You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Niklas Quarfot Nielsen (JIRA)" <ji...@apache.org> on 2015/02/03 23:56:35 UTC

[jira] [Commented] (MESOS-1794) allow CURL download into a pipe

    [ https://issues.apache.org/jira/browse/MESOS-1794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14304231#comment-14304231 ] 

Niklas Quarfot Nielsen commented on MESOS-1794:
-----------------------------------------------

[~kdomanski] [~benjaminhindman] Are you guys working on this, or should we bump again to 0.23.0?

> allow CURL download into a pipe
> -------------------------------
>
>                 Key: MESOS-1794
>                 URL: https://issues.apache.org/jira/browse/MESOS-1794
>             Project: Mesos
>          Issue Type: Improvement
>          Components: stout
>    Affects Versions: 0.20.0
>            Reporter: Kamil Domański
>            Assignee: Kamil Domański
>              Labels: features, patch, performance
>
> This divides {{Try<int> net::download(const std::string& url, const std::string& path)}} into three overloads:
> # {{Try<int> net::download(const std::string& url, FILE* file)}} downloads a file and sends it to FILE stream
> # {{Try<int> net::download(const std::string& url, int fdout)}} opens a file stream based on a file descriptor and calls 1)
> # {{Try<int> net::download(const std::string& url, const std::string& path)}} opens a file at _path_ and passes the descriptor to 2), effectively working as before this change
> This will allow to download into any file descriptor, such as a pipe, e.g. directly to extraction, as proposed by [~bernd-mesos] in MESOS-1667.
> Under similar rationale it's necessary to implement {{hadoop fs -cat}} into a file descriptor, but I'm unable to deploy HDFS for testing.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)