Re: Case Insensitive File Systems -- Torvalds Hates Them

On Wed, 30 Apr 2025 10:29:17 -0000 (UTC), Borax Man wrote:

On 2025-04-29, rbowman <bowman@montana.com> wrote:

On Mon, 28 Apr 2025 15:46:56 -0400, CrudeSausage wrote:
>

The biggest issue I envision anyone having with case sensitivity is
not remembering what they saved a certain file as. However, if you
know that the document deals with green eggs, searching for it despite
not remembering the document filename should be trivial.

>
I was talking about databases. In DB2
>
'SELECT author FROM books WHERE title LIKE 'green eggs%'
>
is only going to match 'green eggs' exactly. If you're not sure how the
record was stored you would need LOWER(title).
>
The equivalent directory search would be something like
>
find . -name "*.txt" | xargs grep -i "green eggs"
>
The difference is the directory search is probably going to be a
one-shot.
The SQL statement may be executed thousands of times and LOWER() is
costly.
>
I'm curious how SQL Server or Access handles case insensitivity in the
internals. However it's done it's faster than explicit conversions at
runtime.

 
Why not just use the -iname option for find?  Its the same as -name, but
not case sensitive.

In the example I'm not worried about finding .TXT, etc. A more realistic
example would be  'find . -name "*.cpp"'  at least on my machines where I
would never have 'foo.CPP',  'foo.PY', and so forth.

After find locates foo.cpp, bar.cpp, and baz.cpp, xargs passes the file
name to grep.  More likely the search string would be something like
"*landmark*".