On the topic of tests and corner cases: http://hypothesis.works is definitely work one or several looks, especially in this type of software that is mostly a fancy frontend and language for doing basic calculus operations.
I also agree that it is far too easy to get the "happy path" working and then throw others in a world of pain and debugging for their slightly off use case. Am Dienstag, 6. Februar 2018 07:07:11 UTC+1 schrieb Martin Blais: > > There's no ticket but it definitely merits one: > > https://bitbucket.org/blais/beancount/issues/213/implement-average-cost-booking > > It's probably not super difficult to implement, but the problem is that > I'm afraid there may be very many corner cases and it will probably require > a ton of unit testing scenarios. Not something to hash out on just a quick > Sunday afternoon. > > Always open to contributions, but as always... a lot of us use that thing > twice / week and we simply _cannot_ break it, so heavy coverage in unit > testing - especially for stuff like this - is super important if you'd like > to bring it to completion. There's little more frustrating to me than > needing to update my Ledger and finding bugs (it immediately throws off the > next couple of hours of my life into a fit of rather unhappy "unplanned" > coding session). Often people make submissions without tests, and of course > as soon as you write tests you realize there's always a /lot/ more to it > than it seemed... like 10x more, so these inevitably go on the backburner. > Anyhow, just saying. I've been very conservative about patches in order to > avoid risking the stability of Beancount. If you understand and you think > you can replicate a simliar abundance of test scenarios as those which are > in the codebase, you're welcome to take a shot at it, I'd look at it. OTOH > a patch without any tests is almost guaranteed to sit there for a long time > (it's very easy and quick to hack something that appears to work). > > > > On Thu, Feb 1, 2018 at 9:21 PM, Robert Sesek <rse...@bluestatic.org > <javascript:>> wrote: > >> Hi Martin, >> >> Is there an issue that tracks the remaining work items for the average >> cost booking method (I searched and couldn't find one)? I see that there's >> at least a partial implementation of booking_method_AVERAGE >> <https://bitbucket.org/blais/beancount/src/9c9de54513aba0b55db8e7be626c8cdf9071fa67/beancount/parser/booking_method.py?at=default&fileviewer=file-view-default#booking_method.py-168>, >> >> but I have to imagine it's incomplete since it is currently disabled. >> >> Upthread you noted this is not an area you would want contributions for. >> But if there were ways to make incremental progress on average booking, I'd >> be interested in helping out. >> >> Thanks, >> Robert >> >> On Tue, Jan 23, 2018 at 9:58 PM Martin Blais <bl...@furius.ca >> <javascript:>> wrote: >> >>> Actually, it gets parsed, but it doesn't get used, as the general lot >>> merging functionality should be common in AVERAGE booking and that's the >>> one booking method not implemented. >>> I'll make it generate an error for now. >>> >>> On Mon, Jan 22, 2018 at 1:15 PM, Dennis Megaffin <dennism...@gmail.com >>> <javascript:>> wrote: >>> >>>> Hi Martin, >>>> >>>> I am a new beancount user. Just wondering if average cost booking is >>>> working yet? >>>> >>>> Is the merge flag now working? {*} >>>> >>>> Thanks for any and all update! >>>> >>>> -Dennis M >>>> >>>> P.S. Thanks for all your contributions! >>>> >>> -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/f9bbf4bc-03ae-4179-aec4-e7e7d8a57d37%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
