't make much difference, but push/pop is shorter to type too. :)

#31 fixed .remove() should be usefully chainable john
Description

Right now this.cur is set to an empty set, but it should probably be left as the existing set for future operations (such as appendTo).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Note: See TracQuery for help on using queries.