The more I think about it, the more I think --download should imply --market. As you said, there is no use case for only --download, and it really does nothing without --market (the code is never executed and the internet query is never attempted). So, stating that you want ledger to query for prices implies you are interested in those prices and should use them if available, ergo --download implies --market.
ledger will error if --download is specified but not (--getquote and --price-exp ). (--getquote "path" is the new option to specify the location of the script) As far as stale prices go, the internet is only queried if there is no price already known, or that price's age in minutes exceeds --price-exp. If the user-supplied price script returns the latest value it can find and there was no price already known, that price should be used. If that price is older than the price already known, the newer price should be used, and no more queries should be made during this run of ledger. If the age of the price returned from the script exceeds --price-exp, ledger should issue a warning, but continue executing. Are there cases where people want it to error (stop processing) on getting a stale price? On Thu, Dec 8, 2011 at 08:30, thierry <[email protected]> wrote: > > > On Dec 8, 1:31 pm, Craig Earls <[email protected]> wrote: > > I agree about warnings if options don't make sense. I thought about > making > > --download imply --market since that is the only way info from downloaded > > quotes is used. > > A warning is enough for me. About implying, my personal way of working > is to prefer the explicit way to the implicit way, so I would prefer > here to not imply --market if --download is used. On the other side, > is there a use case of using --download without --market ? I do not > think so... > > > As far as stale prices I think you are saying that if getquote returns a > > stale price it should be rejected as if it it were never downloaded. Is > > that correct? > > My first thoughts are: > - If price-exp=20111201, and stale price is *before* expiry date, then > silently do *not* accept that stale price. > - If *no* price-exp=20111201, then any stale price *is* acceptable. > > Thierry > -- Craig, Corona De Tucson, AZ enderw88.wordpress.com
