You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@jclouds.apache.org by "Maksim Hadalau (Jira)" <ji...@apache.org> on 2024/03/18 18:57:00 UTC
[jira] [Updated] (JCLOUDS-1631) AWS S3, sign for authorization header failed if query part contains special chars
[ https://issues.apache.org/jira/browse/JCLOUDS-1631?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Maksim Hadalau updated JCLOUDS-1631:
------------------------------------
Description:
Problem description:
Can not list blobs for aws-s3 provider if prefix contains special chars %/&
Steps to reproduce:
try to list blobs with following prefix:
"Folder (`~!@#$%^&*-_+[]'|<>.?) Name/"
Actual behavior:
Error: URLDecoder: Incomplete trailing escape (%) pattern
Expected behavior:
provided prefix must be listed
Problem location:
`AWSRequestAuthorizeSignatureV4.signForAuthorizationHeader()`
Multimap<String, String> queryMap = queryParser().apply(request.getEndpoint().getQuery());
request.getEndpoint().getQuery() - returns a decoded query string
however queryParser() require encoded one
Fix:
Multimap<String, String> queryMap = queryParser().apply(request.getEndpoint().getRawQuery());
When jclouds generates a request to the AWS it encodes prefix (encoding all special chars in it, including % and &), however calling `getQuery()` returns decoded version of query string which lead to unpredictable behavior.
P.S. required patch in 2.5.x if possible
PR https://github.com/apache/jclouds/pull/200
was:
Problem description:
Can not list blobs for aws-s3 provider if prefix contains special chars %/&
Steps to reproduce:
try to list blobs with following prefix:
"Folder (`~!@#$%^&*-_+[]'|<>.?) Name/"
Actual behavior:
Error: URLDecoder: Incomplete trailing escape (%) pattern
Expected behavior:
provided prefix must be listed
Problem location:
`AWSRequestAuthorizeSignatureV4.signForAuthorizationHeader()`
Multimap<String, String> queryMap = queryParser().apply(request.getEndpoint().getQuery());
request.getEndpoint().getQuery() - returns a decoded query string
however queryParser() require encoded one
Fix:
Multimap<String, String> queryMap = queryParser().apply(request.getEndpoint().getRawQuery());
When jclouds generates a request to the AWS it encodes prefix (encoding all special chars in it, including % and &), however calling `getQuery()` returns decoded version of query string which lead to unpredictable behavior.
P.S. required patch in 2.5.x if possible
> AWS S3, sign for authorization header failed if query part contains special chars
> ---------------------------------------------------------------------------------
>
> Key: JCLOUDS-1631
> URL: https://issues.apache.org/jira/browse/JCLOUDS-1631
> Project: jclouds
> Issue Type: Bug
> Components: jclouds-blobstore
> Affects Versions: 2.5.0
> Reporter: Maksim Hadalau
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Problem description:
> Can not list blobs for aws-s3 provider if prefix contains special chars %/&
> Steps to reproduce:
> try to list blobs with following prefix:
> "Folder (`~!@#$%^&*-_+[]'|<>.?) Name/"
> Actual behavior:
> Error: URLDecoder: Incomplete trailing escape (%) pattern
> Expected behavior:
> provided prefix must be listed
> Problem location:
> `AWSRequestAuthorizeSignatureV4.signForAuthorizationHeader()`
>
> Multimap<String, String> queryMap = queryParser().apply(request.getEndpoint().getQuery());
> request.getEndpoint().getQuery() - returns a decoded query string
> however queryParser() require encoded one
> Fix:
> Multimap<String, String> queryMap = queryParser().apply(request.getEndpoint().getRawQuery());
> When jclouds generates a request to the AWS it encodes prefix (encoding all special chars in it, including % and &), however calling `getQuery()` returns decoded version of query string which lead to unpredictable behavior.
> P.S. required patch in 2.5.x if possible
> PR https://github.com/apache/jclouds/pull/200
--
This message was sent by Atlassian Jira
(v8.20.10#820010)