[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 Jeremy Boynes changed: What|Removed |Added Status|NEEDINFO|RESOLVED

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #11 from Tomas Hofman --- I can no longer reproduce the issue with patched version of jstlel. Thanks Jeremy! -- You are receiving this mail because: You are the assignee for the bug.

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 Tomas Hofman changed: What|Removed |Added CC||thof...@redhat.com

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #10 from Tomas Hofman --- Yes, I think I do, but I'm afraid it's gonna involve Wildfly. I will check it out and get back here. -- You are receiving this mail because: You are the assignee for the bug.

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 Jeremy Boynes changed: What|Removed |Added Status|NEW |NEEDINFO ---

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #8 from Jeremy Boynes --- Created attachment 35845 --> https://bz.apache.org/bugzilla/attachment.cgi?id=35845=edit Adds new public method that allows the cache to be bypassed when parsing expressions. Based

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #7 from Tomas Hofman --- Jeremy, what do you think about the PR? https://github.com/jboss/jboss-jstl-api_spec/pull/16/files If you don't want to change current behaviour, what about at least preventing the

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #6 from Tomas Hofman --- Linking the JBoss issue for reference: https://issues.jboss.org/browse/JBEE-183 -- You are receiving this mail because: You are the assignee for the bug.

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #5 from Tomas Hofman --- Sorry for not elaborating before - in fact my main motivation was that currently the parseExpressionString() method can throw a NPE because of the race, which breaks the request

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #4 from Jeremy Boynes --- The current implementation disables caching while expressions are being validated during the translation process. This has the side effect of disabling the cache for evaluations that

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #3 from Jeremy Boynes --- I believe this patch reduces the window for the race but does not fully close it as another thread could still modify the field between when it is set and then read. It would be

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #2 from Tomas Hofman --- Created attachment 35602 --> https://bz.apache.org/bugzilla/attachment.cgi?id=35602=edit Patch -- You are receiving this mail because: You are the assignee for the bug.

[Bug 57434] Race condition in EL1.0 validation

https://bz.apache.org/bugzilla/show_bug.cgi?id=57434 --- Comment #1 from Tomas Hofman --- This concurrency issue was also hit in EAP/Wildfly. It is an issue even if the library is included inside a web application, e.g. when two different JSPs are requested and evaluated