V2 try_file redir question

[John_Siu]

1. Caddy version (caddy version):

/srv # caddy version
v2.1.1 h1:X9k1+ehZPYYrSqBvf/ocUgdLSRIuiNiMo7CvyGUQKeA=

2. How I run Caddy:

Docker

a. System environment:

Docker on Ubuntu 20.04

b. Command:

n/a

c. Service/unit/compose file:

n/a

d. My complete Caddyfile or JSON config:

johnsiu.com {
	encode gzip
	file_server
	import my_redir_files /blog/{path}
	root * /www/site/johnsiu.com
	@try_files {
		file {
			try_files /blog/{path}index.html /blog/{path}/index.html
		}
	}
	redir @try_files {http.matchers.file.relative}
}

3. The problem I’m having:

The above work according to syntax:

https://johnsiu.com/caddyfile is redirected to https://johnsiu.com/blog/caddyfile/index.html

What I want to do is:

https://johnsiu.com/caddyfile is redirected to https://johnsiu.com/blog/caddyfile/

Moreover, it will redirect even the file exist:

https://johnsiu.com/myip (the filename is myip) actually exist at root. But right now it is redirected to https://johnsiu.com/blog/myip/index.html

4. Error messages and/or full log output:

No error.

5. What I already tried:

I already tried following code:

	@try_files {
		file {
			try_files /blog/{path} /blog/{path}/
		}
	}
	redir @try_files {http.matchers.file.relative}

But no redirect happen and give 404 in browser. It feel like file.try_files cannot match directory.

6. Links to relevant resources:

[francislavoie]

I think the best thing to do is combine it with a path_regexp matcher that you can craft to match just the part you need, then redirect using the capture groups.

Maybe something like this:

@blog {
	path_regexp blog ^/([^/]+)/?$
	file /blog/{path}index.html /blog/{path}/index.html
}
redir @blog /blog/{re.blog.1}/

Note that the file matcher allows for a shortcut by putting the try_files arguments on the same line if you don’t need to configure any of its other arguments.

Basically the regexp will match a single path segment, with an optional trailing /. I couldn’t figure out on short notice how to use a non-greedy regexp for this, so if you need it to match more than one path segment, you might need to do some figuring out for the ideal regexp for you.

[John_Siu]

Just tested your code, that works perfectly.

Is there a way to make it only execute if the request path doesn’t exist?

For example, if /caddyfile(file) or /caddyfile/(directory) exist, then the redir code should not execute.

[francislavoie]

Yep, you can add not file {path} as a third matcher in there, so if the path exists as-is, it won’t match. At least I think that’s what you’re asking for.

[John_Siu]

Just tried as follow:

@blog {
	path_regexp blog ^/([^/]+)/?$
	file /blog/{path}index.html /blog/{path}/index.html
	not file {path}
}

But /myip is still redirecting. Or did I misunderstand.

[francislavoie]

So your root is /www/site/johnsiu.com.

Say you get a request with path /foo.

The not file {path} matcher will look for a file /www/site/johnsiu.com/foo. If that exists, then the @blog matcher should not match.

Does that make sense? Maybe I’m misunderstanding your question.

[John_Siu]

Your understanding is correct.

Both of following exist:
/www/site/johnsiu.com/myip
/www/site/johnsiu.com/blog/myip/index.html

Redirect is happening for https://johnsiu.com/myip even with not file {path}.

[John_Siu]

Will split the later question into a separate post once I setup a clean testing site.

[John_Siu]

For the 2nd part: make it only execute if the request path doesn’t exist?

@francislavoie solution works.

I just have to remove /index.html from my original file matcher. So following works:

@blog {
	path_regexp blog ^/([^/]+)/?$
	file /blog/{path} /blog/{path}/
	not file {path} {path}/
}