Hacker News Reader

Comment by richfnelson | Hacker News Reader

In DOS you use `dir` to list files in a directory. You can also use `dir` to list files that match a string in a directory. `dir 1.png` would list all PNGs with a "1" in their name. You need a separate command with options and a flag to do that in bash. In DOS, it's less than a dozen characters, and it's far more intuitive than having to use `find`. Sure, `find` and `ls` might be more flexible, but definitely not more intuitive.

View on HN

You need a separate command with options and a flag to do that in bash.

No you don't; whether ls is a builtin is mostly irrelevant, but

    ls *1*.png

does exactly what you (I assume) are saying,

    dir *1*.png

which AFAIK is a new extension introduced with WinNT's command processor, since the old algorithm...

https://blogs.msdn.microsoft.com/oldnewthing/20071217-00/?p=...

...would effectively produce a pattern matching all files ending in '.png'.

using ls with globbing is redundant: the shell will expand * 1* .png to a list of files, which will then be passed to the argument vector of ls, looking up the files twice. You could just do echo * 1* .png.

ls doesn't do glob matching. ls '*' would try to fstatat (FreeBSD) that character.

EDIT: formatting

This use of the word "new" is a bit of a stretch. Consider that the "new" wildcard algorithm has been around for as long as Linux has. Possibly longer, although I'd have to check when 4DOS and 4OS2 started providing extended wildcards.

Does DOS have anything like "locate x"? Or do you have to use the horribly inefficient graphical search which takes AGES. The Unix terminal takes a little while to learn, but it is powerful. Windows is mind-numbing.

Of course you can: http://stackoverflow.com/questions/8677628/recursive-file-se...

I've used that before and the functionality isn't even close. The UNIX command is fast, while the recursive PS command takes forever and prints tons of garbage to your console when it tries to read from certain directories. It also isn't near as intuitive as locate and takes longer to write-out. You're right that it is better than the graphical way though!

locate is fast because there's a cron job that does the heavy lifting of scanning and indexing the filesystem. Without that, find is about as fast as PS.

Yes and it puts it in a database updated regularly for fast access. I'm not sure why this isn't default on Windows. How often does one need to search for a random file in the OS?

That's for PowerShell, not DOS. IIRC, for DOS you could do something like "dir /s *.txt" to list all files with .txt extension in all subdirectories of the current directory, so "dir /s x" would be somewhat equivalent to Unix's "find -name x" (not as fast as "locate x", but should return the same if run on the root directory, and if there were no changes since the locate database was last rebuilt).

Teaching an old dog new tricks.

It's very common to think of the first thing you learnt as more intuitive. Learning the second thing you have baggage and expectations of how something is supposed to work.

Reply on news.ycombinator.com