Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
Jörn Zaefferer schrieb: > Klaus Hartl schrieb: >> Brian Miller schrieb: >> >>> Is this patch generalized enough to add to the core? Just a thought... >>> >> Brian, it's actually just a check if the object is not null before >> accessing it: >> >> As it is: >> ret = >> document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); >> >> >> The patch (with Dave's proposal to avoid two calls): >> >> var computedStyle = document.defaultView.getComputedStyle(this, null); >> ret = computedStyle && computedStyle.getPropertyValue(prop); >> >> I think that one does no harm. ret may be null in certain cases in >> Safari but that's better than throwing a nullpointer, right? >> >> It could also return an empty string if that's more robust: >> ret = computedStyle && computedStyle.getPropertyValue(prop) || ''; >> > Could you file that as a bug report if you think it should be fixed in core? I worked around that bug at Plazes, so I have an unpatched jQuery now. Nonetheless I have created a ticket because I think it makes sense to fix this: http://jquery.com/dev/bugs/bug/467/ You can decide on this. -- Klaus ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
Klaus Hartl schrieb: > Brian Miller schrieb: > >> Is this patch generalized enough to add to the core? Just a thought... >> > > Brian, it's actually just a check if the object is not null before > accessing it: > > As it is: > ret = > document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); > > > The patch (with Dave's proposal to avoid two calls): > > var computedStyle = document.defaultView.getComputedStyle(this, null); > ret = computedStyle && computedStyle.getPropertyValue(prop); > > I think that one does no harm. ret may be null in certain cases in > Safari but that's better than throwing a nullpointer, right? > > It could also return an empty string if that's more robust: > ret = computedStyle && computedStyle.getPropertyValue(prop) || ''; > Could you file that as a bug report if you think it should be fixed in core? -- Jörn Zaefferer http://bassistance.de ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
Brian Miller schrieb: > Is this patch generalized enough to add to the core? Just a thought... Brian, it's actually just a check if the object is not null before accessing it: As it is: ret = document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); The patch (with Dave's proposal to avoid two calls): var computedStyle = document.defaultView.getComputedStyle(this, null); ret = computedStyle && computedStyle.getPropertyValue(prop); I think that one does no harm. ret may be null in certain cases in Safari but that's better than throwing a nullpointer, right? It could also return an empty string if that's more robust: ret = computedStyle && computedStyle.getPropertyValue(prop) || ''; -- Klaus ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
Is this patch generalized enough to add to the core? Just a thought... - Brian > John Resig schrieb: >> Unfortunately, the hack to work around that is too monstrous to >> comprehend. For now, it works fine in Safari for most cases, and >> that's what matters. > > That is exactly the problem I have. So I will have to live with my > patched version for Plazes? > > > -- Klaus ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
John Resig schrieb: > Yeah, this has been fixed for a while now. The one case that isn't > accounted for (yet) is having a parent element that has a display set > to none - that kills computedStyle for all sub-elements. > Unfortunately, the hack to work around that is too monstrous to > comprehend. For now, it works fine in Safari for most cases, and > that's what matters. That is exactly the problem I have. So I will have to live with my patched version for Plazes? -- Klaus ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
Yeah, this has been fixed for a while now. The one case that isn't accounted for (yet) is having a parent element that has a display set to none - that kills computedStyle for all sub-elements. Unfortunately, the hack to work around that is too monstrous to comprehend. For now, it works fine in Safari for most cases, and that's what matters. --John On 11/29/06, Klaus Hartl <[EMAIL PROTECTED]> wrote: > Hi all, > > A while ago I reported a bug, which I couldn't really narrow down, so it > never went into a ticket: > http://jquery.com/discuss/2006-August/010401/ > > The problem was that in line 1315: > > ret = > document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); > > document.defaultView.getComputedStyle(this,null) returns null causing an > error in Safari obviously. > > I fixed it with the following: > > ret = document.defaultView.getComputedStyle(this,null) && > document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); > > A few minutes ago I stumbled upon a post, which exactly explains that bug: > http://snook.ca/archives/javascript/safari2_display-none_getcomputedstyle/ > > So I wonder if we should merge the fix into jQuery finally? > > > -- Klaus > > ___ > jQuery mailing list > discuss@jquery.com > http://jquery.com/discuss/ > ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
Re: [jQuery] Bug: "display: none" plus getComputedStyle in Safari
> A while ago I reported a bug, which I couldn't really > narrow down, so it never went into a ticket: > http://jquery.com/discuss/2006-August/010401/ > > The problem was that in line 1315: > ret = document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); > > document.defaultView.getComputedStyle(this,null) returns > null causing an error in Safari obviously. > > I fixed it with the following: > > ret = document.defaultView.getComputedStyle(this,null) && document.defaultView.getComputedStyle(this,null).getPropertyValue(prop); > > A few minutes ago I stumbled upon a post, which > exactly explains that bug: > http://snook.ca/archives/javascript/safari2_display-none_getcomputedstyle/ I think the code has changed since August though. Notice that now it's in a swap that sets display:block. So if it was invisible it should now be visible and not return null. Unless visibility:hidden does it too, and in that case the swap should add visibility:visible to the swap list. I can't test on Safari so I've depended on the kindness of others in tracking down these sort of bugs. If the workaround is still needed I'd avoid two calls to getComputedStyle, which may be expensive. ___ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/
