MC doesn't find phrases and sentences any more

[Skinman]

Hi,

MC doesn't find phrases and sentences within files any more, at least not reliably.

I have been using MultiCommander for some time as a workaround of a deficiency of Windows 7, compared to Windows XP:

The search function of the Win XP explorer was able to do a full-text search within files and find phrases and sentences. The search function in Windows 7 is apparently not able to do that. It does a Boolean 'AND' search and is generally somewhat buggy and useless to me.

However, of late, MC has become unable to find files that contain a certain phrase or sentence.

I'm currently using version 6.9.1 build 2306 and it's at least the second automatic update that has failed in this way.

Example: the directory c:\Test_1 contains the subdirectory c:\Test_1\Test_2. Both contain copies of the two files "New Text.doc" and "Old - Text.doc". All of these files contain the search phrase "gas injection".

Searching for that phrase though only turns up the file "New Text.doc" but not the file "Old - Text.doc".

Am I doing something wrong here or what is it? Because, as I said, MC used to do this just fine until recently (though I couldn't say when it started to fail.) Any help appreciated.

[Ulfhednar]

I got caught out by 'ignore case' (fall ignorien) not being checked by default (or being persistent if checked) in these newer builds, + setting dir depths to search.
Maybe that applies?

As I am so familiar with the old dialog, the new one has slapped me in the face 

[Skinman]

Thanks for the tipps but none of those seem to be it. Attached find screenshots of the settings.

However, I noticed a typo in the localisation, in the German word for 'directory depth'. It should be "Verzeichnistiefe" but there is one extra 'r' that doesn't belong.

Could that be it? Maybe I'll just quickly test the English version.

Edit: I get the same failure with the language set to English.

[Ulfhednar]

Just tried on pdf & docx - & I did not find anything - This happens in both simple & advanced dialogs.
I normally use the content search for things like txt lua html xml etc so I hadn't noticed this behavior.

Do you remember in which build it stopped working for you?
Maybe there is some code for accessing these proprietary formats that is now unhooked?

(forgot to say I'm using admin rights & searching unprotected files.)

[Skinman]

Sorry, I don't recall the build it last worked.

[Mathias (Author)]

MC do not and have never decoded doc files.  doc files are binaries files with the own special format.
If you are lucky it might find the wanted text in the file, What it does is a raw binary search of the files. And it depends on how words have saved the file.
You can try to force the binary matching where you select what type of content it is. If it detects is as UTF8 it will try to convert text and then it will go wrong.

[Ulfhednar]

Thanks for that info Mathias. 

Could MC be able to read pdf? (i.e if they have had OCR or input text contents.)
I'm curious to know if MC would see & use this http://www.pdflib.com/products/tet-pdf-ifilter/ if installed or is your search process totally discrete from Win search processes.
Thanks

[Skinman]

MC do not and have never decoded doc files.  doc files are binaries files with the own special format.

Both wrong. There is lots of plain text in doc files and it used to work perfectly with earlier versions, as I wrote. Still does with some files, as I wrote.

[Mathias (Author)]

It depends on word format and it also depends and what format MC thinks the file is(text,utf8, unicode, binary), and it also depends on how the text in the word doc is formated. Modern Word files are actually zip files. and with them you will not find anything since there is no chance the text exists as plain text.

That you find it before is pure luck.

But if you force it to do a binary check it might find the text as long as the text only have A-Z , 0-9 and no special characters.
and that the text document is not encoded with newer word versions.