Vista 64 bit, Foxit iFilter and Adobe Acrobat Professionel

[Hafnium]

Hi,

I'm interested in buying the iFilter PDF for use in my company. We have recently deployed Windows Vista x64 on a number of machines, and also installed Acrobat Professional version 8. Adobe does not, however, include a working 64 bit iFilter for PDF-files. Also we have found the performance of the Foxit iFilter to surpass that of Adobe's.

We therefore tried Foxit iFilter; however, it seems that Acrobat somehow disables the function of the Foxit iFilter.

Our purchase of the iFilter for use in our company is dependent on the iFilter to coexist with Acrobat Professional on Vista x64.

Can you please provide a solution?

Also, you state in your FAQ that the Adobe iFilter needs to uninstalled prior to installing the Foxit iFilter. You do not, however, provide guidelines on how to do so. Could you please post a guide?

Thanks.

Hafnium

[AmyLin]

Thank you for your concern for our products.
I will report this issues to our development team, and give you a solution as soon as possibile.

[markrlondon]

I'm seeing the same problem on Vista x64.

What is particularly annoying is that this worked previously: In August 2007 I installed Adobe Acrobat 8.0 Professional as part of Adobe Creative Suite 3 Design Premium. In order to be compatible with Vista x64 this required a Microsoft hotfix (KB930627) and then the upgrade to Acrobat 8.1.0 Professional. Later, in November 2007, I installed the Foxit PDF IFilter x64 and it all worked fine - PDFs were indexed by Windows Search. After that time I never upgraded any Adobe software and the Foxit PDF IFilter continued to work perfectly.

However, a few days ago I noticed that PDF thumbnails were no longer being generated and so I decided to uninstall and reinstall the Adobe software to try and fix the problem (which was unnecessary I later found out). After this I found that the Foxit PDF IFilter no longer worked!

I've tried installing the Foxit PDF IFilter x64 first and then installing the Adobe software, as well as vice versa, and nothing seems to make any difference: The Foxit PDF IFilter won't work when the Adobe software is installed.

I don't know what's changed. I'm using the same Adobe software that I was before (I'm intentionally not updating it to the latest versions since I did not do so previously). I'm using the Foxit PDF IFilter x64 1.0.0.1122 but I can't be sure if this was the version I had working correctly the first time round.

[markrlondon]

I see there is an (inconclusive) thread on this on the Adobe forum: http://www.adobe.com/cfusion/webforu...&enterthread=y

[markrlondon]

I have found what appears to be a workaround, but it's not satisfactory because it's too easy to break by accident. Here it is.

The order of the following steps might not be critical. This is simply what I tried.

1) Install Adobe's software first (e.g. Reader, Acrobat, etc.).

2) Then install Foxit PDF IFilter x64 1.0.0.1122. I believe this is the current bversion as at the time of writing.

3) Do NOT change any settings for .pdf in the 'File Types' tab of Windows Search's 'Advanced Options' dialog. You can look but you must not change anything. The Filter description should be shown as "PDF Filter" but neither the 'Index Properties Only' nor the 'Index Properties and File Contents' radio buttons will be selected. You should find that PDFs are correctly indexed with the settings in this (seemingly confused and incorrect) state.

4) It's tempting to click 'Index Properties and File Contents' but do not do this. If you do so then the Filter Description will change to "Registered IFilter is not found" and PDFs will not be indexed. What actually occurs when this happens is that Windows deletes the PersistentHandler entry in the Registry for .pdf. Clearly this should not be happening and I can't see why it does happen. As far as I can tell, Foxit's PDF IFilter sets up the Registry correctly according to http://msdn2.microsoft.com/en-gb/lib...s_installation. Clearly Foxit PDF IFilter does work correctly when Adobe's software is not installed.

5) If you do make a mistake then do the following to get PDF indexing to work again:
(a) Set HKCR\.pdf\PersistentHandler
(Default) = "{3C3DCD30-E62D-49a4-A746-039334AFDE59}"
OriginalPersistentHandler = "{3C3DCD30-E62D-49a4-A746-039334AFDE59}"

(The OriginalPersistentHandler value isn't critical. It's just a handy reminder for if/when the (Default) value gets deleted. Also Windows will use it if it's working correctly).

(b) Stop and start the Windows Search service (WSearch) using whatever method you prefer (e.g. Task Manager, Services applet, net stop wsearch/net start wsearch).

(c) You can now safely force a rebuild of the index, if need be, by clicking the Rebuild button on Windows Search's 'Advanced Options' dialog.

