Cloning bug in spec?
I have hit on what I thought was a bug, but after looking at the spec I'm
not sure.
What I'm doing is cloning a persistence-capable class after is has been
loaded from the database, but before fields that contain references to
other persistence-capable classes have been accessed.
The clone receives a shallow copy of the fields of the original (null at
this point), and has its jdoStateManager set to null. After the clone
operation is finished, the clone neither has the value of the field, or a
jdoStateManager to later get the value of the field.
Is this really how clone should work? The spec doesn't really address
this situation. It only says that clones should receive "standard shallow
copy of the fields of the cloned instance".
My quetion is this: Is this a bug in the spec that I have to work around
(be sure to at least access all fields that may need to be loaded from the
database during or before clone), or is this something that JDO
implementations should be taking care of for me?
I have a very simple test case (three classes) that exhibits this behavior
if it will help with diagnosis.
I am running Kodo 2.2.5 on JSDK 1.3.1_03.
Thanks!
-- Chris
Chris --
This is more of a 'bug' in the spec. Prior to serialization, a persitent object will suck in all of its persitent fields and relations, so the entire object graph is serialized. Prior to cloning, though, nothing special is done. And of course, a clone is always transient (otherwise you'd get into situations where two objects in memory represented the same persistent state in the same PersistenceManager, which could lead to all sorts of problems).
In the 1.0 JDO specification, supported by Kodo 2.3, there is a PersistenceManager.retrieve(Object pc) you can use to load all persistent fields into an instance prior to making it transient or cloning it. Look for a new beta of 2.3 in the next few days.
Similar Messages
-
Bug in spec or compiler?
Hello!
I have some problems with the jsr14 specification and the compiler implementation. I am not a native english speaker, but I think I can understand the specification correctly.
In the spec8.pdf document "Adding Generics to the Java Programming Language - Participant Draft Specification" on page 16 example 20: this should not compile but it does compile with jsr14 prototype 1.2 (javac 1.4.0_01)! Why? Anyone tried this?
In the same spec page 14: 6.1: "the erasure of a type variable is [...] if the bound consists of interface tyes only, the interface among the erasures of all interface types which has the least canonical name, using lexicographic ordering."
This sounds crazy, and the compiler does it different. It always uses the first parameter (following extends)! You can see this using "javap -c" on the bytecode. So, what�s right?
In the spec, page 3: "if a type parameter X has more than one bound, then it is a compiletime error to reference a member of an object whose declared type is X, unless that member is an accessible member of class java.lang.object"
Well, why is this an error? It compiles without any errors!
And another point: in the spec are examples of type variables with the "A implements I" syntax, which is not allowed. The compiler uses "extends" only. The example code is even on the same page (Page 3). Oh, oh!
Maybe I am wrong in some parts, and you can help me. (I cant believe, that I found three bugs in five hours) Any comments?There were no answers to my questions, yet. Here is some sample code to show my problems:
In the spec8.pdf document "Adding Generics to the Java
Programming Language - Participant Draft
Specification" on page 16 example 20: this should not
compile but it does compile with jsr14 prototype 1.2
(javac 1.4.0_01)! Why? Anyone tried this?Here is the code:
class C<A> { A id (A x) {} }
interface I<A> { A id(A x); }
// should not compile, but does compile!
public class D extends C<String> implements I<Integer> {
String id(String x) {}
Integer id(Integer x) {}
In the same spec page 14: 6.1: "the erasure of a type
variable is [...] if the bound consists of interface
tyes only, the interface among the erasures of all
interface types which has the least canonical name,
using lexicographic ordering."
This sounds crazy, and the compiler does it different.
It always uses the first parameter (following
extends)! You can see this using "javap -c" on the
bytecode. So, what�s right?Here is the code:
interface IA{ public void testA(); }
interface IB{ public void testB(); }
interface IC{ public void testC(); }
public class Test<A extends IB & IA & IC>{
public A value; // javap tells me, that value has type IB!
In the spec, page 3: "if a type parameter X has more
than one bound, then it is a compiletime error to
reference a member of an object whose declared type is
X, unless that member is an accessible member of class
java.lang.object"
Well, why is this an error? It compiles without any
errors!Here is the code:
public class Test2<A extends IA & IB & IC>{
private A value;
public static void main(String[] args){
value.testA(); // where is the problem
value.testB(); // it works!
And another point: in the spec are examples of type
variables with the "A implements I" syntax, which is
not allowed. The compiler uses "extends" only. The
example code is even on the same page (Page 3). Oh,
oh!Here is the code:
class ReprChange<A implements ConvertibleTo<B>,
B implements ConvertibleTo<A>> {....
Maybe I am wrong in some parts, and you can help me.
(I cant believe, that I found three bugs in five
hours) Any comments?Again: is anyone here who can tell me what is right? -
Weird issue occurring after cloning the R12 Vision instance.
Hi everyone,
We are currently evaluating R12 using the Vision DB instance. Since I'm seeing that the apps processes like to have at least 2gb of memory, the initial server just isn't beefy enough for R12 so I cloned it over to a better server following metalink Note:406982.1 (Cloning Oracle Applications Release 12 with Rapid Clone)
There were various cloning issues that came up that fortunately, I was able to resolve myself .. still with tarring up directories, ftp'ing the tar files, etc etc it took almost 2 days to finish the clone.
