> Has anyone out there done any "real world" > testing of this package?
Since you asked... Being the spec lead for JSR-052 (JSPTL), I like to verify by myself that what we are trying to promote makes sense for the community (nothing like eating your own dog food). When JSPTL EA1 was first released, I've therefore converted quite a few pages of a commercial web site so it would use JSPTL EA1. This definitely cleaned up the JSP pages from all their control flow scriptlets, as well as how application data was accessed (using the SPEL). No problems to report. >From that experience, and for further cleanup of some pages, I also developed a JSPTL-based control flow taglib that complements the current set of standard control flow tags in JSPTL EA1 (see http://marc.theaimsgroup.com/?l=taglibs-dev&m=99714098228774&w=2). This complementary taglib is also being used at that commercial web site. Please note that this all runs on a JSP1.2 platform (tomcat 4.0). There could be some quirks on a JSP1.1 platform, depending on how the port of the taglib is done. You know all the disclaimers, but since you asked :-) -- Pierre Stuart Robertson wrote: > > Hello, > > I'm looking into using the SPEL packages in the JSPTL EA release for a > project. They would be used by a custom taglib, running on IBM's WebSphere > 3.5.x. I realize SPEL is pre-alpha, and likely to change or even be > replaced when JSPTL is released. But the introduction of an EL offers so > much flexibility that it would be nice to use this if the risks are > acceptable. Again, this is not a request for any guarantees or even implied > support... > > My question is, does anyone know of any issues with using SPEL in real > systems? I would just like to know if there are problems already known > that would make this package something to avoid until some future version. > I've done some initial testing, creating a JSP 1.1 version of JSPTL and > everything seems to work fine. Has anyone out there done any "real world" > testing of this package? > > Thanks, > > Stu
