Reliable workaround for long standing ATV2 bug

Cracked Home Share photo viewing at last !!
Windows XP Pro, iTunes11.2.0.115, ATV2 6.1.1 (DHCP) wired ethernet.
ATV2 will only make one attempt to connect to my PC, and once failed nothing in the world will persuade it to try again.
You can turn HomeSharing off/on on one or the other or both devices in any order - achieving absolutely nothing, always the message "turn on home sharing on your computer".
The solution - works every time - pull the mains on ATV2. The connection is established during reboot and works perfectly for the rest of the day.
Any chance of fixing this bug, Apple ??
If you can't fix it, give us a reboot option ?? (Do I really have to put a switch in the mains supply ?)

More info:
ATV maintains the connection through a manual Sleep/Wakeup sequence, but it you let it time-out of its screensave mode then a reboot is required.

Similar Messages

[svn] 4910: Implementing workaround for history manager rendering issue ( Firefox/Mac) caused by a long standing player bug.

Revision: 4910
Author: [email protected]
Date: 2009-02-10 11:51:58 -0800 (Tue, 10 Feb 2009)
Log Message:
Implementing workaround for history manager rendering issue (Firefox/Mac) caused by a long standing player bug.
Bugs: SDK-17020.
QE Notes: None
Doc Notes: None
Reviewer: Alex
Tests: DeepLinking
Ticket Links:
http://bugs.adobe.com/jira/browse/SDK-17020
Modified Paths:
flex/sdk/branches/3.x/frameworks/projects/framework/src/mx/managers/BrowserManagerImpl.as
flex/sdk/branches/3.x/templates/html-templates/client-side-detection-with-history/history /history.js
flex/sdk/branches/3.x/templates/html-templates/express-installation-with-history/history/ history.js
flex/sdk/branches/3.x/templates/html-templates/no-player-detection-with-history/history/h istory.js

Workaround for JSFL shape selection bug?

There seems to be a bug in the document selection reporting in JSFL in CS4 (haven't tested earlier versions). I submitted it as a bug to Adobe but I'd really like to find a workaround for it. I've included my bug report below. Has anyone else encountered this? If so, have you figured out a workaround? It's pretty annoying, making the tool I'm working on really difficult to manage.
******BUG******
After performing a publish preview, fl.getDocumentDOM().selection reports an incorrect selection of raw shapes.
Steps to reproduce bug:
1. Create a JSFL command with the following contents:
doc = fl.getDocumentDOM();
fl.trace("there are " + doc.selection.length + " items selected");
2. Start the authoring environment fresh, with a new document.
3. Draw several shapes on the stage, not touching each other, all in the same frame and layer.
4. Select one of those shapes but leave the others unselected.
5. Run the previously created JSFL command. It will send the following text to the output panel, as one would expect:
"there are 1 items selected"
6. Do a publish preview (either Flash or HTML).
7. When it comes up, close the preview window.
8. Deselect all and then select one of the shapes (again, keeping the others unselected).
9. Run the JSFL command again. This time, it will say that there are n items selected (where n is the number of shapes you drew in step 3). So if you drew three shapes, it will print out the following:
there are 3 items selected
Note that this result will be the same even if you go to a different document, do a publish preview on that document, then return to the original. It seems that simply doing a publish preview alters Flash's state to report the selection incorrectly.
Results: The JSFL command reports that all the shapes are selected, despite the fact that only one of them is. The only way I've found to get Flash back to its normal behavior is to restart the authoring environment.
Expected results: In the steps above, the JSFL command should always print out that there is one element selected. There's no reason that doing a publish preview should change that.

