Resource could not be found error

[olegkon]

Hi,

Trying to run the sample with sparky.sendMessage()
and getting the following error:

{ "errors": [ { "message": "Resource could not be found" } ] }

Looks like the code is adding an extra slash (before transmissions) into the url calling
https://api.sparkpost.com/api/v1//transmissions?num_rcpt_errors=0

We are using apache httpclient-4.5.3.jar (on Win 10).

Can anyone please fix it?

TIA,
Oleg.

[yepher]

@olegkon I just ran this sample after setting my API key, proper from address and a proper too address and it worked fine: sendMessage()

Can you provide a little more details for how you are using the library. Version and if you are using Maven will be helpful too.

To make sure I set a break point here RestConnection.java#L112 and I do not see the extra / that you are seeing.

Is it possible you set a trailing slash for SPARKPOST_BASE_URL=https://api.sparkpost.com/api/v1 in config.properties or have that set as an environment variable with the trailing slash?

[yepher]

@olegkon, I've spent time looking through the library code for how we might get an extra / and I am not finding any suspicious code yet. In fact the code guards against that case.

Early in the libraries history there was a case where there was a duplicate / but that was fixed long ago. I did a quick search but did not find the exact commit where that was fixed. But here is the code that is meant to address it: RestConnection.java#L90

public RestConnection(Client client, String endpoint) throws SparkPostException {
        this.client = client;
        if (StringUtils.isAnyEmpty(endpoint)) {
            this.endpoint = defaultApiEndpoint;
        } else {
            if (endpoint.endsWith("/")) {   // Here we protect agains a trailing slash if not from SPARKPOST_BASE_URL property
                this.endpoint = endpoint;
            } else {
                this.endpoint = endpoint + '/';
            }
        }
    }

My best guess is there is a / on the end of this variable SPARKPOST_BASE_URL that your environment has set by one or more of the following:

* JVM argument
* Env variable
* `config.properties` file.

OR

You are running a very old version of the library.

I am going to leave this issue open until next week to give you some time to provide additional feedback. I would be glad to help you get it working and or fix the code if there is indeed a bug in the library. If you are on "http://slack.sparkpost.com/" mention me there (@chris.wilson-sp) and I will jump on and try and help you real-time.

Starting countdown to close issue as not an issue if no addition details are provided...

[dangerusslee]

@yepher If that code is supposed to do what it says it in the comment, then it's flipped. That code is making sure that there is a slash at the end of the endpoint.

So when there's a leading slash in the path as well, it combines to form //

I believe createConnectionObject should have a guard to make sure there's no // or you need to make sure that the path doesn't have a leading slash.

I believe a difference in the version of httpclient may be causing you to not see the issue.

[yepher]

Sorry. I mis-typed the comment when responding to the issue. I meant to say it only adds the / if it not already there.

I will test with that httpclient version and report back shortly.

[olegkon]

Guys, it adds another part to it (path?) with transmissions, which has another '/' (/transmissions?num_rcpt_errors=0) Sorry, don't have debugger now, can't recall exactly, saw it yesterday.

I don't have any config file. Should I use it? Please give more details, I am totally new to SparkPost.

[dangerusslee]

@yepher It's also possible that you have a config file and it gets the defaultApiEndpoint from there which doesn't have a trailing slash.

[yepher]

hmm, EndPoint unit tests fail with httpclient-4.5.3.

Will do more investigation:

Running com.sparkpost.EndPointTest
Tests run: 4, Failures: 4, Errors: 0, Skipped: 0, Time elapsed: 5.51 sec <<< FAILURE!
testEndPointWithMultipleParameters(com.sparkpost.EndPointTest)  Time elapsed: 0.028 sec  <<< FAILURE!
org.junit.ComparisonFailure: expected:<[]transmissions?num_rc...> but was:<[/]transmissions?num_rc...>
	at com.sparkpost.EndPointTest.testEndPointWithMultipleParameters(EndPointTest.java:84)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at org.apache.maven.surefire.junit4.JUnit4Provider.execute(JUnit4Provider.java:252)
	at org.apache.maven.surefire.junit4.JUnit4Provider.executeTestSet(JUnit4Provider.java:141)
	at org.apache.maven.surefire.junit4.JUnit4Provider.invoke(JUnit4Provider.java:112)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at org.apache.maven.surefire.util.ReflectionUtils.invokeMethodWithArray(ReflectionUtils.java:189)
	at org.apache.maven.surefire.booter.ProviderFactory$ProviderProxy.invoke(ProviderFactory.java:165)
	at org.apache.maven.surefire.booter.ProviderFactory.invokeProvider(ProviderFactory.java:85)
	at org.apache.maven.surefire.booter.ForkedBooter.runSuitesInProcess(ForkedBooter.java:115)
	at org.apache.maven.surefire.booter.ForkedBooter.main(ForkedBooter.java:75)

[yepher]

I am working a new release. I will fix this and get it into the next release. Thank you @dangerusslee and @olegkon.

I am going to make sure it works with httpclient-4.5.3 and httpclient-4.5 but I will also bump the pom dependency up the latest version of HTTP Client.

[yepher]

Back testing HTTP CLIENT 4.5.x

• 4.5 - OK
• 4.5.1 - OK
• 4.5.2 - OK
• 4.5.3 - Broken

[yepher]

This is the HTTP Client change that break backward behavior:
apache/httpcomponents-client@0554271

[yepher]

I was getting ready to send an email to olegk at apache.org to verify the new HTTP client URIBuilder is performing the expected behavior when I noticed the similarity in names:

---

ChangeLog:

• [HTTPCLIENT-1803] Improved handling of malformed paths by URIBuilder.
  Contributed by Oleg Kalnichevski

---

@olegkon are you the same Apache olegk that made URIBuilder change or is that coincidence?

[olegkon]

No, I am olegkon@yahoo.com Oleg is a popular name ;-)

[yepher]

From Oleg at Apache:

This change was made per this bug report [1]. You have already found
out that. There was a follow up change request post 4.5.3 that adjusted
URIBuilder behavior slightly [2]. 

Going forward (4.5.4) URIBuilder will make sure that a path if non-null
will always starts with /. Null path will not be altered or converted
to /.

Hope this helps

Oleg

[1] https://issues.apache.org/jira/browse/HTTPCLIENT-1803
[2] https://issues.apache.org/jira/browse/HTTPCLIENT-1810

So I will add code to handle <4.5.3 and >=4.5.3 HTTP Client. Eventually will remove support for HTTPClient 4.5.3.