Fix issue #669 and add test case.

[willson-chen]

No description provided.

[codecov]

Codecov Report

Merging #721 into master will increase coverage by 0.04%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #721      +/-   ##
==========================================
+ Coverage   96.02%   96.07%   +0.04%     
==========================================
  Files          17       17              
  Lines        2668     2700      +32     
  Branches      310      310              
==========================================
+ Hits         2562     2594      +32     
  Misses         87       87              
  Partials       19       19              

Impacted Files	Coverage Δ
jsonschema/cli.py	96.99% <100.00%> (+0.04%)	⬆️
jsonschema/tests/test_cli.py	99.19% <100.00%> (+0.11%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e48d56c...d106511. Read the comment docs.

[willson-chen]

Add test cases based on PR #680 and fix the base_uri layer problem.

The code before the rectification is as follows:

os.path.abspath(arguments["schema"])) 

Code after the rectification:

os.path.abspath(os.path.dirname(arguments["schema"]))

[Julian]

Thanks for trying to tackle this one as well. Have a look at the comment here from that PR -- specifically, I'd like to support this with a --base-uri flag instead, and if someone wants to use the working directory as that base URI, you'd specify --base-uri . essentially.

[willson-chen]

@Julian Sorry, I ignored some of the details in the comment #680, now I have replaced -r with --base-uri.

[willson-chen]

@Julian I have already pushed the repair part, comments are welcome.

[willson-chen]

We can specify a remote or local path through --base-uri, where the local path can be a relative path or an absolute path. Through scheme = urllib.parse.urlsplit(arguments["base_uri"]).scheme, if the scheme is an empty string or a drive letter identifier from 'a' to 'z' is considered to be a local path, otherwise it is considered a remote path.
@Julian How about this solution?

[Julian]

I'd have to look more carefully myself but that doesn't sound right to me either I think.

I don't really want to change any default behaviors of things (what behavior folks get should match what behavior they get from normal other things).

Or, said differently, I don't want to maintain code that's trying to guess what kind of URI was provided as the base URI.

The base URI as provided needs to be suitable for use exactly as is, or be unambiguous.

Will leave a quick comment on the PR in the section that looks "suspicious" to me, let me know what you think.

[willson-chen]

@Julian Modified to the following simple implementation, let users pass in the correct URI, for example --base-uri file:///c:/ --base-uri http://example, also supports --base-uri .
Is this implementation possible?

    if arguments["base_uri"] is None:
        resolver = None
    elif arguments["base_uri"] == ".":
        file_prefix = "file:///{}/" if "nt" == os.name else "file://{}/"
        resolver = RefResolver(
                    base_uri=file_prefix.format(os.getcwd()),
                    referrer=schema,
                )
    else:
        resolver = RefResolver(
                        base_uri=arguments["base_uri"],
                        referrer=schema,
                    )

[willson-chen]

@Julian I have modified it according to your last review, could you please take a look again?

[willson-chen]

@Julian Sorry, I have been away for a while, but I never forget that there is still something unfinished. Is it possible to modify as follows? 1bfbb90

[Julian]

Hey! No worries (trust me I understand the busyness...)

From a quick look that definitely looks more like what I was expecting! Will have a closer look and leave any comments but think we may be close.

[willson-chen]

@Julian Can this PR be merged? Ask a rather abrupt question, do you welcome new maintainers to take care of the community. I am willing to participate if needed.

[Julian]

Apologies for the delay again here...

There were a few things I wanted to write up comments on which I didn't get a chance to, so I just fixed them in a few commits.

I think this should be good to merge now as soon as CI passes.

On your offer, 100%! I definitely could use more help. If you see pull requests from other folks that aren't yet fully mergeable, or waiting themselves for review, or issues that could use triaging, etc. it would be super super appreciated.

[RomainTT]

A year passed and I did not thank you both yet !

Thank you @willson-chen and @Julian for finishing this work! That is a very useful feature.