When selected all shapes in selection are treated as one. You can see it if you run this script with selected multiple shapes:
fl.trace( fl.getDocumentDOM().selection[0].contours.length );
It will output you twise bigger number then selected shapes (because each Shape has two contours - one clockwise, other counterclockwise).
Of course this. implementation is strange for me too. Thats how i found your post

Workaround for keyword button sets bug

Thought I'd post this in case anyone else is beating their head against a wall trying to use custom keyword button sets.
BUG: Control bar sometimes displays the wrong captions for the keyword buttons. Cause: Deselecting a keyword button set in the "Edit buttons" dialog will cause any active button sets after it to appear corrupted in the control bar. For example, with the default set of "Photo Descriptors, People, Stock Categories, Snapshots, ..." try unchecking the Stock Categories button set. When you return to the control bar, use the popup button set selector (just under "Add Keyword") to select Snapshots. Instead of the Snapshot keyword buttons, you will instead see the button captions from Photo Descriptors, though if you click one of those buttons, it will apply the appropriate keyword from the Snapshots set.
Workaround: Rearrange the order of the button sets list in the Edit Buttons dialog to make all checked button sets appear before all unchecked button sets. In other words, don't have any gaps where you have an unchecked button set followed by a checked button set.
Cheers,
Paul

You are my hero!!!! This is exactly what I was looking on here for. Thanks again.

Is there a workaround for the transform/crop bug?

Happy Tuesday!
Does anyone know a way to get around this bug?
- I have two video clips.
- One video clip (on top) is a talking head that need to be placed in a box on the other clip.
- Apply transform to talking head clip. Scale and change position.
- Apply crop to talking head clip. Bring in left and right so image fits in the box.
Everything is good so far...
- Apply same transform and crop to other talking head clips in same project.
- Crop fails terribly for second and all subsequent attempts. It seems like there is a magic area for which the clip cannot render.
The first transform/crop works fine...
All subsequent transform/crops, either mannually applied or from copied attributes, do this: (The collection of icons are inside the box and is showing through from the bottom clip.)
If I drag the image around the screen, there seems to be a "magic area." If the image is in the magic area, everything is fine. If the image approached any edge of the screen, the bad crop effect happens. Here I have dragged the transform near the bottom left corner of the screen:
I have played with crop settings and the way the crop jumps, there seems to be some kind of logic bug in the cropping code, particularly with transform/crop combinations. Any help would be appreciated! This bug is a show-stopper, and I'm on a fairly tight schedule. Does anyone know of a way to get around this bug?
Tom

I guess, I'm of no help here.. I can crop and transform like crazy - no issues
I also tried your 'special' method, to crop/transform the main storyline (=what I wouldn't do) ... but the attached other clips appear without problems:
and after that 'arrangement', another one - no issues:
tsphillips wrote:
… Perhaps an interesting experiment would be to see what it takes to make a clip render outside of it's crop box.
here you lost me - in translation, sorry:
a clip OUTside its crop??
outside a crop is nothing, that's what a crop is for… shouldn't it?
cropping means; cutting-off parts of the pic, that icon is a passépartout..
so, outside a crop is … nothing to see here, Sir..
Sorry, of being no help here......
Would be interesting, what Cupertino will andswer you..........

Any simple workaround for the 'umount -a' bug?

On Arch 0.7.2 the installer seems to use a busybox version that has a bug (according to what I've read) or maybe busybox isn't built with the umount -a feature.
The thing is that 'umount -a' doesn't work so, after installation, the partitions aren't umounted which results in FS errors on the next boot.
As I'm developing an automated installation I don't use ArchLinux installation scripts but I stumble into this problem.
Anyone can think of a simple shell snippet to umount all mounted filesystems?
Something like a loop would do, but I can't think of anything simple.
Maybe I can even use an updated ramdisk or update/patch busybox on the 0.7.2 initrd.

detto wrote:dunno exactly whats the problem :S
all the time ive installed arch using ext3 as fs it does indeed a quick check during first boot but that always looks like wanted to me, and also is only maybe 3-4 secs long. what fs are u using in installation?
well that check is probably the result of the filesystem not being umount ed correctly. The problem is supposed to be a bug in 'busybox', the program that's behind all basic comands on the installation. If you try to do 'umount -a' it doesn't work (it should umount everything).
My filesystem of choice is reiserfs, and it does handle problems like this. But that's not the point. There's a bug that needs to be fixed.

