RE: RequestUtils.computeURL()
Only that unnecessary calls to the URLEncoder.encode within the computeURL method adds excess overhead that is not always needed. If encoding a url could be optionally turned on / off, performance would be better. The URLEncoder instantiates a ByteArrayOutputStream and a BufferedWriter which can add significant process cycles when called repetitively. Also, any routines that allocate memory buffers like the URLEncoder or MessageFormat should be questioned for performance. They add significant overhead over time on highly repetitive calls. Also, if a machine begins to run low on memory, the constant allocation and release of these buffers only makes the situation worse. -Original Message- From: James Mitchell [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 10:01 AM To: Struts Developers List Subject: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? It seems that ImgTag does it all by itself, and since it hard-codes '' instead of the standard 'amp;', my new tests (which I haven't even committed yet) are failing. Should we refactor this now? I cannot setup tests for this tag with the current code that's there. Thoughts? -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: RequestUtils.computeURL()
Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? computeURL also encodes the url with session information which isn't desirable for images. Maybe ImgTag should just replace with amp; itself without calling computeURL. David _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
That was my initial thought, but wanted to bounce it off a few other people first. -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) -Original Message- From: David Graham [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 10:28 AM To: [EMAIL PROTECTED] Subject: Re: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? computeURL also encodes the url with session information which isn't desirable for images. Maybe ImgTag should just replace with amp; itself without calling computeURL. David _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
Except that the doc for html:img says: Renders an HTML img element with the image at the specified URL. Like the link tag, URL rewriting will be applied automatically to the value specified in src or page, to maintain session state in the absence of cookies. This will allow dynamic generation of an image where the content displayed for this image will be taken from the attributes of this tag. So ImgTag already does URL rewriting, doesn't it? Quoting James Mitchell [EMAIL PROTECTED]: That was my initial thought, but wanted to bounce it off a few other people first. -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) -Original Message- From: David Graham [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 10:28 AM To: [EMAIL PROTECTED] Subject: Re: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? computeURL also encodes the url with session information which isn't desirable for images. Maybe ImgTag should just replace with amp; itself without calling computeURL. David _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Kris Schneider mailto:[EMAIL PROTECTED] D.O.Tech http://www.dotech.com/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
Ok, so make the image tag call RequestUtils.computeURL() to escape the and encode session data if needed. David Except that the doc for html:img says: Renders an HTML img element with the image at the specified URL. Like the link tag, URL rewriting will be applied automatically to the value specified in src or page, to maintain session state in the absence of cookies. This will allow dynamic generation of an image where the content displayed for this image will be taken from the attributes of this tag. So ImgTag already does URL rewriting, doesn't it? Quoting James Mitchell [EMAIL PROTECTED]: That was my initial thought, but wanted to bounce it off a few other people first. -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) -Original Message- From: David Graham [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 10:28 AM To: [EMAIL PROTECTED] Subject: Re: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? computeURL also encodes the url with session information which isn't desirable for images. Maybe ImgTag should just replace with amp; itself without calling computeURL. David _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Kris Schneider mailto:[EMAIL PROTECTED] D.O.Tech http://www.dotech.com/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] _ MSN 8 helps eliminate e-mail viruses. Get 2 months FREE*. http://join.msn.com/?page=features/virus - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
