On 27 Jan 2009, at 5:46 PM, Maxwell, Adam R wrote:

>
>
>
> On 01/27/09 02:54, "Christiaan Hofman" <[email protected]> wrote:
>
>>
>> On 27 Jan 2009, at 2:05 AM, Maxwell, Adam R wrote:
>>
>>> On 01/26/09 16:00, "Christiaan Hofman" <[email protected]> wrote:
>>>
>>>> Yes, but that's all about the window title, not the represented  
>>>> file.
>>>> I can only find one vague reference to the represented file, and as
>>>> soon as they get specific they forget to mention it.
>>>
>>> Well, for my original comment I was actually going off of the API
>>> docs,
>>> which say this:
>>>
>>> "synchronizeWindowTitleWithDocumentName
>>> Synchronizes the displayed window title and the represented filename
>>> with
>>> the information in the associated document."
>>>
>>> I think that's pretty clear in referring to "the represented
>>> filename."
>>> Indeed, if you set a breakpoint on -[NSWindow setRepresentedURL:],
>>> it's
>>> called from -[NSWindowController
>>> synchronizeWindowTitleWithDocumentName].
>>>
>>
>> And it's the one I was referring to. When it gets explicit, it
>> completely misses it.
>
> Misses what?  It tells you in the summary that it's going to update  
> the
> window title and represented filename, then in the description it  
> tells
> which document methods are called.
>
>> And the method name is absolutely anti-
>> suggestive.
>
> I do agree on that point.
>
>> Moreover, my biggest point: how do I know this is THE way
>> it happens, in other words, how do I know that overriding this method
>> will not bite me somewhere else?
>
> The AppKit release notes link I posted says that it's called  
> whenever the
> document has changed such that window title needs updating.  Since  
> that's
> dependent on the filename (by default), it must be called whenever the
> filename changes as well.
>

Indeed, the window title. That this will go together with the proxy  
icon is something we should guess from some contextual info. That's  
how I read it.

>> The point is: the docs don't tell me,
>> so it's just an assumption and a guess. We've run into the trap of
>> making assumptions from the docs that later proved wrong often  
>> enough.
>> That's why I think this really is a documentation bug: the docs  
>> should
>> be explicit about what needs to be done to override the default
>> behavior.
>
> I think the docs are pretty explicit on this, but I guess we can read
> anything we want into them.  I would add that no invented examples are
> necessary to describe how the category override + swizzling can  
> break...
>

Again, they're explicit only about the title, they're not at all  
explicit about the proxy icon.

>>
>>> Anyway, it's probably not worth this discussion; it was just a
>>> suggestion
>>> that I thought might avoid future breakage.  (apologies to the users
>>> list
>>> for dragging this out here)
>>
>> As a case in point why I don't like making assumptions when the docs
>> are not explicit: when you set the representedFilename to nil, you'll
>> get a proxy icon for something weird like /path/to/home/path/to/app/ 
>> ~/
>> path/to/app.
>
> That didn't match my recollection, so I did a quick test.  Passing nil
> raises an exception, so if you don't crash, you'd certainly get weird
> behavior:
>
> 2009-01-27 08:12:40.997 FileView Program[22970:10b] *** Assertion  
> failure in
> -[NSWindow setRepresentedFilename:],
> /SourceCache/AppKit/AppKit-949.43/AppKit.subproj/NSWindow.m:1985
> 2009-01-27 08:12:41.001 FileView Program[22970:10b] An uncaught  
> exception
> was raised
> 2009-01-27 08:12:41.003 FileView Program[22970:10b] Invalid  
> parameter not
> satisfying: filename != nil
> 2009-01-27 08:12:41.004 FileView Program[22970:10b] *** Terminating  
> app due
> to uncaught exception 'NSInternalInconsistencyException', reason:  
> 'Invalid
> parameter not satisfying: filename != nil'
>
> So if you want to eliminate the proxy icon, use  
> setRepresentedFilename:@""
> or setRepresentedURL:nil (this is inconsistent, but there is no  
> empty URL).

Than this behaves very differently on your system, because I  
definitely don't get an exception, but rather just the behavior I  
described. Consistently. Moreover, the docs don't say you can't pass  
nil, so it should be valid, as it makes sense (I think more than en  
empty string).

Christiaan


------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Bibdesk-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop

Reply via email to