Upon starting the clone I'm seeing the following errors reported when the db listener and the apps fndfs listener start:
TNS-01114: LSNRCTL could not perform local OS authentication with the listener
TNS-01115: OS error 22 creating shared memory segment of 129 bytes with key 2044160132
However, the listener processes actually do start even though the above errors are reported and I verify that I can see the server processes for both. In addition, I can login to apps and do everything without any issues and login from sqlplus to the database remotely and everything appears to be working.
But if I try to look at the listener status or try to stop either listener process I get the above errors. So, if I need to bounce the app or the db I have to actually kill the listener processes in order to do so. Then upon startup, the above situation repeats.
So, I've opened a SR in regards to this as it appears to be a cloning bug. The only metalink hits I found were talking about the shmsys:shminfo_shmmni setting (currently at 200) but I don't think that's it since I've had 6 apps instances up and running on this particular server before.
Currently the R12 Vision instance is running with RUP 1 applied. I did see there was a new Techstack update and a new RUP 2 update for apps released just last week but thought I would ask if anyone else who's used rapid clone with R12 experienced similiar issue?
Thanks.
Chad
Message was edited by:
Chad SmithI am usually seen on the on ramps with a sign that says will clone for food...
I think your perl version is a little low for R12. Here is what I have in an upgraded R12 from 11.5.10.2 Vision:
which perl
/lv01/oracle/ltest/apps/tech_st/10.1.3/perl/bin/perl
perl -version
This is perl, v5.8.3 built for i686-linux-thread-multi
Copyright 1987-2003, Larry Wall
Perl may be copied only under the terms of either the Artistic License or the
GNU General Public License, which may be found in the Perl 5 source kit.
Complete documentation for Perl, including FAQ lists, should be found on
this system using `man perl' or `perldoc perl'. If you have access to the
Internet, point your browser at http://www.perl.com/, the Perl Home Page.
I was clearing 150 GB for an Oracle Enterprise Linux VM for installing/cloning 11i or R12 into, so I will clone a scratch R12.0.2 upgrade instance there. The prepare scripts completed without error, and the cpios are running now.
prepare DB:
ApplyDBTier Finished at Mon Jul 23 22:00:38 PDT 2007
Status: Completed Successfully
perpare AppsTier:
Successfully prepared appl_top for staging. Mon Jul 23 22:07:05 PDT 2007
Creating zip files for appltop.
Mon Jul 23 22:07:05 PDT 2007
Done...
Successfully created zip files.
Mon Jul 23 22:07:05 PDT 2007
StageAppsTier Finished at Mon Jul 23 22:07:05 PDT 2007
Status: Completed Successfully
I will update thread with the results of the config clone scripts.
Make sure you link to your environment files correctly:
lrwxrwxrwx 1 oracle dba 43 2005-02-10 20:29 vis9.env -> /lv03/oracle/vis9db/9.2.0/vis9_hostname.env
lrwxrwxrwx 1 oracle dba 62 2007-01-26 22:59 visr12.env -> /lv04/oracle/visr12/db/tech_st/10.2.0/visr12_hostname.env
ln -s /lv04/oracle/visr12/db/tech_st/10.2.0/visr12_hostname.env visr12.env
lrwxrwxrwx 1 applmgr appl 33 2005-02-10 20:27 vis9.env -> /lv03/oracle/vis9appl/APPSORA.env
lrwxrwxrwx 1 applmgr appl 66 2007-01-26 22:56 visr12.env -> /lv04/oracle/visr12/apps/apps_st/appl/APPSvisr12_hostname.env
ln -s /lv04/oracle/visr12/apps/apps_st/appl/APPSvisr12_hostname.env visr12.env
Yes, mythtv is fantastic. I also like mythvideo and mytharchive. I needed to move unneeded backups and edelivery downloads to make room for recordings... -
Flash CS6 not responding every few seconds.
I've been working on a few animation projects for the past few days, and Flash CS6 always tends to show the spinning blue wheel and white screen for like 10 seconds. This is very irritating. I would probably be done with the project by now if it weren't for this bug.
Specs:
Windows 7 64-bit
Nvidia GeForce GT 540m
8gb ram
Intel Core i7
Thanks in advance.I'm also having this problem. Every now and then (probably with a 5-10 min interval) Flash CS6 freezes for 10 or more seconds. The larger the Flash-file, the longer it freezes. I've turned off the auto-recovery mode, I'm not working on a file that's in a network folder (like Dropbox, since I heard somebody had trouble with this), and I've quit all programs. Still freezes. I've also got all the latest updates.
It's very annoying, and I think I would've been able to work a lot more if this weren't happening all the time. Perhaps I'll try to import all of the content into a new file?
Any know solutions?
I'm on a late 2009 iMac:
Mac OSX 10.8.3
ATI Radeon HD 4850 512 MB
8GB RAM
2,8 GHz Intel Core i7 -
Hi,
I have been hunting down a NullPointerException for half a day to come to
the following conclusion.
My constructor calls a method which uses static variables. Since an intance
of my class is created in the static block when the class is loaded, those
statics are probably not fully initialized yet and the constructor called
from the static block has those null pointer problems.
I've considered moving the initialization of the static variables from the
declaration to the static block. But your code is inserted BEFORE any other
code. Therefore not solving my problem.
Two questions:
1) what would be a solution to my problem? How can I make sure my static
variables are initialized before the enhancer generated code in the static
block calls my constructor? Short of decompiling, changing the code and
recompiling.
2) Why is the enhancing code inserted at the beginning of the static block
and not at the end? The enhancements would be more transparent that way if
the static variables are initialized in the static block.
