MLS Startup

[hdradio]

I want to ask what is MLS doing at startup.
I know it scans for the files in libraries but does it rescan ALL the files and read ALL the tags from beginning or looks for changed files and updates that files only ?
Every time I update PF, MLS scans my library it that takes about an hour.
Sure it's a background task but why to scan all the files and not only the modified and new ?

[radio42]

How do you know that MY FILE is absolutely correct ?

If they are not correct an error would indicate that. And in addition I just tried the notepad++ XML Tool here on my 'correct' playlist files and the tool showed the same error - so it is proven, that the tool shows an XML validation error on a correct XML file!

If you remember there was the same issue in past with _synced_.pfp XML file and you provided a fix. We had a teamviewer session and you had copied my _synced_.pfp file and confirming that there was a problem in that xml file

No, that was not a problem in the _synced_.pfp (xml) file!, but just an issue with evaluating its content for different UNC and local paths! That's why I could fix it via a new software version only!

XML is NOT well formed...

This has nothing to do with a valid XML or not! It is still totally valid, as you can enable or disable the character checking for an XML element note. And ProppFrexx CAN fully and CORRECTLY read its own data!
As another example: you might run an CSS or HTML validator on various web-sites...many of them are working 100% perfect with all different browsers, but a strict HTML/CSS validation would fail...
Note, that the extension is .pfp and NOT .xml !

There is also NO such way to break another once ProppFrexx system!
What you are try to construct is NOT possible, as PF would automatically convert the given comment data!
Try it out yourself before you claim that there is a serious or general problem - there is no!
You can NOT break ProppFrexx by sending someone an mp3 file!Please stop claiming things, which are based on guessings and are simply false.
You might be frustrated because the rescan performance is not good with the latest version, but that is no reason to say things which is totally incorrect.

And in the end you wouldn't even have to care about how ProppFrexx saves its data...or do you care how exactly Oracle, Outlook, Dalet, Microsoft Word or any other system saves its data...?
The ONLY important part is, that the system works as expected.
So I wouldn't care, if any XML validation tool claims, that invalid characters are being used - just because these tools do use by default a strict character checking of XML version 1.0, while ProppFrexx does NOT use the strict character checking and also uses the newer XML v1.1. standard! And this for a very good reason ...but I guess this goes beyond any interest here, as it has pure performance and internal development reasons...
Please see here: http://www.w3.org/TR/xml11/#charsets

Long talk, just try the new v3.0.14.0...

[radio42]

It takes exactly the same time as before.

Which seems to be okay, as you claimed, that in the last version it took hours and never completed!
So that is back to normal then...! And is the main bug-fix to the v.31/v.32 versions! So that seems to be solved then?!

Rescanning of a folder based media lib involves scanning the hard disk folder and all sub-folders for all contained tracks. The time it takes is mainly limited by the speed of the hard disk.
In my tests here on my machines I achieve around 10.000 tracks in 45 sec.!
(that is e.g. on a standard laptop Win7-64bit, i5-CPU, 8GB RAM with a 7.200 rpm SATA III hard disc)
On the same system with an SSD drive I achieve around 10.000 within 30 sec.!

So when you say, that on your machine a rescan takes 45 min. for just 50.000 tracks (and if I understand you correctly, that it was always taking that time also in previous versions),
then I can only assume, that your I/O sub-system is the real limiting factor and your hard disk might be quite slow?
What I/O sub-system are you using, what type of hard drive, how is it connected etc.?
And what machine and OS are we talking about?

Further in MY tests I do can perform a search on a remote media lib which is currently being rescanned.
Of course that search is a bit slower (as the MLS and the I/O system on the server is utilized), but is does return quite timely, i.e. I experience a delay of around 1 or 2 sec. maybe!
And it does return valid tracks, which are matching the filter (which was the other bug you posted - so at least here in my tests this bug is also solved).

So it is a bit hard to say, why in your case resp. on your hardware the search would be so slow if the MLS is rescanning.
Maybe your I/O system is so much utilized, that it blocks everything?
At least I can not reproduce the issue on my systems.
When you open the Windows Performance Monitor what does it show for I/O and CPU?

[hdradio]

Take a look because the new version shows "rescanning..." and have passed 4 hours since start.

[radio42]

The MLS does a 'Rescan' of all media libs at startup.
This to detect any changed, updated, removed or new files.
For folder libs it has to scan all files of that folder, but only reads TAGs for new or changed files.

But I can take a look, if it is possible to speed that up...

[hdradio]

Bernd there must be something wrong with v.32
A few hours ago I updated my on air pc to v.32 Unfortunately I can not use find anymore. I wait for half an hour without any results.
The same to Find/Expression builder.
Usually when I start MLS it takes about 45 minutes to scan my folders. During that time I see the Hard disk LED flashing.
But now after update the LED does not flash and almost two hours have passed and MLS shows "rescanning..."
At least it selects tracks from scripts, but I am very worried if I have to let PF on air or switch to v.31

Any ideas what is wrong ?

Edit:
I restored v.31
I don't feel confidence to let this version on air.

