4 messages in org.apache.tomcat.devRe: Fixing bug 33453| From | Sent On | Attachments |
|---|---|---|
| Mark Thomas | May 20, 2011 9:30 am | |
| Tim Funk | May 23, 2011 5:22 am | |
| Konstantin Kolinko | May 23, 2011 2:01 pm | |
| Mark Thomas | Jun 20, 2011 10:47 am |
| Subject: | Re: Fixing bug 33453 | |
|---|---|---|
| From: | Tim Funk (funk...@apache.org) | |
| Date: | May 23, 2011 5:22:26 am | |
| List: | org.apache.tomcat.dev | |
How about calling it getDependantsMap() instead of getDependants() ?
Then catch the NoSuchMethodException (I in the case where the compiled jsp doesn't have getDependants() and that would force a recompile (and then the new version would be OK)
-Tim
On 5/20/2011 12:31 PM, Mark Thomas wrote:
All,
I've been looking at [1]. Ignoring the flames, there do appear to be several use cases where the current time-stamp checks are insufficient (although there are simple work-arounds). I have a patch [2] but I don't particularly like the fact that it breaks binary compatibility with JSPs compiled with an earlier version. My instinct is that this is bad. What does everyone else think?
I do have an idea for addressing this: - Leave JspSourceDependent as is in 7.0.14 but deprecate it - Add a new JspSourceDependent2 interface (better names welcome) - Compilation always uses JspSourceDependent2 - Isoutdated checked for JspSourceDependent as well as JspSourceDependent2 and any classes implementing JspSourceDependent are treated as outdated.
Thoughts?
Mark
[1] https://issues.apache.org/bugzilla/show_bug.cgi?id=33453 [2] https://issues.apache.org/bugzilla/attachment.cgi?id=27040&action=diff