Thanks,
EricHi Eric,
JDO calls the no-args constructor. Your application should regard this constructor as belonging
primarily to JDO. For example, you would not want to initialize persistent fields to nondefault
values since that effort is wasted by JDO's later initilization to persistent values. Typically all
you want to initialize in the no-args constructor are the transactional and unmanaged fields. This
rule means that you need initialization after construction if your application uses the no-args
constructor and wants persistent fields initialized. On the other hand, if your application really
uses constructors with arguments, and you're initializing persistent fields in the no-args
constructor either unintentionally through field initializers or intentionally as a matter of
consistency, you will find treating the no-args constructor differently helpful.
On the other hand, if Kodo puts its static initializer code first as you report, then it is a bug.
Spec Section 20.16: "The generated static initialization code is placed after any user-defined
static initialization code."
David Ezzio
Eric Borremans wrote:
>
Hi,
I have been hunting down a NullPointerException for half a day to come to
the following conclusion.
My constructor calls a method which uses static variables. Since an intance
of my class is created in the static block when the class is loaded, those
statics are probably not fully initialized yet and the constructor called
from the static block has those null pointer problems.
I've considered moving the initialization of the static variables from the
declaration to the static block. But your code is inserted BEFORE any other
code. Therefore not solving my problem.
Two questions:
1) what would be a solution to my problem? How can I make sure my static
variables are initialized before the enhancer generated code in the static
block calls my constructor? Short of decompiling, changing the code and
recompiling.
2) Why is the enhancing code inserted at the beginning of the static block
and not at the end? The enhancements would be more transparent that way if
the static variables are initialized in the static block.
Thanks,
Eric -
Chapters and Playlists.
I have a one piece video from P Pro. I have set the menus and chapters within that timeline. I have a button to play all and one for each chapter and after each chapter the end action to the menu. Each of the chapter sections works well. However whren I "play all" it only plays the first chapter and returns to the menu. Where did I go wrong
ThanksScenarist is a $30,000 application for the full, unrestricted full featured version - so perhaps not a fair comparison.
Maestro is defunct - swallowed up by Apple & turned into the dreadful DVDSP which is chock full of bugs and spec fudges.
And lets be honest, Maestro may be remembered fondly, but so is the Victorian era. "The Good old Days" never really were that good, same with "Golden Ages".
It's all dreadfully subjective.
However, full scripting would indeed be a serious boon, and will you please file a feature request? The more requests a feature gets, the more chance it will get included next time around.
http://www.adobe.com/support/feature.html -
Premiere CC is crashing anytime I try to move a clip on the timeline
It also will crash by just selecting a clip. I've tried to identify any problem clips through a process of elimination, but it doesn't seem to be realetd to any specific footage.
I am getting this error messge:
[/ppro721/releases/2013.10/shared/adobe/MediaCore/Backend/Make/Mac/../../Src/Sequence/Trac kItemSelection.cpp-108]
Thanks for any suggestions!
Scott
Premiere Pro CC
Mac
OSX 10.8.5Hello,
I was having similar trouble with a project I synced in Pluraleyes (on Windows).
The XML timeline I imported was extremely glitchy - it would not visually refresh and there was the window drag glitch (only this time with the Premiere timeline).
Here's what I mean (randomly Googled image to serve as an example):
http://www.marcofolio.net/images/stories/programming/webdesign/winxp_error/winxp_error_nos talgia.png
Instead of the window being repeated, it was the playhead.
I exported the sequence as a separate Premiere project and reimported that into my main project. It fixed the visual glitchiness, but then produced the same crash as described in the first post (TrackItemSelection.cpp).
I deleted the entire video track and using Match Frame on the audio clips, I manually put the video clips back and linked them with their audio.
Not ideal, but fortunately it was a short sequence. I shudder to think of the pain on a larger project.
Please fix this bug.
Specs:
Windows 8.1
Intel Core i7-3770
12 GB RAM
GeForce GTX 560 (I disabled Mercury Playback but no cigar) -
Bug SQL Developer 3.0.04 Save Package Spec and Body to ips file bug
Hi.
I have problem with export package body end header. Exported ips file have some rows switched compared to original source code.
SQL Developer version: 3.0.04. Build MAIN-04.34
Error simulation:
- Click on package header with right mouse button in SQL Developers Connections bar
- In context menu choose 'Save Package Spec and Body...'
- Save ips file... (Small bug: Offered file name is wrong (= last saved package/file, not actual name of package) :-( )
- This file(source code) compare with original source code in Developer and u will see switched rows every +-90. row in source code.
Can u fix this in next version of SQL Developer?
Thanks...
Edited by: 880809 on 22.8.2011 6:57Bug 12904494 has been created.
Michael -
JSC IDE bug or has the spec changed?
I have been fighting this for days now trying to find a way around it, asking questions in Sun's JSC, JSP, JS forums as well as on JavaRanch.
For some reason, the Java Studio Creator IDE does not like the following code:
<jsp:params>
<jsp:param name="URL" value='<%=(String)request.getAttribute("urlValue")%>'/>
</jsp:params>It says that: The value of attribute "value" must not contain the '<' character.
The thing is, I have seen this time and time again in examples on various sites, and this is what I have been told to use by users on Javaranch and Sun's JSF forums.
So, is it a spec change, an IDE bug, or are all of those people wrong?
BTW, what I am trying to do (and banging my head into the wall over the past three days now) is to get a value from my backing page bean read into the JSP so I can pass it to an applet. Surely this simple thing is possible isn't it?
