Bug report: Skipping and Scanning files with $FileType

[Hank]

Version: 10.40
OS: Linux, Windows
Contents of the directory:

Code Select Expand


gsdll32.dll
gsdll32.lib
gsdll64.dll
gswin32c.exe
gswin32.exe

(Ghostscript in this example)

Command:

Code Select Expand

exiftool -r -SourceFile -AssemblyVersion -Comments -CompanyName -FileDescription -FileName -FileSize -FileType -FileVersion -FileVersionNumber -InternalName -LegalCopyright -LegalTrademarks -if '$FileType=~/DLL|EXE|LIB|OCX|SO|AXF|BIN|ELF|PRX|PUFF|DYLIB|MSI/i' -json .


Output:

Code Select Expand


[{
  "SourceFile": "./gsdll32.dll",
  "FileName": "gsdll32.dll",
  "FileSize": "11 MB",
  "FileType": "Win32 DLL"
},
{
  "SourceFile": "./gsdll64.dll",
  "FileName": "gsdll64.dll",
  "FileSize": "12 MB",
  "FileType": "Win64 DLL"
},
{
  "SourceFile": "./gswin32.exe",
  "FileName": "gswin32.exe",
  "FileSize": "144 kB",
  "FileType": "Win32 EXE"
},
{
  "SourceFile": "./gswin32c.exe",
  "FileName": "gswin32c.exe",
  "FileSize": "136 kB",
  "FileType": "Win32 EXE"
}]
    1 directories scanned
    4 image files read

Skipped the LIB file.

Command:

Code Select Expand

exiftool -r -SourceFile -AssemblyVersion -Comments -CompanyName -FileDescription -FileName -FileSize -FileType -FileVersion -FileVersionNumber -InternalName -LegalCopyright -LegalTrademarks -json .

Output:

Code Select Expand


[{
  "SourceFile": "./gsdll32.dll",
  "FileName": "gsdll32.dll",
  "FileSize": "11 MB",
  "FileType": "Win32 DLL"
},
{
  "SourceFile": "./gsdll64.dll",
  "FileName": "gsdll64.dll",
  "FileSize": "12 MB",
  "FileType": "Win64 DLL"
},
{
  "SourceFile": "./gswin32.exe",
  "FileName": "gswin32.exe",
  "FileSize": "144 kB",
  "FileType": "Win32 EXE"
},
{
  "SourceFile": "./gswin32c.exe",
  "FileName": "gswin32c.exe",
  "FileSize": "136 kB",
  "FileType": "Win32 EXE"
}]
    1 directories scanned
    4 image files read

Command:

Code Select Expand

exiftool -r -SourceFile -AssemblyVersion -Comments -CompanyName -FileDescription -FileName -FileSize -FileType -FileVersion -FileVersionNumber -InternalName -LegalCopyright -LegalTrademarks -json *.lib

Output:

Code Select Expand


[{
  "SourceFile": "gsdll32.lib",
  "FileName": "gsdll32.lib",
  "FileSize": "7.8 kB",
  "FileType": "Static library"
}]


Also I have noticed that exiftool has another bug with OCX files.  Microsoft Word files (*.DOCX) are captured as well.

Thanks

[Phil Harvey]

Hi Hank,

Skipped the LIB file.

Yes.  This is not currently in the list of supported extensions.  You can force processing of all extensions with -ext "*" if you want.

Also I have noticed that exiftool has another bug with OCX files.  Microsoft Word files (*.DOCX) are captured as well.

I don't understand.  Do you mean that DOCX files are processed when you didn't expect?  What was the command you used?

- Phil

[StarGeek]

Also I have noticed that exiftool has another bug with OCX files.  Microsoft Word files (*.DOCX) are captured as well.

If your command includes the above -if '$FileType=~/DLL|EXE|LIB|OCX|SO|AXF|BIN|ELF|PRX|PUFF|DYLIB|MSI/i' then the OCX part of the regex is going to match DOCX.  Without any boundary limits to your regex, it will match any string with the listed character sequence.  For example, your regex will match "Staypuff Marshmallow Man" because of the PUFF character sequence.