Also closing PF had a crash:

Code: Select all

02/07/2014 01:42:19 [v3.0.13.32]:********** Error/Exception Information **********
Error Message: Object reference not set to an instance of an object.
Error Type: System.NullReferenceException
Error Source: DevExpress.XtraEditors.v14.1.Void set_EditValue(System.Object)
********** System Stack Trace **********
   at DevExpress.XtraEditors.PictureEdit.set_EditValue(Object value)
   at DevExpress.XtraEditors.PictureEdit.set_Image(Image value)
   at ProppFrexx.p.a(Object A_0, FocusedNodeChangedEventArgs A_1)
   at DevExpress.XtraTreeList.TreeList.RaiseFocusedNodeChanged(TreeListNode oldNode, TreeListNode newNode)
   at DevExpress.XtraTreeList.TreeList.AppendNode(Object nodeData, TreeListNode parentNode, Object tag)
   at ProppFrexx.p.a(Object A_0, ProgressChangedEventArgs A_1)
   at System.ComponentModel.BackgroundWorker.OnProgressChanged(ProgressChangedEventArgs e)
   at System.ComponentModel.BackgroundWorker.ProgressReporter(Object arg)
********** End **********

[hdradio]

I am afraid that v.32 had made a harm to my system. Now even v.31 does not work normally.
Since last night that I restored v.31 (it's now 09am) MLS still shows "rescanning..."
Any help ?

Update:
On my backup pc I have v.32 (MLS v.14)
The same problems here.
I did a small test to check if there is something wrong with _synced_.pfp file.
So I deleted it and run MLS.
It took about 45 minutes to scan the tags, after that a new _synced_.pfp file was created but MLS is still showing "rescanning..."
Still find takes ages to give results.
Closing PF same error like another user reported:

Code: Select all

02/07/2014 09:42:08 [v3.0.13.32]:********** Error/Exception Information **********
Error Message: Object reference not set to an instance of an object.
Error Type: System.NullReferenceException
Error Source: DevExpress.XtraTreeList.v14.1.Void RaiseInvalidValueException(DevExpress.XtraEditors.Controls.InvalidValueExceptionEventArgs)
********** System Stack Trace **********
   at DevExpress.XtraTreeList.TreeListContainerHelper.RaiseInvalidValueException(InvalidValueExceptionEventArgs e)
   at DevExpress.XtraEditors.Container.ContainerHelper.OnInvalidValueException(IWin32Window owner, Exception sourceException, Object fValue)
   at DevExpress.XtraTreeList.Data.TreeListDataHelper.InvalidValueException(Exception e, Object value)
   at DevExpress.XtraTreeList.Data.TreeListData.SetValue(Int32 nodeID, Object columnID, Object value, Boolean initEdit)
   at DevExpress.XtraTreeList.TreeList.SetNodeValue(TreeListNode node, Object columnId, Object val, Boolean initEdit)
   at DevExpress.XtraTreeList.TreeList.OnSetValue(TreeListNode node, Object columnID, Object val)
   at DevExpress.XtraTreeList.Nodes.TreeListNode.SetValue(Object columnID, Object val)
   at DevExpress.XtraTreeList.Nodes.TreeListNode.set_Item(Object columnID, Object value)
   at ProppFrexx.p.a(Object A_0, ProgressChangedEventArgs A_1)
   at System.ComponentModel.BackgroundWorker.OnProgressChanged(ProgressChangedEventArgs e)
   at System.ComponentModel.BackgroundWorker.ProgressReporter(Object arg)
********** End **********

[hdradio]

Bernd I am still looking for a solution because I am afraid letting PF on air.
What I did now was to open _synced_.pfp in notepad++ and using the xml tools plug in, I validated it.

The result :

Code: Select all

---------------------------
XML Tools plugin
---------------------------
XML Parsing error at line 565596: 
Extra content at the end of the document

I don't see something strange on that line :

Code: Select all

<pf:bitRate>128</pf:bitRate>

I am afraid that those plain text XML databases you use, cause more problems that good

Any help now ?

[hdradio]

I further do not know, what notepad++ resp. the XML Tool plugin is trying to validate or finds incorrect with the XML file - as it is absolutely correct!

How do you know that MY FILE is absolutely correct ?

There have never been any issues with the XML files and ProppFrexx reads them correctly! As such I can only disagree with you, that XML databases cause problems!

If you remember there was the same issue in past with _synced_.pfp XML file and you provided a fix. We had a teamviewer session and you had copied my _synced_.pfp file and confirming that there was a problem in that xml file.

PM dated 29 Apr 2014 :
The "_synced_.pfp" playlist file was indeed corrupted in a certain way.
...
Note the sudden nesting of tracks...I have no real clue how that cold have happened.
When looking to the code a filename/path normalization function must have failed internally, which should/could normally never happen (when calculating an UNC path)! This is very strange and might infact again another Microsoft bug.

It might be again another Microsoft bug.

[hdradio]

Checked also the XML with that tool :
http://download.cnet.com/XML-Validator/ ... tml?tag=bc
and result is :
Invalid XML document:
At line 158905 column 71, invalid character reference
At line 1464864 column 71, invalid character reference
XML is NOT well formed, 2 error(s)!

That line has :

Code: Select all

<pf:comment>..:: blablabla ::.&#x4;</pf:comment>

As you see, the comment has not acceptable characters for XML file.
Do you understand now why XML databases can cause problems ?
Unless you write a bible informing the users what characters not to use in tags.

Comment by MODERATOR:
WHAT THE AUTHOR OF THIS POST TRYS TO CLAIM HERE IS INCORRECT.
There is no need to write a bible informing the users what characters not to use in tags!
The user can simply use ANY characters in the TAG data - ProppFrexx handles (read/writes) that correctly!
It is just that most XML validators handle those characters differently, as certain characters are marked as 'encouraged to avoid' - but they are still fully legal!
XML 1.1 explicitly defines: "XML processors MUST accept any character in the range specified for Char; which is any Unicode character, excluding the surrogate blocks, FFFE, and FFFF".
And ProppFrexx is ensuring that!