(d) This workaround also survives clicking the 'Restore Defaults' button and rebooting.



There we go, this seems to work. But clearly it's not very comfortable. I wonder what has changed to cause this? It can't be Adobe's software since the version I'm using is the same as the one that worked previously. I'm not sure if the version of the Foxit PDF IFilter is the same. Perhaps there is an emergent effect due to a hotfix that Microsoft have supplied since last November, when all this worked perfectly for me. (N.B. I'm still not on Vista SP1 here).

[Hafnium]

I wonder if you've installed SP1 for Windows? I have and maybe MS has changed something...

So far, your solution does not seem to work for me. I will explore it in greater detail, though.

Hafnium

[markrlondon]

I've not installed SP1 yet.

[markrlondon]

There is a problem with the workaround I described: Although the workaround works for me to an extent, it doesn't seem to be stable. I inevitably get an APPCRASH for SearchIndexer.exe in the TQUERY.DLL module whilst the initial crawl is taking place. This *never* happened until I started using this workaround.

Unfortunately, after SearchIndexer is restarted, it always thinks that the crawl is complete. The only way to continue is to start over. In practice this means that I can never get all my files indexed - SearchIndexer.exe always crashes before it finishes the crawl.

There has got to be a proper solution for this weird problem, somehow.

[markrlondon]

I think I may have found a solution to this weird problem. I'll describe it below but I want to make it clear that I don't know if I've found the complete solution. I am currently testing to see how stable it is.

Here is the possible solution: This is a registry permissions issue. I traced all references to the Adobe PDF IFilter in the registry and I discovered that there are some extra keys that are not (as far as I know) documented. These keys must be set correctly in order for Windows Search to be able to use an IFilter for content indexing but, due to the wrong permissions, the values in the keys could not be set. I do not know if the Foxit PDF IFilter installer attempts to set these keys - I haven't tested this.

Here are the details:
1) See HKCR\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000} and HKCR\Wow6432Node\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000}. These two keys are a class definition for the "Adobe Acrobat Document".

2) Both keys have a sub-key called PersistentHandler whose (Default) value contains the GUID for the Persistent Handler for this document class (.pdf). Even though I had installed Foxit's PDF IFilter, the GUID was still pointing to Adobe's 32bit PDF IFilter: {F6594A6D-D57F-4EFD-B2C3-DCD9779E382E}. See HKCR\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000}\PersistentHandler and HKCR\Wow6432Node\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000}\PersistentHandler.

3) I tried to change the GUID to the one for Foxit's x64 PDF IFilter, {3C3DCD30-E62D-49a4-A746-039334AFDE59}, but Regedit prevented me from doing so.

4) I checked the permissions on the two keys and I found that they had both had their permissions set to not inherit from their parent, and Administrators only had Allow Read privs. The default that they should have is to inherit permissions from parent and for Administrators to have Allow Full Control.

5) I changed the permissions back to Administrators to have Allow Full Control.

6) I was then able to set the (Default) values for both keys to the Foxit IFilter's GUID, {3C3DCD30-E62D-49a4-A746-039334AFDE59}. [Note that for 64bit purposes it should only be necessary to set the (Default) value of the HKCR\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000}\PersistentHandler key. I did the Wow6432Node version too just to be absolutely sure]

7) I looked in the 'File Types' tab of Windows Search's 'Advanced Options' dialog and I found that I was able to select 'Index Properties and File Contents' option for .pdf as normal! It all seems to be back to normal.


Summary: Something changed the permissions on the HKCR\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000} and HKCR\Wow6432Node\CLSID\{B801CA65-A1FC-11D0-85AD-444553540000} keys such that the values could not be changed. This prevented any PDF IFilter (except for Adobe's) from being properly configured.

I'm currently testing the stability of this configuration and will report back with my findings.

================================================== ============================
Update 20080304 13:30 GMT:
This does not appear to be any more stable that the workaround I suggested. I found that SearchIndexer.exe crashed with TQUERY.DLL about a quarter of the way into its initial crawl, just as before. Sorry!

Obviously there's still something I've missed.

[Hafnium]

Too bad. Thanks for your time and effort. Hopefully, you'll be able to investigate further.

I'm not sure if I got a SreachIndexer.exe crash but I was able to get the iFilter working as indicated by the "Properties and File Contents"-setting working correctly. The keys you mentioned did not have permissions set to denied; they were, however, set to the Adobe 32 bit iFilter.

I do not get thumbnails of PDF files. I haven't tested for too long, but it does not appear that PDF files are indexed properly just yet. I will return when I the indexer has hopefully indexed all files.