fix: use re.match instead of re.search for mounted app path (#3501)

[0xE111]

Description

See #3501 for detailed description.

TL;DR: When mounting an app, path resolution uses re.search instead or re.match, thus mounted app matches any path which contains mount path.

Original code used re.search and had explicit "sub-path match" (path[match.start() : match.end()]) which makes me think there was some reasoning behind re.search but tbh I don't get it:

if mount_paths_regex and (match := mount_paths_regex.search(path)):
    mount_path = path[match.start() : match.end()]

Closes

Fixes #3501

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.28%. Comparing base (17db66b) to head (1ceb480).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3511   +/-   ##
=======================================
  Coverage   98.28%   98.28%           
=======================================
  Files         328      328           
  Lines       14882    14882           
  Branches     2367     2367           
=======================================
  Hits        14627    14627           
  Misses        116      116           
  Partials      139      139           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[peterschutt]

Original code used re.search and had explicit "sub-path match" (path[match.start() : match.end()]) which makes me think there was some reasoning behind re.search but tbh I don't get it:

I'm trying to figure this out for myself - its almost as though it is designed to match against a sub-path and then extract the mount path component.. but why.. I'm not sure.. I'll spend more time trying to figure it out. On the plus side, your implementation doesn't break any tests - so maybe it is just a weirdness that exists due to code evolution or whatever, but I'd like to take a bit of time to think it through.

[peterschutt]

Original code used re.search and had explicit "sub-path match" (path[match.start() : match.end()]) which makes me think there was some reasoning behind re.search but tbh I don't get it:

if mount_paths_regex and (match := mount_paths_regex.search(path)):
    mount_path = path[match.start() : match.end()]

@provinzkraut My best guess with this is that it was required before we started handling root_path in #3039. That is, before that PR, the path we matched the mount path against, may still have had the root path prepended to it, and so indexing both ends of the slice would have effectively stripped the root path from the path we pass through to the mounted app.. this is my best guess.. any thoughts?

[peterschutt]

Thanks @0xE111 - I'm just waiting on another set of eyes on that path slicing business before we merge.

[github-actions]

Documentation preview will be available shortly at https://litestar-org.github.io/litestar-docs-preview/3511

[provinzkraut]

@provinzkraut My best guess with this is that it was required before we started handling root_path in #3039. That is, before that PR, the path we matched the mount path against, may still have had the root path prepended to it, and so indexing both ends of the slice would have effectively stripped the root path from the path we pass through to the mounted app.. this is my best guess.. any thoughts?

Can't come up with a better explanation, so I'd say it's okay to merge this as is!

[provinzkraut]

@all-contributors add @0xE111 for code

[allcontributors]

@provinzkraut

I've put up a pull request to add @0xE111! 🎉