Workaround for iPhone smart playlist bug

There is a bug (do a search for smart playlist in this forum) with smart playlists not displaying the correct songs on the iPhone.
I just tried something that seems to have worked...
I have a playlist in iTunes called "Recently Added" which I want to keep synced with my iPhone.
I created a new smart playlist called "Recently Added (iPhone)". The playlist has one rule: "Playlist is Recently Added" with live updating checked.
I now sync this second playlist with the iPhone and all the correct songs show up in it.

Excellent idea Jeff.
Your approach works for the same reason that my original idea works.
The inclusion of another playlist as a condition in a smart playlist, causes the playlist to become "rasterized" or static.
The developers wanted smart playlists to be really smart - instead of copying a static list of songs from the smart playlist to the iPhone/Pod, they simply send the definition of the playlist to the iPhone and let it calculate which songs from it's own library belong on it.
The nice thing about their approach is that the playlists will continue to live update on the Phone in-between syncs - so if you have a playlist called "Last 25 played" it will have the last 25 played songs on your iPhone, instead of the last 25 played songs that you played in iTunes.
The problem of course, is that there's a bug in the way the iPhone calculates these rules (probably related to the porting of code to a new architecture).
So why does this fix work?
Well, my theory is that when you create a smart playlist that references another playlist, iTunes simply calculates the songs that should appear on the smartplaylist and sends a literal list to the iPhone/Pod.
This solves the problem that would be created if you only sync the parent playlist, and not the playlists it references. (IE, if you would want to be able to sync the "25 most played" playlist without also having to sync the "MusicOnly" playlist).

Workaround for IMAP mailbox import bug.

The Import function in Mail doesn't work properly with mailboxes created using an IMAP account. It will refuse to import them. After some fiddling, I discovered that's its the mailbox's file extension that chokes the Import function. It expects ".mbox", and IMAP mailboxes have ".imapmbox". The workaround is to manually change the extension on them to ".mbox", and then Import will be able to read them.

Whew, great tip, thanks!

WARNING: workaround for critical clustering bug in OC4J

I found a workaround for the HTTP clustering bug described in my previous posting of August 16, 2001 "Urgent: serious HTTP clustering bug in OC4J?" (http://technet.oracle.com:89/ubb/Forum99/HTML/000261.html)
The workaround:
When adding the <cluster config/> tag in orion-web.xml specify and ID (numeric) with no more than 9 figures. The autogenerated ID is longer (14 figures on my test system)
and will cause the error.
Luciano
null

burricall wrote:
..Did you find a solution?See [this post|http://forums.sun.com/thread.jspa?threadID=5426321&messageID=11001580#11001580].

Inaccurate timing for long slides causes early or late advance. Error 2%

Dear Apple,
There are many users of Keynote who wish to embed movie clips and audio files into their slides as part of a looping presentation, for example: in a museum or corporate lobby.
Both Keynote and PowerPoint have the ability to hold a slide in place for a specified length of time, simply by turning on a "Transition Delay" and entering a time value.
So, if I wanted to play a 2 minute video and advance the slide precisely at the movie's endpoint, I would enter a delay of 120 seconds. Likewise, if the slide were 10 minutes long, I would expect the slide to advance at 10 minutes, one second.
Sadly, this is not the case, as demonstrated by these threads. Even worse: the clock error is different from machine to machine.
http://discussions.apple.com/thread.jspa?messageID=11073048&#11073048
http://discussions.apple.com/thread.jspa?messageID=11336657&#11336657
http://discussions.apple.com/thread.jspa?messageID=11877994&#11877994
http://discussions.apple.com/thread.jspa?messageID=11198872&#11198872
http://discussions.apple.com/thread.jspa?messageID=11517877&#11517877
All of these threads have gone unanswered by an Apple representative for months. If there was a knowledge base article, nobody's posted it.
Can you please give the Macintosh community an update regarding this long-standing Keynote bug?
Thank you.

