How to stop strict filename sanitization?

[mcandre]

Calibre insists on renaming my books, dropping non-alphanumeric characters including even a basic plus sign (+). These are being converted to underscores (_) which I find annoying and more confusing than helpful.

How do I customize Calibre's filename sanitization rules so that I can preserve more of the book titles and authors in the filename as regular old characters?

[ownedbycats]

https://www.mobileread.com/forums/sh...d.php?t=119175

the sanitization is for cross-platform compatibility. It cannot be change.

[mcandre]

Which file system can't handle plus sign? Unless you want to explain how Calibre supports MSDOS but not titles with "X++".

[DNSB]

Originally Posted by mcandre

Which file system can't handle plus sign? Unless you want to explain how Calibre supports MSDOS but not titles with "X++".

Have you looked at the character in filename restrictions for all versions of Windows, Mac OSX and Linux with their multiple file systems? Not to mention the additional restrictions on filenames and web services. You might want to check the Characters to Avoid in Directories and Filenames on the MTU website.

What is renamed is the copy of the book stored in calibre's library which you should not be looking at. The calibre GUI, command line tools, etc. will display the book title. So stop peeking under the skirts and treat the calibre library as the black box it is intended to be.

[BetterRed]

Originally Posted by mcandre

Which file system can't handle plus sign? Unless you want to explain how Calibre supports MSDOS but not titles with "X++".

There are other factors at play, some platforms don't support characters such as Ü and è, some platforms have name length restrictions etc - so the approach taken is to translate everything into ASCII, substitute all special characters with underscore ('_') and truncate as required.

Calibre puts format files etc into folders, rather than the database, for performance reasons. The reason it derives the arbitrary folder and file names from the title and first author names is because… the alternative would have been something like database row numbers, or UUIDs.

The core of calibre is its database, not the file system.

BR