[prev in list] [next in list] [prev in thread] [next in thread] 

List:       jakarta-commons-dev
Subject:    [jira] [Commented] (FILEUPLOAD-197) ServletFileUpload isMultipartContent method does not support HTT
From:       "Vinicius da Costa Pires (JIRA)" <jira () apache ! org>
Date:       2015-02-26 18:25:05
Message-ID: JIRA.12514116.1310672942000.24983.1424975105493 () Atlassian ! JIRA
[Download RAW message or body]


    [ https://issues.apache.org/jira/browse/FILEUPLOAD-197?page=com.atlassian.jira.plu \
gin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14338857#comment-14338857 \
] 

Vinicius da Costa Pires commented on FILEUPLOAD-197:
----------------------------------------------------

Hi, Simone, Roy and Jochen. I know this is an old discussion, but I just wanted to \
add a thought on this, to see if you agree with me...

I agree with you and Roy that ASF has a bigger role in the scene, and that many \
frameworks like Spring rely on stable architectural code as the Commons File Upload \
project, and to accomplish that stability and cleaness you should stick to the specs.

But you have to agree with me that the thing here is not about accepting PUT, POST, \
PATCH or whatever method someone is using, it's about a concise architectural code \
that does exactly what is expected of it.

If the purpose of the "ServletFileUpload.isMultipartContent" method is to verify if \
the "HttpServletRequest" object has a multipart content type and therefore the \
request is multipart, shouldn't this method only verify that and exit? What is that \
"if" doing there? Is it checking if request has content? If it's preventing a problem \
that we can't see, maybe it could be the same "if", but checking the right thing, \
like "if the request method is the kind of method that has content"...

And as David stated in his comment, neither POST and PUT specs are clear about the \
content type, if it's form-data, multipart, json, xml, or any other content. HTTP was \
designed to not opinionate on content type, am I right?

I understand this requires a certain amount of work for everybody, but all around the \
web in forums and sites like StackOverflow, this method haunts every HTTP1.1 java API \
developer, and the only choices are either not PUT or PATCH files to the server, or \
not using Commons FileUpload at all. Implementing a way to .zip the file just to \
upload on PUT and PATCH requests in javascript for me is out of question.

I personally like ASF's work and trust the stability of its packages, so I don't \
think that implementing my own "isMultipartContent()" is the way to go (although it's \
my only choice now with the tight schedule).

I hope that my comment helps with something. Thanks for your time.

> ServletFileUpload isMultipartContent method does not support HTTP PUT
> ---------------------------------------------------------------------
> 
> Key: FILEUPLOAD-197
> URL: https://issues.apache.org/jira/browse/FILEUPLOAD-197
> Project: Commons FileUpload
> Issue Type: Bug
> Affects Versions: 1.2.2
> Reporter: David Wolverton
> Assignee: Simone Tripodi
> Fix For: 1.3
> 
> 
> This method explicitly checks for method POST. I believe the PUT method can also \
> have multipart requests, and there may be others. In our case we are receiving rest \
> calls using Spring Framework's CommonsMultipartResolver which in turn uses this \
> method of the Commons FileUpload library. Here is the offending code...
> if (!"post".equals(request.getMethod().toLowerCase())) {
> return false;
> }



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


[prev in list] [next in list] [prev in thread] [next in thread]