Mailing-List: contact issues-help@struts.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@struts.apache.org
Date: Fri, 17 Nov 2017 10:19:00 +0000 (UTC)
From: "M. Manna (JIRA)" <jira@apache.org>
To: issues@struts.apache.org
Message-ID: <JIRA.13119184.1510913623000.265509.1510913940191@Atlassian.JIRA>
In-Reply-To: <JIRA.13119184.1510913623000@Atlassian.JIRA>
References: <JIRA.13119184.1510913623000@Atlassian.JIRA> <JIRA.13119184.1510913623548@jira-lw-us.apache.org>
Subject: [jira] [Updated] (STR-3224) DiskFile getInputStream()
 implementation returns ByteInputStream or FileInputStream
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Fri, 17 Nov 2017 10:19:05 -0000


     [ https://issues.apache.org/jira/browse/STR-3224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

M. Manna updated STR-3224:
--------------------------
    Description: 
i have posted the question original on SO. Here is the link:
[https://stackoverflow.com/questions/47347194/struts-diskfile-getinputstream-returns-bytearrayinputstream-and-filearrayinput?noredirect=1#comment81646937_47347194]

The summary is that this "Switchover" behaviour should be documented somewhere if it returns ByteArrayInputStream or FileInputStream based on file size.

The files for my case was class files (byte code files). and my code is given below


{code:java}
/**
 ** Get Uploaded file from Struts Multipart form upload
 ** org.apache.struts.upload.FormFile
 **/
FormFile file = form.getTheFile();

/**
 ** Since we never had any issue, we didn't explicitly check using
 ** instanceof, but even then it's confusing. We do know that 
 ** ByteArrayInputStream has mark supported. And it works upto
 ** 256KB file size on Windows OS.
 **/
 InputStream markStream = file.getInputStream(); // Doc says it sends "FileInputStream"

 try {
       
// Do something with the markStream object

markStream.reset(); // for ByteArrayInputStream mark is supported
      
// Do something else with markStream

} Catch (IOException ioe) {
       
ioe.printStackTrace();
     
} finally  {
        
markStream.close(); // no effect for certain types
       
}
{code}


Our issue is that this behaviour, even if expected, should be well documented. If we have missed this somewhere in Oracle JDK docs or in ASF docs, please could you point us there?
 

  was:
i have posted the question original on SO. Here is the link:
[https://stackoverflow.com/questions/47347194/struts-diskfile-getinputstream-returns-bytearrayinputstream-and-filearrayinput?noredirect=1#comment81646937_47347194]

The summary is that this "Switchover" behaviour should be documented somewhere if it returns ByteArrayInputStream or FileInputStream based on file size.

The files for my case was class files (byte code files). and my code is given below

/**
 ** Get Uploaded file from Struts Multipart form upload
 ** org.apache.struts.upload.FormFile
 **/
FormFile file = form.getTheFile();

/**
 ** Since we never had any issue, we didn't explicitly check using
 ** instanceof, but even then it's confusing. We do know that 
 ** ByteArrayInputStream has mark supported. And it works upto
 ** 256KB file size on Windows OS.
 **/
 InputStream markStream = file.getInputStream(); // Doc says it sends "FileInputStream"

 try {
       
// Do something with the markStream object

markStream.reset(); // for ByteArrayInputStream mark is supported
      
// Do something else with markStream

} Catch (IOException ioe) {
       
ioe.printStackTrace();
     
} finally  {
        
markStream.close(); // no effect for certain types
       
}

Our issue is that this behaviour, even if expected, should be well documented. If we have missed this somewhere in Oracle JDK docs or in ASF docs, please could you point us there?
 

> DiskFile getInputStream() implementation returns ByteInputStream or FileInputStream
> -----------------------------------------------------------------------------------
>
>                 Key: STR-3224
>                 URL: https://issues.apache.org/jira/browse/STR-3224
>             Project: Struts 1
>          Issue Type: Bug
>    Affects Versions: 1.2.9
>         Environment: Windows OS
>            Reporter: M. Manna
>            Priority: Critical
>
> i have posted the question original on SO. Here is the link:
> [https://stackoverflow.com/questions/47347194/struts-diskfile-getinputstream-returns-bytearrayinputstream-and-filearrayinput?noredirect=1#comment81646937_47347194]
> The summary is that this "Switchover" behaviour should be documented somewhere if it returns ByteArrayInputStream or FileInputStream based on file size.
> The files for my case was class files (byte code files). and my code is given below
> {code:java}
> /**
>  ** Get Uploaded file from Struts Multipart form upload
>  ** org.apache.struts.upload.FormFile
>  **/
> FormFile file = form.getTheFile();
> /**
>  ** Since we never had any issue, we didn't explicitly check using
>  ** instanceof, but even then it's confusing. We do know that 
>  ** ByteArrayInputStream has mark supported. And it works upto
>  ** 256KB file size on Windows OS.
>  **/
>  InputStream markStream = file.getInputStream(); // Doc says it sends "FileInputStream"
>  try {
>        
> // Do something with the markStream object
> markStream.reset(); // for ByteArrayInputStream mark is supported
>       
> // Do something else with markStream
> } Catch (IOException ioe) {
>        
> ioe.printStackTrace();
>      
> } finally  {
>         
> markStream.close(); // no effect for certain types
>        
> }
> {code}
> Our issue is that this behaviour, even if expected, should be well documented. If we have missed this somewhere in Oracle JDK docs or in ASF docs, please could you point us there?
>  


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)