Actually, Tapestry's Response interface and ResponseImpl implementation (a thin facade over HttpServletResponse) don't include an addHeader method, so we would either have to 1) combine the two setHeader calls or 2) use HttpServletResponse directly, or 3) change the Response interface (a backward-incompatible API change, although maybe worth doing, since I guess very few people would be affected).
On Fri, Jan 24, 2014 at 2:56 PM, Dimitris Zenios <dimitris.zen...@gmail.com> wrote: > It would be nice also to provide a patch file.The only change needed is to > replace the second setHeader with addHeader. After that we can inform a > person with commit access to merge your patch. > > > On Fri, Jan 24, 2014 at 1:22 AM, Ilya Obshadko <ilya.obsha...@gmail.com>wrote: > >> Okay, https://issues.apache.org/jira/browse/TAP5-2275 >> >> >> On Fri, Jan 24, 2014 at 8:46 AM, Dimitris Zenios >> <dimitris.zen...@gmail.com>wrote: >> >> > You should start by writing a jira bug report >> > On 23 Jan 2014 00:09, "Ilya Obshadko" <ilya.obsha...@gmail.com> wrote: >> > >> > > So I have discussed the issue with Kaptcha library author and found out >> > > that the problem is, in fact, with Tapestry implementation itself. >> > > >> > > https://code.google.com/p/kaptcha/issues/detail?id=72 >> > > >> > > Tapestry component uses response.setHeader("Cache-Control") twice, and >> > the >> > > second call obviously overrides the first one. Is it going to be fixed? >> > > >> > > >> > > >> > > On Tue, Jan 21, 2014 at 12:59 PM, Ilya Obshadko < >> ilya.obsha...@gmail.com >> > > >wrote: >> > > >> > > > I'm not sure if anyone else has that issue: reloading the zone >> > containing >> > > > kaptcha component does not trigger image reloading, so the whole >> thing >> > > > becomes unusable (because internal CAPTCHA text is already different, >> > but >> > > > image is not). This affects ONLY Firefox, all the other major >> browsers >> > > are >> > > > OK. Don't have an idea why (probably Firefox doesn't respect cache >> > > control >> > > > headers because of some reasons). >> > > > >> > > > After spending some time to fix it, I've ended up with simple >> > > > copy-and-paste versions of KaptchaField and KaptchaImage classes with >> > > only >> > > > one small difference in KaptchaImage: >> > > > >> > > > >> > > > boolean beginRender(MarkupWriter writer) >> > > > >> > > > { >> > > > >> > > > captchaText = producer.createText(); >> > > > >> > > > >> > > > >> > > > Link link = resources.createEventLink("image", >> > > > RandomStringUtils.randomAlphanumeric ( 8 ) ); >> > > > >> > > > >> > > > writer.element("img", >> > > > >> > > > >> > > > "src", link.toURI(), >> > > > >> > > > >> > > > "width", producer.getWidth(), >> > > > >> > > > >> > > > "height", producer.getHeight()); >> > > > >> > > > >> > > > resources.renderInformalParameters(writer); >> > > > >> > > > >> > > > writer.end(); >> > > > >> > > > >> > > > return false; >> > > > >> > > > } >> > > > >> > > > It might be a good idea to fix it in the main branch. >> > > > >> > > > -- >> > > > Ilya Obshadko >> > > > >> > > > >> > > >> > > >> > > -- >> > > Ilya Obshadko >> > > >> > >> >> >> >> -- >> Ilya Obshadko >> --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