Thanks.I think the problem is that JSC uses 1.2 and it defaults to isELIgnored="true" where as 2.0 defaults to false. Anyway, the underlying issue is solved here: http://forum.java.sun.com/thread.jspa?messageID=3574258�
-
Revision: 9892
Author: [email protected]
Date: 2009-09-01 16:06:51 -0700 (Tue, 01 Sep 2009)
Log Message:
Panel skin changes per XD spec, and a few other Panel Skin bugs. Default Panel size changed to match spec. Panel titles are now single-line and show ellipsis.
QE Notes: Panel-related bitmap compares will need upgrading
Panel_Properties_title5
Panel_Properties_DfrdInst_none
Panel_Properties_DfrdInst_auto
Doc Notes: None
Bugs: SDK-21680, SDK-21671, SDK-21666, SDK-19962, SDK-19959
Reviewer: Ryan
API Change: No
Is noteworthy for integration: No
tests: checkintests mustella/managers/Panel
Ticket Links:
http://bugs.adobe.com/jira/browse/SDK-21680
http://bugs.adobe.com/jira/browse/SDK-21671
http://bugs.adobe.com/jira/browse/SDK-21666
http://bugs.adobe.com/jira/browse/SDK-19962
http://bugs.adobe.com/jira/browse/SDK-19959
Modified Paths:
flex/sdk/trunk/frameworks/projects/spark/src/spark/components/Panel.as
flex/sdk/trunk/frameworks/projects/spark/src/spark/skins/spark/ApplicationSkin.mxml
flex/sdk/trunk/frameworks/projects/spark/src/spark/skins/spark/PanelSkin.mxml
flex/sdk/trunk/frameworks/projects/wireframe/defaults.css
flex/sdk/trunk/frameworks/projects/wireframe/src/WireframeClasses.as
flex/sdk/trunk/frameworks/projects/wireframe/src/spark/skins/wireframe/PanelSkin.mxml
Added Paths:
flex/sdk/trunk/frameworks/projects/wireframe/src/spark/skins/wireframe/ApplicationSkin.mx mlNew forum: http://forums.adobe.com/community/flash_builder
-
Hi hussein,
I have been cloning our EBS 11i for many times before at it was good. Lately I applied set of patches. Then what happened is every time I clone, I have to edit the http.conf
The patches have affected the public access of the system, in a way that it seems “disabling” the public IP access. This was caused by the patch updating the “httpd.conf” file.
HTTPD.conf is a server configuration file which typically contains directives that affect how the server runs, such as user and group IDs it should use, and location of other files. Because the server configuration file is the main file that the server starts with, Oracle HTTP Server does not include any directive that says where to locate it. The location is passed on command line when the server starts.
This file is located at:
TRNG
/uo1/oracle/oatrngora/iAS/Apache/Apache/conf/httpd.conf
PROD
/u02/oracle/oaprodora/iAS/Apache/Apache/conf/httpd.conf
To resolve the public IP access problem, edit the httpd.conf file and find the following lines:
==========================================
# Port: The port to which the standalone server listens. For
# ports < 1023, you will need httpd to be run as root initially.
Port 8000
Listen oracletrng.abc.local:8000
===========================================Remove the hard-coded Server.DNS name as shown below.
===========================================
# Port: The port to which the standalone server listens. For
# ports < 1023, you will need httpd to be run as root initially.
Port 8000
Listen 8000
============================================Then restart the Oracle Apps Services (apps tier only).
Any idea how to fix this?
Thanks a lot
Ms Inday
Edited by: crystal on Jan 29, 2011 8:14 PM
Edited by: crystal on Jan 29, 2011 8:57 PMHow do I know which is AD and ATG?
hostfile is:
192.168.100.5 oracletrng.abc.local oracletrng
To resolve issues about the Purchasing – Work in Process Module, the following 34 patches were applied at TRNG and PROD servers.
Patching Order Sorted by Number
4291053 3218526
3218526 3315676
3854951 3761838
3761838 3854951
4206794 4206794
4216942 4216942
4393740 4291053
4608023 4355172
6064563 4393740
5610530 4608023
5527236 4635071
7172682 4649626
4355172 4651420
6603716 4665900
4946700 4676589
6642331 4946700
5400654 5024341
4665900 5400654
4651420 5452096
5923665 5527236
4635071 5610530
6360253 5647736
4649626 5739724
5024341 5923665
5647736 6064563
7429271 6360253
6975727 6603716
3315676 6642331
7660107 6975727
5452096 7172682
4676589 7429271
5739724 7660107
8403254 8362391
8362391 8403254
The above patches were test applied at the test sever OATRNG, and after checking that it solved the issues it was then carried out to the Production (PROD)
Server.Thanks
Edited by: crystal on Jan 30, 2011 12:00 AM
Edited by: crystal on Jan 30, 2011 12:03 AM -
I am not good at English. But but... please hear my JSF Problem.
<h:form>
<h:inputText id="input1" required="#{true}"/>
<h:message for="input1"/>
<h:inputText id="input2" value="#{testBean.value}">
<f:convertNumber integerOnly="true"/>
</h:inputText>
<h:message for="input2"/>
<h:commandButton/>
</h:form>From browser:
1.Input empty string at input1. (this cause validation error.)
2.Input empty string at input2. (convet to null.)
3.Click commandButton.
I hope:
Display value at input2 is empty string. (My input value is keeped.)
But result is:
Display value at input2 is "#{testBean.value}". (My input value is forgotten...)
In my research:
1.converter converts empty string to null.
2."local value" is set to null.
3.Go to Rnder Response Phase By Input1 validation error.
4.getValue ignore "local value" if it is null and see VE.