Also, you're making the assumption that the Filetype is the extension.  As you can see in the last output that Lib files return "Static library", not LIB.  This will be a match due to LIB being in LIBrary, but it's not the result you think is happening.

[Hank]

Thank you Phil for explaining the LIB ext. being not supported.

For the DOCX files, yes, DOCX files were scanned when I would not have expected it.
From Windows, but it also occurs in Linux:

Code Select Expand


exiftool -if "$FileType=~/DLL|EXE|LIB|OCX|SO|AXF|BIN|ELF|PRX|PUFF|DYLIB|MSI/i" -json .
[{
  "SourceFile": "./P_U_D.docx",
  "ExifToolVersion": 10.28,
  "FileName": "P_U_D.docx",
  "Directory": ".",
  "FileSize": "169 kB",
  "FileModifyDate": "2017:05:18 06:01:24-04:00",
  "FileAccessDate": "2017:05:22 12:16:26-04:00",
  "FileCreateDate": "2017:05:18 06:01:24-04:00",
  "FilePermissions": "rw-rw-rw-",
  "FileType": "DOCX",
  "FileTypeExtension": "docx",
  "MIMEType": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
  "ZipRequiredVersion": 20,
  "ZipBitFlag": "0x0006",
  "ZipCompression": "Deflated",
  "ZipModifyDate": "1980:01:01 00:00:00",
  "ZipCRC": "0x2578eb05",
  "ZipCompressedSize": 411,
  "ZipUncompressedSize": 1605,
  "ZipFileName": "[Content_Types].xml",
  "Template": "Normal",
  "TotalEditTime": "7.9 days",
  "Pages": 8,
  "Words": 1730,
  "Characters": 9865,
  "Application": "Microsoft Office Word",
  "DocSecurity": "None",
  "Lines": 82,
  "Paragraphs": 23,
  "ScaleCrop": "No",
  "HeadingPairs": ["Title",1],
  "TitlesOfParts": "",
  "Company": "",
  "LinksUpToDate": "No",
  "CharactersWithSpaces": 11572,
  "SharedDoc": "No",
  "HyperlinksChanged": "No",
  "AppVersion": 14.0000,
  "Creator": "admin",
  "LastModifiedBy": "admin",
  "RevisionNumber": 105,
  "CreateDate": "2016:01:05 21:22:00Z",
  "ModifyDate": "2016:01:22 20:54:00Z"
}]
    1 directories scanned
    1 files failed condition
    1 image files read


[Phil Harvey]

OK.  You're looking for a substring in your -if expression.  Try this:

-if "$FileType=~/\b(DLL|EXE|LIB|OCX|SO|AXF|BIN|ELF|PRX|PUFF|DYLIB|MSI)\b/i"

This will look for the complete word instead (word breaks before and after the type).

- Phil

[StarGeek]

OK.  You're looking for a substring in your -if expression.  Try this:

-if "$FileType=~/\b(DLL|EXE|LIB|OCX|SO|AXF|BIN|ELF|PRX|PUFF|DYLIB|MSI)\b/i"

But this will no longer match several other files.  DLL files can return Win32 DLL/Win64 DLL.  EXE files can return Win32 EXE/Win64 EXE.  MSI files aren't ever going to match because they return FPX.

This part of the command should be dropped and replaced with -ext unless there are going to be some files that are of the correct type but don't have proper extensions.  If that is the case, then a list of actually possible Filetypes results needs to be researched.

[Phil Harvey]

But this will no longer match several other files.  DLL files can return Win32 DLL/Win64 DLL.  EXE files can return Win32 EXE/Win64 EXE.

My expression will match these, because DLL and EXE are a separate word within FileType.

MSI files aren't ever going to match because they return FPX.

True.

- Phil

[StarGeek]

My expression will match these, because DLL and EXE are a separate word within FileType.

D'oh again.  Yep, I should have seen that.

[Phil Harvey]

I should have mentioned:

If you want to ExifTool to process LIB files, add this to your .ExifTool_config configuration file:

Code Select Expand

%Image::ExifTool::UserDefined::FileTypes = (
    LIB => 'EXE',
);

- Phil