Hi all! $ cat | mtn automate stdio l6:select5:k:blae mtn: warning: unknown selector type: k 0:0:l:0:select5:d:fooe 2:2:l:53:misuse: selector 'foo' is not a valid date
My question here: Should warnings in selectors.cc be actually errors so they can be properly catched up over stdio? From a user's perspective I see no real difference between both types of messages. Then I thought "maybe if the warning is issued, the invalid selector is just skipped, e.g. if multiple selectors are combined", so I tried l6:select11:k:bla/d:20*e mtn: warning: unknown selector type: k 0:2:l:53:misuse: selector '20*' is not a valid date and yes, apparently the parsing goes further. Now I tried something which should bring up a revision: l6:select11:k:bla/i:341e mtn: warning: unknown selector type: k 0:0:l:0: Huh? But the revision exists, the "i" selector alone brings it up: l6:select5:i:341e 1:0:l:41:341326bf4f07f6438d478f7db9b6a8dccb448e40 (I don't know if this is a bug or a feature, but as I acknowledged that I noticed that it is actually irrelevant to continue parsing anything if an unknown selector has been catched previously.) So in the end I see no reason why warnings should not just be handled as errors there as well... Opinions? Thomas. -- ICQ: 85945241 | SIP: 1-747-027-0392 | http://www.thomaskeller.biz > Guitone, a frontend for monotone: http://guitone.thomaskeller.biz > Music lyrics and more: http://musicmademe.com _______________________________________________ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel