Hi Thiago! On Tue, Mar 3, 2020 at 6:37 PM Thiago H. de Paula Figueiredo < thiag...@gmail.com> wrote:
> Hello! > > On Sat, Feb 29, 2020 at 6:31 AM Ilya Obshadko <xf...@xfyre.com> wrote: > > > My application still runs on 5.4-beta-6 due to various breaking component > > changes introduced in later versions. > > > The ones in the other thread? If not, please post them here so we can > discuss them. > Long story short, I made a couple of attempts to upgrade (I think I try at least once every year) since 2015. Every time I ran into different issues. The last one with non-working nested FormFragment is one of those issues. Another one I have encountered recently is that my Typeahead mixin doesn't work anymore in last release versions. Typeahead duplicates an input field with some CSS classes that effectively hide it, but client-side validation 'unhides' the field and then fails. Since my application heavily relies on complex form logic, I just don't have enough time to go through all the issues, so every next attempt usually stops after I spend 4-6 hours without substantial progress. > I'm quite happy with it, but I'm > > getting concerned about outdated upstream dependencies (like Hibernate). > > > > Please file a Jira ticket about them. > I don't know if that's something that can be reasonably fixed by you guys. > > So I did the following: > > > > - upgraded everything to 5.4.5, adding tapestry-core to exclusions > > - added tapestry-core 5.4-beta-6 as a separate Maven dependency and > > excluded its upstream dependencies to other components. > > > The resulting hybrid of 5.4.5 and 5.4-beta-6 seems to work fine (although I > > didn't test everything yet). > > > > I have a question to Thiago, Lance and other committers - do you guys see > > any potential issues with this approach? Anything I might be missing? > > > > I don't see any potential issues, although I wouldn't guarantee they > wouldn't happen, even if I think the probability is very low. > I'm still running this hybrid branch locally every day. So far so good - I didn't encounter anything yet. I think I'll deploy it to production on the weekend. > On a related note: would it be possible to consider releasing > > *tapestry-corelib* (core components) as a completely separate artifact? > > This would make the maintenance a ton easier for edge cases like mine > (when > > updating to the latest components has a prohibitive cost). > > > > This is something we can do and it's been already suggested by someone else > in this mailing list which likes the Tapestry support for webapps (request > filters, dispatchers, resource support, etc), but doesn't actually use > pages and components. Since there's a plan to support Java 9 modules, if we > want to do a tapestry-corelib, we should do it sooner rather than later. > This would be really, really nice to have. This would allow developers to create independent foundation component libraries (I'm thinking of one for MaterialUI, for instance). -- Ilya Obshadko