Uhh... Why getValue method doesn't see isLocalValueSet?
I can't understand... Please Help me...Unfortunately I've had no luck with this solution. In my case I am performing a search using a session scope bean. My results are listed in a dataTable and I have an actionListener on a request scoped bean. All of that works fine.
faces-config.xml
<managed-bean>
<managed-bean-name>searchBean</managed-bean-name>
<managed-bean-class>com.package.SearchBean</managed-bean-class>
<managed-bean-scope>session</managed-bean-scope>
</managed-bean>
<managed-bean>
<managed-bean-name>detailBean</managed-bean-name>
<managed-bean-class>com.package.DetailBean</managed-bean-class>
<managed-bean-scope>request</managed-bean-scope>
</managed-bean>
public class DetailBean
private String field1;
private String field2;
//Bind the dataTable component to the controller
protected transient UIData dataTable;
//The object used to populate the row in the table that was selected
protected transient Object row;
public DetailBean()
super();
FacesContext facesContext = FacesContext.getCurrentInstance();
Boolean isPostback = facesContext.getRenderKit().getResponseStateManager().isPostback(facesContext);
if(!isPostback) //always seems to be true
if(dataTable != null)
row = dataTable.getRowData();
public void detail(ActionEvent event)
row = dataTable.getRowData();
if(row != null)
Map<String,String> record = (Map<String,String>)row;
setField1(record.get("field1"));
setField2(record.get("field2"));
public String edit()
//do the edit
return "success";
<h:inputText id="field1" value="#{detailBean.field1}" required="true"/>
<h:inputText id="field2Input" value="#{detailBean.field2}"/>
<h:outputText id="field2Text" value="#{detailBean.field2}"/>
<h:inputText id="field2Hidden" value="#{detailBean.field2}"/>
<h:commandLink id="edit" action="#{detailBean.edit}" value="edit"/>
<h:messages errorStyle="color:red" showSummary="true"/> The values from the table are successfully populated and the page is shown with the values:
field1=test1
field2Input=test2
field2Text=test2
field2Hidden=test2when I clear field1 (required) and submit I have the following:
field1=
field2Input=test2
field2Text=
field2Hidden=test2
message about field1 being requiredI would expect field2Text to still retain the value.
Furthermore if I set the field2Text to readonly="true", it, too, does not retain its value. However in both cases once the required field has a value and is submitted, all values show as expected.
My only workaround for this is to set my detailBean session scope, but that avoids the problem. -
What i thought was a bug..
Problem solved, don't know how to delete threads..
Message was edited by: DavidGunnardoChris --
This is more of a 'bug' in the spec. Prior to serialization, a persitent object will suck in all of its persitent fields and relations, so the entire object graph is serialized. Prior to cloning, though, nothing special is done. And of course, a clone is always transient (otherwise you'd get into situations where two objects in memory represented the same persistent state in the same PersistenceManager, which could lead to all sorts of problems).
In the 1.0 JDO specification, supported by Kodo 2.3, there is a PersistenceManager.retrieve(Object pc) you can use to load all persistent fields into an instance prior to making it transient or cloning it. Look for a new beta of 2.3 in the next few days. -
Bugs I've Encountered In PSE 7
Here are all the reproducible bugs I’ve encountered in PSE 7. Problems added since my last posting are marked [N]. In the coming weeks, I’ll evaluate PSE 8 to see how many of these have been fixed.
I’ve reported all of these problems to Adobe. Almost all of them are in the Organizer. Based on all the reports on the forums, it appears that the Editor has far fewer problems than the Organizer. Perhaps that’s because the Editor shares the same code with the professional Photoshop, while the Organizer appears to be an orphaned consumer product whose development Adobe has moved offshore.
For reference, my computer is a midrange Vista laptop with a dual-core 1.9 GHz processor, 4 GB of memory, and a 7200 RPM disk (the same speed as desktop disks).
File Management and Tagging
Major Problems
[N] The conversion of a catalog from PSE 6 to PSE 7 doesn’t properly handle photos that are stored within the PSE 6 catalog folder (as occurs when PSE 6 was used to restore a catalog from a backup). As a result, it isn’t possible to save photos in version sets, and PSE 7 will create duplicate thumbnails in the Organizer for an edited file.
File > Backup/Restore Catalog doesn’t backup and restore audio in slide shows.
Using Folder Location view to move a folder containing hidden files imported in the catalog will silently fail to move the hidden files and will leave the unhidden files disconnected in the catalog.
When converting a previous-version catalog whose photos are stored inside the catalog folder, as occurs when you’ve previously used File > Restore Catalog with New Location, PSE 7 needlessly copies all those photos. If you have a very large catalog, this will waste huge amounts of disk space; and if there isn’t enough disk space to do the copy, the conversion will fail instantly with a generic non-explanatory error message.
After reconnecting a file on a network drive, PSE will let you import the file a second time, failing to recognize it as a duplicate. The catalog is left in an inconsistent state, with two entries in the volume table for the network drive, one upper case and one lower case.
The Organizer fails very ungracefully when two drives have the same volume serial number, e.g. because of the use of a disk-cloning utility.
I’m very disappointed that Adobe didn’t add the three lines of code that would have issued a warning when it discovered two drives with the same number. While most people will never encounter this, when it strikes someone, its symptoms are very mysterious.
The Reconnect command fails to reconnect when: you move your photos to a new drive, assign the drive the letter of the old drive to the new one, and leave both drives connected to your computer.
