buildr-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Assaf Arkin (JIRA)" <j...@apache.org>
Subject [jira] Commented: (BUILDR-67) download(art => 'http://example.com/foo?query') fails to send query part of URI when performing the HTTP GET
Date Wed, 07 May 2008 15:52:55 GMT

    [ https://issues.apache.org/jira/browse/BUILDR-67?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12594948#action_12594948
] 

Assaf Arkin commented on BUILDR-67:
-----------------------------------

That was the right thing to do.  On my side, I applied the first patch (git am), ran the tests
(failed), applied the second patch, ran again (pass).  At that point I had the patch committed
locally, basically the same commit log that you have on your machine (including the same commit
message).

The commit messages didn't transfer to SVN, I did something wrong, still trying to figure
it out.  But that's the correct way to send patches.

You can't git push directly because we still need to apply those patches manually, but (when
I get the second part figured out), it would have had the same effect.


> download(art => 'http://example.com/foo?query') fails to send query part of URI when
performing the HTTP GET
> ------------------------------------------------------------------------------------------------------------
>
>                 Key: BUILDR-67
>                 URL: https://issues.apache.org/jira/browse/BUILDR-67
>             Project: Buildr
>          Issue Type: Bug
>          Components: Core features
>    Affects Versions: 1.2.10
>            Reporter: Tommy Knowlton
>         Attachments: 0001-Add-a-failing-spec-test-demonstrating-issue-BUILDR.patch, 0002-fix-for-BUILDR-67.patch
>
>   Original Estimate: 0h
>  Remaining Estimate: 0h
>
> This appears to be caused by inappropriate use of the URI's path member at the point
where the GET is performed.
> I have already tested a fix that works against 1.2.10 (released) code.
> Also, looking at the trunk code, it appears that this bug remains, although the code
affected in 1.2.10 has since been rewritten to use Net::HTTP (i.e., the rewrite still uses
uri.path when uri.request_uri seems to be more correct).
> I *have not* yet created a unit test to fail.
> I'm not clear in my mind about the 'priority' on this, but I'm leaving it at 'Major'
for now, because it breaks my organization's builds. I don't particularly care if the project
management team sees fit to move the priority down on this.
> I expect to submit one or more patches Real Soon Now (TM).

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message