Yes, well, my concern is that we have duplicate code doing the same thing. I assume computURL was added after the code for ImgTag was written, and then perhaps overlooked. I changed the '' to amp; in the ImgTag and my tests are now passing. That sounds fine until you actually look at the test. I use computeURL in my tests to verify that Tags are calculating URLs correctly when using: Map defined by name in any scope Map defined by name and property in any scope Map defined by name in application scope Map defined by name and property in application scope Map defined by name in session scope Map defined by name and property in session scope Map defined by name in request scope Map defined by name and property in request scope I don't believe refactoring it would help speed-wise, but I just hate seeing duplicate code out there. -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) -Original Message- From: Kris Schneider [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 11:13 AM To: Struts Developers List Subject: RE: RequestUtils.computeURL() Except that the doc for html:img says: Renders an HTML img element with the image at the specified URL. Like the link tag, URL rewriting will be applied automatically to the value specified in src or page, to maintain session state in the absence of cookies. This will allow dynamic generation of an image where the content displayed for this image will be taken from the attributes of this tag. So ImgTag already does URL rewriting, doesn't it? Quoting James Mitchell [EMAIL PROTECTED]: That was my initial thought, but wanted to bounce it off a few other people first. -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) -Original Message- From: David Graham [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 10:28 AM To: [EMAIL PROTECTED] Subject: Re: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? computeURL also encodes the url with session information which isn't desirable for images. Maybe ImgTag should just replace with amp; itself without calling computeURL. David _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Kris Schneider mailto:[EMAIL PROTECTED] D.O.Tech http://www.dotech.com/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: RequestUtils.computeURL()
On Mon, 10 Mar 2003, James Mitchell wrote: Date: Mon, 10 Mar 2003 10:01:16 -0500 From: James Mitchell [EMAIL PROTECTED] Reply-To: Struts Developers List [EMAIL PROTECTED] To: Struts Developers List [EMAIL PROTECTED] Subject: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? It seems that ImgTag does it all by itself, and since it hard-codes '' instead of the standard 'amp;', my new tests (which I haven't even committed yet) are failing. Should we refactor this now? I cannot setup tests for this tag with the current code that's there. Sounds like we missed refactoring this when computeURL() was created in the first place, so +1 for refactoring it now. As of last night, I got all the test.tomcat.41 tests to run on my Linux box (yah :-) -- thanks for all the hard work on creating these tests. I ran into line-end problems on the code we generate for html:form and html:errors, so I tweaked them a bit. I hope that didn't mess up running the tests on Windows, and would be interested to find out if it did. Thoughts? Craig -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: RequestUtils.computeURL()
On Mon, 10 Mar 2003, David Graham wrote: Date: Mon, 10 Mar 2003 08:27:32 -0700 From: David Graham [EMAIL PROTECTED] Reply-To: Struts Developers List [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? computeURL also encodes the url with session information which isn't desirable for images. Maybe ImgTag should just replace with amp; itself without calling computeURL. It is definitely desireable if your image is dynamically calculated based on information stored in the session. It is also desireable if you are using session affinity to track all of a particular user's requests throughout their interaction with the app. Having a boolean switch on the URL rewriting would be a nice 1.2 enhancement, but I think we should definitely call computeURL() now. David Craig - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
