PersistentGenericBag? It is not only a collection, but tightly bound to NHibernate collection persistence logic. I was thinking of the NHibernate.Util.LinkedHashMap example, but preferably truly internal. (Even if testing it is then less convenient.)
In my mind most of the NHibernate.Util namespace is here for supporting implementation details of NHibernate and should be turned to internal, like IdentitySet, JoinedEnumerable, NullableDictionnary, SafetyEnumerable, ... Or the other way round, moved to Iesi instead, and then used from Iesi, which is targeted at supplying custom collections and so way more appropriate than NHibernate for supplying such things. P.S.: I have sent a bug report about the email addresses, hoping they will fix that. This is the only reasonable solution (unless we change settings for forbidding mail responses and enforce responses by the web interface, which nicely reports mails address when trying to publish). Here the change to x...@free.fr was not enough because there was still an underlying mailto:completeAddress. Le dimanche 19 novembre 2017 15:09:19 UTC+1, Fabio Maulo a écrit : > > Ok. *internal* as PersistentBag is (for example). > > On Sun, Nov 19, 2017 at 5:38 AM, Frédéric Delaporte <@free.fr> wrote: > >> I do not care about keeping Iesi or not, I just care on keeping the >> feature working in a satisfactory way. >> So with those conditions, it should be alright. >> >> NHibernate already have some own internal specialized >> collection/dictionary for its internal uses. Why not adding another one. >> But it should be internal, since this is not the purpose of NHibernate to >> supply specialized collections to consume. >> Then if the user needs to instantiate some ordering preserving hashset on >> his side, it will be up to him to use some adequate implementation, be it >> Iesi or something else. >> >> Le dimanche 19 novembre 2017 00:24:29 UTC+1, Fabio Maulo a écrit : >>> >>> Sir, yes sir. >>> >>> Can we agree about this: >>> . The PR should allow the usage of order-by in <set> mapping >>> . As all other *persistent collection* NH may have its own >>> implementation to ensure the behavior just when the collection comes from NH >>> . The user is responsible to use an ISet implementation that >>> maintains the order when he works on its side >>> . The PR should pass at least all existing tests >>> >>> Is it ok ? >>> >>> -- >> >> --- >> You received this message because you are subscribed to the Google Groups >> "nhibernate-development" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to nhibernate-development+unsubscr...@googlegroups.com. >> For more options, visit https://groups.google.com/d/optout. >> > > > > -- > Fabio Maulo > *T* @fabiomaulo > *S fabiomaulo* > -- --- You received this message because you are subscribed to the Google Groups "nhibernate-development" group. To unsubscribe from this group and stop receiving emails from it, send an email to nhibernate-development+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