If a hard drive containing photos in the catalog gets its drive letter reassigned by Windows, then the left-hand folder pane of Display > Folder Location view shows the old drive letter, not the new one, as containing the photos.
This sounds innocuous, but it led a user on the forums to select the wrong drive when dragging and dropping a folder, permanently deleting some files. This bug and the previous one with Reconnect have the same underlying cause – the Organizer doesn’t update drive letters in its catalog.
Folder Location view shows incorrect folder contents when more than one keyword tag is selected in the Keyword Tags pane.
Based on reports on the forums, there appear to be other situations where Folder Location view shows incorrect folder contents, but they haven’t been easy to reproduce.
The Restore Catalog command puts the wrong path into the restored catalog for folders whose name starts with the catalog name, and it will mark files in those folders as missing. For example if the catalog is named “houses”, and a folder containing photos in the catalog is named “houses photos”, files in the restored catalog will have the path “houses\ photos\” instead of “houses photos\”.
Minor Problems
In File > Export, the Common Basename can no longer be empty, and a hyphen is added automatically, so you can no longer get files named 1.jpg, 2.jpg, etc.
You can get files named 1.jpg, 2.jpg, etc. but you can’t get files named p1.jpg, p2.jpg, etc. – a hyphen is still added in that case.
Setting the Apply Metadata > Author or Copyright fields or the option Preserve Current Filename In XMP of the advanced options of the Photo Downloader causes duplicate files to be downloaded and imported.
Integration with Photoshop.com
Major problems
[N] Photoshop.com shows an incorrect date/time taken for photos with unknown month, day, or time.
For photos taken in years prior to 1970, the date/time shows as 12/31/1969 08:00 AM. For photos with year taken 1970 or later, the date/time shows as the the first month/day/minute/second in that time period in UTC that then gets converted to local time; e.g. “7/2/2005” gets shown as “7/1/2005 5:00 PM” for a user in PDT.
[N] Photoshop.com doesn’t obey the EXIF Orientation metadata field and can show photos rotated incorrectly, e.g. if they were rotated with the PSE option Edit > Preferences > Files > Rotate JPEGs/TIFFs Using Orientation Metadata.
Changes to the order of photos in a PSE album aren’t synced to Photoshop.com, and vice versa.
Adding a photo to a second synced album will cause Photoshop.com to lose any changes you’ve made with PSE’s Edit > Adjust Date And Time.
If you you rename a file that’s synced with Photoshop.com, Photoshop.com doesn’t pick up the new name (important if you’re using Photoshop.com as a backup service).
Only the top photo in a version set is synced with Photoshop.com (so you can’t rely on Photoshop.com to backup version sets).
When a stack is synced with Photoshop.com, all the photos in the stack are synced, but the fact that they were in a stack is lost (so you can’t rely on Photoshop.com to backup stacks).
Map View
Major Problems
Map view is unusably slow with more than about 6-8,000 mapped photos.
When I tested PSE 7 by importing 12,000 photos (the size of my catalog) that had GPS locations, the red push-pins didn’t blink uncontrollably as they did in PSE 6. But the Organizer crashed a number of times when I tried to zoom and pan the Map view until I was able to scroll the map to another country and then back again. The CPU continued to consume 80% of a single CPU (40% of two), even 10 minutes after I closed Map view. When I selected Limit Search to Map Area, PSE took over a minute to respond.
It takes 5 seconds to assign a keyword tag to a photo if there are 1,000 or more photos with map locations.
In PSE 7, it takes 3 seconds with 1,000 photos, 12 seconds with 8,000 photos, and 18 seconds with 12,000 photos.
Giving a map location to a tag assigned to hundreds of photos makes both the Keywords pane and the Map view unusably slow.
Giving a map location to a tag assigned to 2000 photos makes assigning any keyword tag to a photo very slow (see above). Searching on keyword tags takes 15 seconds, and a Show All takes 15 seconds.
Removing the map push-pin of a tag assigned to hundreds of photos mistakenly tries to update the metadata of each photo, which can take tens of minutes and fill up your recycle bin.
The weird thing is that assigning a map location to the tag doesn’t try to write the metadata of files with that tag.
Moving map locations always fails, silently (too bad if you want to record backcountry locations not namable via a place name).
If a photo with GPS coordinates is imported, it doesn’t show as a red push-pin on the map.
Scrolling the map or restarting the Organizer causes the pin to appear.
Color Management
Major Problems
The Organizer considers photos produced by cameras set to the color space Adobe RGB as untagged with a color profile, showing their colors incorrectly.
Metadata
Major Problems
When setting the Organizer’s date/time for imported photos containing EXIF:DateTimeOriginal but not XMP:DateTimeOriginal, PSE 7 doesn’t properly handle US Daylight Savings Time in years prior to 2007.
Edit > Adjust Date and Time doesn’t properly handle US Daylight Savings Time when setting date/times many decades ago, e.g. October 5, 1961 12:00 PM.
In Thumbnail View, the Organizer incorrectly orders files that have unknown time. For example, it will show a file dated 12/25/1976 8:00 PM PST after a file dated 12/26/1976 (unknown time) on a computer in time zone PST.
If you set a file's date/time fields to all unknown with Edit > Adjust Date And Time, the value of EXIF:DateTimeOriginal isn't cleared by File > Write Keyword Tag.
The File > Write Keyword Tag command doesn’t create or update .XMP sidecars for Nikon D300 or D80 .NEF raw files (but it does for Canon CR2 raw files).
New dates set by Edit > Adjust Date and Time sometimes get silently lost.
