On Wed, Jul 2, 2014 at 9:35 AM, Simon Michael <[email protected]> wrote:
> On 7/1/14 7:45 PM, Martin Blais wrote: > >> Alright... so here's the punchline: I'm claiming that if Ledger or >> HLedger gave up on balance assertions attached to postings, if you >> separated them as I do, just that change, you would both very likely be >> able to guarantee order-independence w.r.t. your input files! Think >> > > Indeed we would, of course, at the cost of giving up intra-day assertions. > > I don't feel too strongly about intra-day assertions either way. But I > think hledger needs to support Ledger's style of assertions, at least. > (Though I just added a --ignore-assertions flag as an alternative.) > > We could also name and implement all of these three kinds of assertions: > > - ledger's single-commodity, ordered, posting-based assertions > - hledger's multi-commodity, ordered, posting-based assertions (using ==) > - beancount's order-independent, directive-based assertions (do these > assert one commodity or all of them ?) > For now, a single commodity. I'm attracted to the idea of extending this to cover the entirety of the inventory, I think this is going to be added eventually. I'm also vaguely interested in ideas for supporting intra-day balance assertions in a way that preserves order-independence. I'm not sure that it's possible, but it's worth throwing around ideas. This gives most power and compatibility.. but also more feature bloat. > Anyway I probably shouldn't think about supporting beancount format with > hledger's main parser, due to its other differences. Oh yeah... I've walked away from even trying to be compatible years ago. I want to push things forward aggressively, especially w.r.t. to the inventory booking semantics, I have a real need for booking at average cost: https://docs.google.com/document/d/1F8IJ_7fMHZ75XFPocMokLxVZczAhrBRBVN9uMhQFCZ4/ On Wed, Jul 2, 2014 at 9:41 AM, Simon Michael <[email protected]> wrote: > On 7/2/14 6:35 AM, Simon Michael wrote: > >> I don't feel too strongly about intra-day assertions either way. But I >> think hledger needs to support Ledger's style of assertions, at least. >> (Though I just added a --ignore-assertions flag as an alternative.) >> > > To be clear: I mean I'd like hledger to include support for the idiomatic > style of assertions Ledger uses going forward, whatever that ends up being. If you're trying to be compatible, that makes sense. -- --- You received this message because you are subscribed to the Google Groups "Ledger" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