Doublechecking something. In the course of your troubleshooting to date, have you tried running a chkdsk over the Hard drive that the songs are stored on on the PC?
For instructions (and a guided help) see the following Microsoft document:
http://support.microsoft.com/kb/315265

Ora:view support for long columns

Hi, i'm making a documentation tool of my database tables.
I want to generate XML description of tables with Xquery.
When i use this Xquery, gives me error,
ORA-00932: tipos de dato inconsistentes: se esperaba NUMBER se ha obtenido LONG
SELECT XMLQuery(
'for $i in ora:view( "SYS", "ALL_TABLES")/ROW
return <table id="{$i/TABLE_NAME}">
{for $j in ora:view("SYS","ALL_TAB_COLUMNS")/ROW
                 where $j/TABLE_NAME eq $i/TABLE_NAME
                 return ($j/COLUMN_NAME)}
</table>'
RETURNING CONTENT) FROM DUAL;
The view ALL_TAB_COLUMNS have a LONG column for DEFAULT value.
Does ora:view have a workaround for LONG columns?
The only solution that i have found is to create a new view, without the LONG column, but that will make my documentation tool less portable.
Thanks.

Hi,
user625976 wrote:
The only solution that i have found is to create a new view, without the LONG columnThat's the only workaround I can imagine too.
But, what about using SQL/XML functions?
For example :
SELECT xmlelement("tables",
xmlagg(
    xmlelement("table",
      xmlattributes(c.table_name as "id", c.owner as "owner"),
      xmlagg(
        xmlelement("column_name", c.column_name)
        order by c.column_id
    ) order by c.table_name
FROM all_tab_columns c
GROUP BY c.owner, c.table_name;

Long standing bug in ARD

There is a LONG STANDING bug in ARD dating back to at least v2.0, maybe earlier.
While controlling a remote Mac, for seemingly no reason, the display stops refreshing. The connection has not been lost just refresh is stopped. A single refresh workaround is to click the color adjustment bar, no need to move it, just click it and you get one refresh. This happens to me way too frequently! It occurs on PPC as well as Intel Macs, on local LANs and over the Internet through VPNs.
Sometimes you can break out of this by closing the connection and connecting again but it usually requires the remote machine to restart the ARD service. Doing this with ARD itself usually leaves the remote machine unreachable (ARD service stopped.) I have taken to using a near by Windows machine then ssh'ing into the Mac to manually restart the service. This works but the refresh problem re-occurs after an undetermined amount of time.
There seems to be no clues as to why this happens. There is never any particular task I am performing, no specific app I am always running, etc, etc.
Excuse me if this has already been brought up and if someone knows a fix or can point me to a KB article I would be most appreciative!
Thanks,
Josh
G3, G4, G5, Intel iMac, you name... Mac OS X (10.4.6) Mac OS X (10.4.6)

I, too have experienced this problem. However, much more so since I upgraded to ARD 3. This can happen, as you say, on a wide variety of clients: eMacs, G5 Dual, iMac G5, Powerbooks. Most annoying. I have found that if I use TB2 to control the client, I will find, in some cases, the SystemUIServer process is hung. By Force Quitting this process and Re-Starting the client, I am able to get ARD to run again.
This seems a little more severe than your problem, but I have had the same experience as you describe as well--the SystemUIServer is not hung in these cases. I can generally find nothing wrong on the client to account for the behavior.
Nor have I found a satisfactory answer to the problem on this forum. Would love it someone could offer a clue!
PowerBook G4 Aluminum 15 Mac OS X (10.4.6)

Best workaround for editing long AVCHD clips in CS6?