See http://www.johnrellis.com/psedbtool/photoshopelements-6-7-faq.htm#_Date/times_of_files for details and workarounds.
[N] Edit > Adjust Date and Time > Shift To New Starting Date And Time and > Shift By Set Number Of Hours writes the time in UTC rather than local time, as required by the standard, causing other tools to show the wrong date/time.
File > Write Keyword Tags and Properties fails to write the map location (GPS coordinates) of TIFFs (tested with TIFFs produced by the PSE Editor and the Nikon Coolscan V, Nikon Coolscan 5000, and Epson 4490 scanners).
File > Write Keyword Tags doesn’t write the map location (GPS coordinates) into the .xmp sidecar files of Nikon D80 or Canon G9 raw photos.
[N] The Organizer ignores correctly formatted EXIF GPS locations in some files that many other programs can read without trouble; this may be because the EXIF is encoded in little-endian byte order.
Minor Problems
[N] Deselecting the option Use “Last Modified” Date If EXIF Date Is Not Found does not work – the last-modified date is always used for an imported file that doesn’t contain an EXIF date.
The Properties – Metadata window shows the wrong value for Date Time Original when the month, day, or time is unknown.
Edit > Adjust Date and Time applied to a file with time “unknown” decrements the date by one day.
When you save an edited photo as a new file but not in a version set, the edited version picks up the date/time and caption of the original but not the keyword tags, star rating, or notes.
With the Import EXIF Caption option turned off, captions pre-existing in a file’s EXIF:ImageDescription metadata field reappear in the Organizer after invoking the Full Editor.
Searching
Major Problems
The timeline doesn’t correctly display date ranges spanning many years.
Using the Keyword Tags pane, you can no longer exclude a parent category from a search, e.g. exclude all photos tagged with any tag in the People category.
The User Guide gives two examples of how to use category exclusion that just don’t work in PSE 6 or 7. Interestingly, you can almost work around this by doing a text search of “not <category tag>”, e.g. “not People”. But this will also exclude photos that happen to have the word “people” in their filenames, captions, or notes, probably not what you want.
Find > By History > Imported On sorts the photo dates alphabetically, rather than by date/time, and it takes a couple of minutes on a large catalog, making the command almost useless; and it shows a scary message “Deleting Keyword Tags”.
[N] Searching with the Find Bar for “1 star and lower” doesn’t show photos with 0 stars.
Find > By Details > Rating Is Lower Than 1 Star does work.
Searching for “0 stars only” doesn’t work for photos in a catalog converted from PSE 5.
Searching with date ranges doesn’t properly handle time “unknown”.
Minor Problems
The Show All button sometimes doesn’t appear if you quickly type a query into the Search text box and hit Enter.
Setting a date range and then excluding two tags from the search clears the date range.
Show All doesn’t clear a date range set by Find > Set Date Range (can be very confusing).
Save Search Criteria As Smart Album isn’t available if a date range has been set but no other search criteria have been.
When searching, you can exclude a keyword tag and then include an album, but you can’t do it in the other order.
User Interface
Major Problems
On 1024x768 displays, and sometimes at higher resolutions, the Editor Print window is positioned to perfectly obscure the scroll bar of the drop-down list of printer profiles, leading people to think their printer’s profiles aren’t available.
This bug occurred with higher-resolution displays in PSE 6.
The menu bars don’t display if you have changed the screen DPI to be larger than 96 (as many people do on today’s ubiquitous high-res displays).
There is a workaround for both PSE 6 and 7. Set the option > Edit > Preferences > General > Use System Font. The result isn’t that pretty, since some of the application text will be large and some will be small, but it’s functional, and you’ll still be able to use the higher DPI with all your other applications.
Minor Problems
In the Properties window of Full Screen mode, if you click in the Notes field and do Ctrl-A to select all the text, then click in the Caption field and do Ctrl-A to select its text, the notes remain highlighted.
In the Organizer, Help > System Info reports the wrong amount for “Built-in Memory” on computers with more than 2GB of installed memory. (The command reports the correct amount in the Editor.)
The setting of View > Show Borders Around Thumbnails isn’t remembered after you restart PSE.
Copying a paragraph from Microsoft Word and pasting it into the Notes field of a keyword tag causes some of the spaces between words to be deleted.
Ctrl-A to select all text doesn’t work in the text fields of the Properties window.
Interestingly, you can use ctrl-A in the text fields of the full-screen mode Properties window.
In the full Editor, if at least one open photo is minimized to the Project Bin, Ctrl-Tab no longer cycles through the open windows.
Dialogs in the Editor sometimes bounce back when you try to move them.
Dates are shown by Display > Import Batch as 2\20\2008 rather than 2/20/2008.
In Create > Slide Show > Slide Show Preferences dialog, you can't use backspace or delete to clear the text in the Static and Transition Duration fields -- you need to select the text and then type over the selection (non-standard Windows behavior).
Escape doesn’t close the full-screen-mode Properties dialog.
You can’t use the Windows Explorer Tile command to tile the PSE Organizer and Editor windows.
The Editor window can’t be resized the standard Windows way by grabbing any edge, just the lower-right corner.
Alt doesn’t underline the shortcut letters of top-level menu items in the Full Editor (but it does in the Organizer).
The width of the Editor's Palette Bin can't be adjusted by dragging the left edge, as you can with all the other similar panes (Project Bin, Organizer Bin, Map View).
Slide Shows
Major Problems
The duration of a video clip included in a Slide Show is set to the default duration, not the length of the video clip, and right-click Edit Duration doesn’t change the clip’s duration.