Sounds great. I'll take care of that tonight. Can anyone run test.tomcat.all? I guess I purchased some bad memory for my laptop or the OS/JVM is a piece of sh#t, because when I try to run them all, I get somewhere in the .40 tests and my laptop just shuts downno blue screenno OutOfMemory errorsit just stopslike pulling the plug on a desktop.very frustrating. -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) -Original Message- From: Craig R. McClanahan [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 11:35 AM To: Struts Developers List Subject: Re: RequestUtils.computeURL() On Mon, 10 Mar 2003, James Mitchell wrote: Date: Mon, 10 Mar 2003 10:01:16 -0500 From: James Mitchell [EMAIL PROTECTED] Reply-To: Struts Developers List [EMAIL PROTECTED] To: Struts Developers List [EMAIL PROTECTED] Subject: RequestUtils.computeURL() Is there any reason that any tag should NOT call computeURL when evaluating a URL that needs a parameter of Map entries added? It seems that ImgTag does it all by itself, and since it hard-codes '' instead of the standard 'amp;', my new tests (which I haven't even committed yet) are failing. Should we refactor this now? I cannot setup tests for this tag with the current code that's there. Sounds like we missed refactoring this when computeURL() was created in the first place, so +1 for refactoring it now. As of last night, I got all the test.tomcat.41 tests to run on my Linux box (yah :-) -- thanks for all the hard work on creating these tests. I ran into line-end problems on the code we generate for html:form and html:errors, so I tweaked them a bit. I hope that didn't mess up running the tests on Windows, and would be interested to find out if it did. Thoughts? Craig -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
On Mon, 10 Mar 2003, James Mitchell wrote: Date: Mon, 10 Mar 2003 11:44:01 -0500 From: James Mitchell [EMAIL PROTECTED] Reply-To: Struts Developers List [EMAIL PROTECTED] To: 'Struts Developers List' [EMAIL PROTECTED] Subject: RE: RequestUtils.computeURL() Sounds great. I'll take care of that tonight. Can anyone run test.tomcat.all? I guess I purchased some bad memory for my laptop or the OS/JVM is a piece of sh#t, because when I try to run them all, I get somewhere in the .40 tests and my laptop just shuts downno blue screenno OutOfMemory errorsit just stopslike pulling the plug on a desktop.very frustrating. The .40 tests do indeed work -- but only after a small change to each of the pages (CVS commit forthcoming). It's a bug in Tomcat 4.0 that we have to work around. At the bottom of many of the test pages, you've got a throw new JspException() statement if a test was misconfigured. Unfortunately, Tomcat 4.0 does an implicit import of the org.apache.jasper.runtime.* package into each page, and there's a org.apache.jasper.runtime.JspException class that causes JspException to be ambiguous. I changed it to use javax.servlet.jsp.JspException explicitly, which will still work on all the other containers, and am rerunning .41 right now to ensure that nothing got messed up. I'm going to have to go download Tomcat 3.3 to test those. And shouldn't we rip out the Tomcat 3.2 tests since we are no longer supporting it? -- James Mitchell Craig - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: RequestUtils.computeURL()
-Original Message- From: Craig R. McClanahan [mailto:[EMAIL PROTECTED] Sent: Monday, March 10, 2003 12:24 PM To: Struts Developers List Subject: RE: RequestUtils.computeURL() On Mon, 10 Mar 2003, James Mitchell wrote: Date: Mon, 10 Mar 2003 11:44:01 -0500 From: James Mitchell [EMAIL PROTECTED] Reply-To: Struts Developers List [EMAIL PROTECTED] To: 'Struts Developers List' [EMAIL PROTECTED] Subject: RE: RequestUtils.computeURL() Sounds great. I'll take care of that tonight. Can anyone run test.tomcat.all? I guess I purchased some bad memory for my laptop or the OS/JVM is a piece of sh#t, because when I try to run them all, I get somewhere in the .40 tests and my laptop just shuts downno blue screenno OutOfMemory errorsit just stopslike pulling the plug on a desktop.very frustrating. The .40 tests do indeed work -- but only after a small change to each of the pages (CVS commit forthcoming). It's a bug in Tomcat 4.0 that we have to work around. Yes, I can run test.tomcat.33, (or .40 or .41) 'one at a time', but it's when I run test.tomcat.all that I get to take a break while the laptop boots back up :( At the bottom of many of the test pages, you've got a throw new JspException() statement if a test was misconfigured. Unfortunately, Tomcat 4.0 does an implicit import of the org.apache.jasper.runtime.* package into each page, and there's a org.apache.jasper.runtime.JspException class that causes JspException to be ambiguous. I changed it to use javax.servlet.jsp.JspException explicitly, which will still work on all the other containers, and am rerunning .41 right now to ensure that nothing got messed up. Ah, see, I learn something new everyday. I'm going to have to go download Tomcat 3.3 to test those. And shouldn't we rip out the Tomcat 3.2 tests since we are no longer supporting it? Yes, I haven't been using those for a few months now (or least since we last discussed doing that) -- James Mitchell Craig -- James Mitchell Software Developer/Struts Evangelist http://jakarta.apache.org/struts People demand freedom of speech to make up for the freedom of thought which they avoid. - Soren Aabye Kierkegaard (1813-1855) - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
