[Bug 62971] New: Revert the fix for 49464 in tomcat 9

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

[Bug 62971] New: Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

            Bug ID: 62971
           Summary: Revert the fix for 49464 in tomcat 9
           Product: Tomcat 9
           Version: 9.0.13
          Hardware: PC
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Catalina
          Assignee: [hidden email]
          Reporter: [hidden email]
  Target Milestone: -----

I encountered weird issue and I found this
http://tomcat.10.x6.nabble.com/Tomcat-8-5-19-corrupts-static-text-files-encoded-with-UTF-8-td5065877.html
Tomcat introduce such regress in 8.5.19 and reverted it soon in 8.5.20, but
this still exists in tomcat 9.0, and not documented in migration guide
https://tomcat.apache.org/migration-9.html
I think it's bad to assume file encoding as Charset.defaultCharset(),
especially on windows, normally system is not utf-8 encoding, and files like js
are utf-8, if non-ascii characters included, it parse failed by browser.
The change resolved some edge cases but bring a big regression issue, if you
insist on it, please update the migration guide.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

Remy Maucherat <[hidden email]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 OS|                            |All
             Status|NEW                         |RESOLVED
         Resolution|---                         |DUPLICATE

--- Comment #1 from Remy Maucherat <[hidden email]> ---


*** This bug has been marked as a duplicate of bug 49464 ***

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

[hidden email] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|DUPLICATE                   |---
             Status|RESOLVED                    |REOPENED

--- Comment #2 from [hidden email] ---
The fix to bug 49464 potentially missed a scenario in its consideration, and
may have introduced a "new" bug. However, we have not been able to create an
isolated test scenario to demonstrate such yet.

In comments to bug 49464, it listed 3 scenarios considered: "
There are three scenarios to consider:
a) directly returning a file
b) including a file into an output stream
c) including a file into a writer".

However, a fourth scenario <c:import...> for standard tag library core appears
to be impacted.

To my understanding, if a jsp contains a:
<c:import var="xslt" url="/sample.xsl"/>
it is supposed to populate the content from that file "sample.xsl" into the
"variable" "xslt", for later code to use.

The issue behavior that we (in certain conditions) observed, is that the
"variable" "xslt" would become empty, yet the jsp or the result html would
receive the content of "sample.xsl" as part of its own content, making it
behaving as if it is a "include" tag instead of an import tag.

We have not identified the exact conditions that would trigger such behavior
yet - thus no isolated test case created to demonstrate such yet.  But since
the issue behavior looks to match the comments of the bug/fix for 49464, I am
adding this comment to start the communication. - We will update when/if we
have more specific info.

We are using tomcat 9.0.26

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

Mark Thomas <[hidden email]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |NEEDINFO

--- Comment #3 from Mark Thomas <[hidden email]> ---
This sounds more like a concurrency issue with the tag library.

Waiting for more information on how to reproduce this...

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

--- Comment #4 from [hidden email] ---
I am able to have an isolated case to reproduce the "bug" behavior now.
Still need to clean up a bit more before I submit.  But a quick update:
previously in the scenario I couldn't reproduce the "bug", appears to be
because locally a newer version of taglib was in effect. Removing the
taglibs-standard-impl-1.2.5.jar and taglibs-standard-spec-1.2.5.jar, and let
the javax.servlet.jsp.jstl-1.2.3.jar and javax.servlet.jsp.jstl-api-1.2.1.jar
to take effect, would cause the behavior to appear.  - Well, I guess it could
be said as a javax.servlet.jsp.jstl-1.2.3.jar bug...
-A quick update anyway.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

--- Comment #5 from [hidden email] ---
Also it appears to require struts(1) action/tiles to get it happen - a simple
jsp with c:import does not show the issue behavior.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

--- Comment #6 from [hidden email] ---
It turns out the issue behavior is not related to struts/tiles/action. I am
able to develop a simple jsp that encounter the issue - still want to trim down
further before I submit the pacakge.  But it does require to use tag lib
http://java.sun.com/jsp/jstl/fmt, with javax.servlet.jsp.jstl-1.2.3.jar.

In debug, it looks the fmt:setLocale would cause an CharacterEncoding (for
en-US) to be set on the Response, which in turn, at the code for bug 49464,
would cause:
outputEncoding != org.apache.coyote.Constants.DEFAULT_BODY_CHARSET.name()
to be true, which triggers the outputEncodingSpecified pw.flush to happen.

and with the ImportSupport from javax.servlet.jsp.jstl-1.2.3.jar looks not able
to deal with the flushed irw, while the taglibs-standard-impl-1.2.5 appears to
be able to handle it - these I have not debugged into.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

--- Comment #7 from [hidden email] ---
Note that in debug that encountered the problem (in my particular case anyway),
in the DefaultServlet.java, even when:
outputEncoding != org.apache.coyote.Constants.DEFAULT_BODY_CHARSET.name()
becomes true, the contents of both sides are actually the same.  -This and the
flush looks to be the logic for the 49464 fix.

Which makes this fix suspicious.  - Locally changing the "!=" to ".equals(..)"
would make the issue go away.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

[hidden email] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEEDINFO                    |NEW

--- Comment #8 from [hidden email] ---
A simple jsp below could demonstrate the issue behavior - when it is used with
javax.servlet.jsp.jstl-1.2.3.jar (and javax.servlet.jsp.jstl-api-1.2.1.jar)
instead of taglibs-standard-impl-1.2.5.jar and taglibs-standard-spec-1.2.5.jar.
 I have also created a simple webapp with such. - Some environment factors
might play - I am using a Windows 10 pro, with en-US/Eastern Time zone, which
might cause the non-constant CharacterEncoding to be set on the Response.

But now I am wondering maybe I should create a separate bug instead of reopen
this one - so that we could move forward instead of considering rolling it back
- what do you think?

<!--begin of jsp>
<%@ taglib uri="http://java.sun.com/jsp/jstl/core" prefix="c" %>
<%@ taglib uri="http://java.sun.com/jsp/jstl/fmt" prefix="fmt" %>

<fmt:setLocale value='' scope="session"/>

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<title>ShowBug</title>
</head>
<body>
    <div id="xsltnot"> c:import is not expected to get the contents in between
-<br>
        ---------------------------------------------<br>
        <c:import var="xslt" url="/sample.txt"/><br>
        ---------------------------------------------<br>
    </div>
    <br>
    <br>
    <div id="xslt"> Instead, we expect the c:import to get the contents into
the variable, which should be displayed in between *<br>
    ********************************************<br>
        <c:out value="${xslt}" escapeXml="true"/><br>
    ********************************************<br>
    </div>
</body>
</html>
<!-- end of jsp>

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

--- Comment #9 from Mark Thomas <[hidden email]> ---
This situation you describe seems to be wholly unrelated to this bug.

I recommend opening a new bug.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[Bug 62971] Revert the fix for 49464 in tomcat 9

Bugzilla from bugzilla@apache.org
In reply to this post by Bugzilla from bugzilla@apache.org
https://bz.apache.org/bugzilla/show_bug.cgi?id=62971

[hidden email] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #10 from [hidden email] ---
Thanks for the recommendation. New Bug 64081 reported. Closing this one.

--
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]