For video clips, PSE 7 doesn’t enable the command to change the duration as it does for photos. But you can work around this by using Add Media > Photos And Videos From Folder to add the clip to the show.
An audio caption attached to a photo isn’t imported into a slide show even though the option Include Audio Captions as Narration is selected.
Editor
Minor Problems
[N] The Editor crashes if you invoke Quick Fix and use just one Touch Up tool (e.g. Whiten Teeth) and then save the file (but your changes are correctly saved).
The File > Save For Web command doesn’t remember the last settings (e.g. file format and quality) after your restart the Editor.
For a workaround, see http://www.johnrellis.com/psedbtool/photoshopelements-6-7-faq.htm#_Allow_Save_For.Just kidding. But, are you going to give us a bug comparison between 7 and 8 ? I will probably sit version 8 out and get 9 next year. Thanks for the list!
Juergen -
BUG: Forced Restart Via Creative Cloud Desktop Application Installer
Hey guys, I was trying to install a few adobe products and have been encountering strange errors.
The following bug has been experienced when trying to install Adobe Application Manager, Adobe Photoshop, Adobe Dreamweaver, and Adobe Illustrator.
Program will install to about 44% or so and then say "extracting". After extraction, it displays the %installation, increases about 4-10%, then forcefully restarts my computer.
According to my computer, no errors are present (looked in the events log)
Here are the following things I have tried to remedy this issue.
Running windows update for all general updates and going to intel to update graphics driver.
1. a. Removing all Adobe applications via control panel and using the Adobe CCleaner utility.
b. Re-installing Creative Cloud w/ administrator privileges
c. Attempting to reinstall application.
2. Tries the above but in Safe mode.
3. a. Looked at computer resources to make certain CPU wasn't somehow overheating.
b. CPU temperature was not adversely affected nor did the utilization rate of any resources(CPU, GPU, Drive, Network) exceed 25%
c. Viewed the event logs to determine if any errors were present. The even log displayed no errors had occurred during any of the forced restarts.
On a side note- I found that despite Skype being told not to start-up on boot, it is actually reducing my boot time by 6 seconds!
4. Went into advanced settings and unchecked the box that allows windows to automatically restart on error.
I am at a loss as to why it is forcibly restarting my computer. Adobe products are the ONLY products that I have experienced any issues with during installation.
However, in order to remedy this issue here is some information regarding computer specs.
OS Name
Microsoft Windows 8.1
Version
6.3.9600 Build 9600
Other OS Description
Not Available
OS Manufacturer
Microsoft Corporation
System Name
ZAYIK
System Manufacturer
Dell Inc.
System Model
Inspiron 5520
System Type
x64-based PC
System SKU
Inspiron 5520
Processor
Intel(R) Core(TM) i7-3632QM CPU @ 2.20GHz, 2201 Mhz, 4 Core(s), 8 Logical Processor(s)
BIOS Version/Date
Dell Inc. A14, 5/13/2013
SMBIOS Version
2.7
Embedded Controller Version
1.01
BIOS Mode
UEFI
BaseBoard Manufacturer
Dell Inc.
Platform Role
Mobile
Secure Boot State
On
PCR7 Configuration
Binding Not Possible
Locale
United States
Hardware Abstraction Layer
Version = "6.3.9600.17196"
User Name
Zayik\ZAYIK\David
Installed Physical Memory (RAM)
8.00 GB
Total Physical Memory
7.87 GB
Available Physical Memory
5.69 GB
Total Virtual Memory
9.12 GB
Available Virtual Memory
6.76 GB
Page File Space
1.25 GB
Graphics Card
Intel HD 4000
This is quite a fatal bug and I am hoping we can find a solution so others will have a place to turn to if they encounter the same bug.Hi pillguy,
Please let us know the network environment are you in? Which operating system are you using? Do you utilize a proxy server or a software firewall?.
Please try the steps mentioned in the kb: http://helpx.adobe.com/creative-suite/kb/remote-server-responding-installing-aam.html / http://helpx.adobe.com/creative-suite/kb/error-update-server-repsonding-cs4.html .
You may even try download using different internet connection if possible.
Regards,
Romit Sinha
Maybe you are looking for
-
Just bought a new TV - old one died - which has all the whizzy features you expect of a new TV like HD and HDMI inputs so bought some HDMI cables. Well the picture is better but unlike the DVD player which when you turn it on autoswitches the TV to t
-
Weblogic fails to start after new deployment
Hello, We are experiencing problems after a new deployment please see the log snippet below. <Jun 14, 2012 4:14:30 PM PDT> <Error> <HTTP> <BEA-101017> <[weblogic.servlet.internal.WebAppServletContext@6248fe00 - appName: 'SPLWeb', name: '/spl', contex
-
Difference between material(/SAPAPO/MAT1) & material(/SCWM/MAT1)
Hi Experts, I am new to EWM & I am confuse between Product(/SAPAPO/MAT1) & Product(/SCWM/MAT1) ?. Please anyone outline me some brief about above to product in EWM. Regards, Rock
-
So a few years ago, we had one email address that was my user ID and that apps were downloaded to. Then I got a new phone, cancelled that email provider (not thinking about my phone and itunes) and signed up with Gmail, therefore giving myself a new
-
Calendar images: RGB or CMYK? TIFF or JPEG?
Why am I not finding this info anywhere? Prior to making a calendar and/or book, I'm going to edit my images in Photoshop, then save them out, but want to know in advance the ideal size (300 dpi at final size? At what line screen are the books/calend