Hi there, I've recently switched to CS6 and have run into incredibly long render times using native AVCHD clips (13 hours to encode a 30 minute m2v). After reading many posts it seems the issue I am running into is a known bug. So, my question is: is there a good workaround for this issue?
It sounds like this issue is only on long clips. Is it possible to automatically split the .mts files using another program (virtualDub?) prior to importing?
If transcoding from the native format is required (until the fix) what would be the best method using Windows 7 64bit? I have heard Avid's DNxHD is a good lossy codec, but I hate to do this if there is an accepted workflow I should learn instead.
Thanks for any advice!
-Stephen

One of these message threads may have some help
CS6 Bug AVCHD http://forums.adobe.com/thread/1004369?tstart=0
-and http://forums.adobe.com/thread/1004369?start=0
-and LOCK the media http://forums.adobe.com/thread/1077245

[svn] 3105: SDK-15764, SDK-15765 -- Two long standing Ant task bugs.

Revision: 3105
Author: [email protected]
Date: 2008-09-04 15:05:39 -0700 (Thu, 04 Sep 2008)
Log Message:
SDK-15764, SDK-15765 -- Two long standing Ant task bugs.
- Ant Tasks fail when setting fork="true" and there are spaces in the path name
- Setting fork="true" in Ant Tasks leads to warning
The issue with spaces is that forked tasks construct a stringified commandline, since Java doesn't allow you to pass real String[] to a new process. When not forking, we can pass the original String[], which makes spaces irrelevent. So the fix uses a Commandline.toString(String[]) which is designed to wrap arguments with spaces correctly with quotes, so the commandline tokenizes correctly.
Bugs: SDK-15764 SDK-15765
QA: This affects ant calls for _all_ FlexTasks where fork=true. It would be great if we have tests for both fork="true" and "false" where we tested paths and arguments with and without spaces -- off the top of my head, good ones to try: +flexlib=/path/with spaces/, -define=CONFIG::foo,"bar with spaces", and any other arguments you can think of. Also very important: Test passing space arguments without quotes, with single quotes, with double quotes, with a mixture (illegal), and with a mixture escaping the inner set of quotes (not sure if this should work or not, probably a stupid case to test unless quotes are allowed in paths).
Doc: No
Reviewer: Pete
Ticket Links:
http://bugs.adobe.com/jira/browse/SDK-15764
http://bugs.adobe.com/jira/browse/SDK-15765
http://bugs.adobe.com/jira/browse/SDK-15764
http://bugs.adobe.com/jira/browse/SDK-15765
Modified Paths:
flex/sdk/branches/3.0.x/modules/antTasks/src/flex/ant/FlexTask.java

[svn:fx-trunk] 10526: Temporary workaround for A+ bug SDK-23387.

Revision: 10526
Author:   [email protected]
Date:     2009-09-22 21:48:33 -0700 (Tue, 22 Sep 2009)
Log Message:
Temporary workaround for A+ bug SDK-23387.
The LinkNormalFormat, LinkHoverFormat, and LinkActiveFormat classes have been removed from FXG, and the FXG compiler was autogenerating static vars of these types (for linkage reasons?) that no longer compiled. I've commented out the offending lines in FXGCompiler.java, but we need a correct fix to cope with the fact that the way hyperlinks are formatted in FXG has changed.
Instead of formatting hyperlinks by writing
tag) but at least it can once again compile FXG as long as it doesn't involve styled hyperlinks.
QE notes: None
Doc notes: None
Bugs: SDK-23387
Reviewer: Carol
Tests run: ant checkintests
Is noteworthy for integration: No
Ticket Links:
    http://bugs.adobe.com/jira/browse/SDK-23387
    http://bugs.adobe.com/jira/browse/SDK-23387
Modified Paths:
    flex/sdk/trunk/modules/compiler/src/java/flex2/compiler/fxg/FXGCompiler.java

error dateField not selecion date 27/11/2002 ?
This is bug flex 3 ?
thanks

Reliable workaround for long standing ATV2 bug

Similar Messages

Maybe you are looking for