so now.. is there any example of how you can handle url redirects on ktor-client manually? i am deal...

# ktor

[Nikky]

so now.. is there any example of how you can handle url redirects on ktor-client manually?
i am dealing with a api returning me incorrectly escaped urls, and with fuel i just disabled redirects and looped until it finally resulted in a file

[Deactivated User]

your best bet is to check the HttpRedirect feature:
https://github.com/ktorio/ktor/blob/d69d00b7971da8501c33efe296c93cf138ec0028/ktor-client/ktor-client-core/src/io/ktor/client/features/HttpRedirect.kt

[Deactivated User]

have you reported the issue you have with escaping, so we can fix it?

[Nikky]

i can do so.. my fix is https://gist.github.com/NikkyAI/d2820695a112a2b3a681b6333c2189e6

now i am stuck on some jar from jenkins not working due to content-type, seems like the header is lowercase ?

[Nikky]

i have to go soon, i hope i will remember to do a proper report on github

[Deactivated User]

okay, thanks!

[Nikky]

now how to handle files like this: https://media.forgecdn.net/files/2522/248/Controlling-3.0.6.jar

it has a empty header value for content-encoding
i hate people who do this to me

[Deactivated User]

< HTTP/1.1 200 OK
< Content-Type: application/octet-stream
< Content-Length: 26034
< Connection: keep-alive
< Date: Tue, 23 Jan 2018 01:34:55 GMT
< Last-Modified: Sat, 20 Jan 2018 18:12:27 GMT
< ETag: "0d9a0429c72bc1c06358f1166bdd8c59"
< Content-Encoding:
< Accept-Ranges: bytes
< Server: AmazonS3
< Age: 2954
< X-Cache: Hit from cloudfront
< Via: 1.1 ...(CloudFront)
< X-Amz-Cf-Id: ...

Yep, loops like that. What do you mean by how to handle?

[Deactivated User]

Do you have an exception or something when downloading?

[Nikky]

did not fit in here so a gist: https://gist.github.com/NikkyAI/442b4ef3025158ba8aa2694a0d7349d0

[Nikky]

well curl and chrome can download it just fine.. so how can i instruct HttpClient to ignore those "excpected" errors ?

[Deactivated User]

let me check if I can find a workaround

[Nikky]

cool, i will be back later.. i am running late

[Deactivated User]

okay

[Deactivated User]

@e5l it seems that if (range.start == range.end) throw ParserException("No HTTP header value provided for name ${builder.substring(nameStart, nameEnd)}: \n$builder") (or at least should be a way to disable that verification) since the request is still valid, chrome, curl and wget can download it