}\0abody:\0a{<body>}") :::: :: Remarks: :: :: There are two recursions in the log file: :: 1. the outer loop, `commit-loop` which threads downwards into each commit by ancestor :: 2. the inner loop, `ancestor-loop`, which threads left-to-right on reverse-ancestors :::: ++ story-log |= [current-commit=yaki:clay tale=story] ^- tang =/ our p.bec =/ cas=case [%da now] %- flop %- head =| state=[result=tang mergebase=(unit tako:clay)] |- ^- _state =* commit-loop $ =/ reverse-ancestors (flop p.current-commit) |- =* ancestor-loop $ ?- reverse-ancestors ~ :: stop here and return the current message =/ msg=(unit cord) (msg-from-commit current-commit tale) [?~(msg result.state [u.msg result.state]) mergebase=~] :: [tako:clay ~] =/ parent=tako:clay i.reverse-ancestors =/ parent-commit .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv parent)) =/ msg (msg-from-commit current-commit tale) :: proper logic is: if there is a mergebase and the current commit is the mergebase, stop. otherwise, continue ?: ?&(?=(^ mergebase.state) =(u.mergebase.state r.current-commit)) :: stop here and clear mergebase. we skip adding the mergebase's msg itself because it will be added through the other branch [result=result.state mergebase=~] commit-loop(current-commit parent-commit, state [?~(msg result.state [u.msg result.state]) mergebase.state]) :: [tako:clay tako:clay ~] :: :: mainline: ultimate base chain :: nextline: relative mainline :: sideline: side-chain, featurebranch :: :: From the context of e, commit c is on its relative mainline, or nowline, while commit d is on its sideline :: %base a--b-------------X :: mainline :: %new \--c------e--/ :: nowline :: %new2 \--d--/ :: sideline :: :: ~& "in merge commit" =/ sideline=tako:clay i.reverse-ancestors =/ mainline=tako:clay i.t.reverse-ancestors =/ mergebases .^((list tako:clay) %cs /(scot %p our)/base/(scot cas)/base-tako/(scot %uv mainline)/(scot %uv sideline)) =/ next-mergebase=(unit tako:clay) ?~(mergebases ~ (some i.mergebases)) :: take the first valid mergebase (by convention) if exists, else none =/ sideline-commit=yaki:clay .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv sideline)) =/ mainline-commit=yaki:clay .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv mainline)) =/ msg=(unit cord) (msg-from-commit current-commit tale) :: :: 1 - process current commit :: 2 - recur and queue processing on all commits on the sideline :: 3 - recur and queue processing on all commits on the mainline :: :: Because mainline messages are cons'd to result last, they are :: (by definition) towards the less recent side of the (resulting, flopped) list :: =. state [result=?~(msg result.state [u.msg result.state]) mergebase=next-mergebase] :: 1 =. state commit-loop(current-commit sideline-commit) :: 2 =. state commit-loop(current-commit mainline-commit) :: 3 state :: [tako:clay tako:clay tako:clay *] ~& "in 3+ ancestor commit" :: take the first two (sideline/mainline). calculate that mergebase :: recur through sideline (state-b). but instead of state-c line for mainline, =/ sideline=tako:clay i.reverse-ancestors =/ nowline=tako:clay i.t.reverse-ancestors =/ mergebases .^((list tako:clay) %cs /(scot %p our)/base/(scot cas)/base-tako/(scot %uv nowline)/(scot %uv sideline)) =/ next-mergebase=(unit tako:clay) ?~(mergebases ~ (some i.mergebases)) :: take the first valid mergebase (by convention) if exists, else none =/ sideline-commit=yaki:clay .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv sideline)) =. mergebase.state next-mergebase =. state commit-loop(current-commit sideline-commit) =. state ancestor-loop(reverse-ancestors t.reverse-ancestors) state == --

:: :::: :: /- *story :- %say |= $: [now=@da eny=@uvJ bec=beak] [[syd=desk ~] ~] == |^ =/ our p.bec =/ cas `case`[%da now] =/ tale .^(story %cx /(scot %p our)/[syd]/(scot %da now)/story) =/ current-tako .^(tako:clay %cs /(scot %p our)/[syd]/(scot cas)/tako/~) =/ current-yaki .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv current-tako)) :- %tang (story-log current-yaki tale) :: ++ msg-from-commit |= [commit=yaki:clay tale=story] ^- (unit cord) =/ proses (~(get by tale) r.commit) ?~ proses ~ =/ prose-list ~(tap in u.proses) :: TODO signal conflict =/ prose-first (snag 0 prose-list) =/ title title:prose-first =/ body body:prose-first %- some (crip "commit: {<`@uv`r.commit>}\0atitle: {}\0abody:\0a{<body>}") :::: :: Remarks: :: :: There are two recursions in the log file: :: 1. the outer loop, `commit-loop` which threads downwards into each commit by ancestor :: 2. the inner loop, `ancestor-loop`, which threads left-to-right on reverse-ancestors :::: ++ story-log |= [current-commit=yaki:clay tale=story] ^- tang =/ our p.bec =/ cas=case [%da now] %- flop %- head =| state=[result=tang mergebase=(unit tako:clay)] |- ^- _state =* commit-loop $ =/ reverse-ancestors (flop p.current-commit) |- =* ancestor-loop $ ?- reverse-ancestors ~ :: stop here and return the current message =/ msg=(unit cord) (msg-from-commit current-commit tale) [?~(msg result.state [u.msg result.state]) mergebase=~] :: [tako:clay ~] =/ parent=tako:clay i.reverse-ancestors =/ parent-commit .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv parent)) =/ msg (msg-from-commit current-commit tale) :: proper logic is: if there is a mergebase and the current commit is the mergebase, stop. otherwise, continue ?: ?&(?=(^ mergebase.state) =(u.mergebase.state r.current-commit)) :: stop here and clear mergebase. we skip adding the mergebase's msg itself because it will be added through the other branch [result=result.state mergebase=~] commit-loop(current-commit parent-commit, state [?~(msg result.state [u.msg result.state]) mergebase.state]) :: [tako:clay tako:clay ~] :: :: mainline: ultimate base chain :: nextline: relative mainline :: sideline: side-chain, featurebranch :: :: From the context of e, commit c is on its relative mainline, or nowline, while commit d is on its sideline :: %base a--b-------------X :: mainline :: %new \--c------e--/ :: nowline :: %new2 \--d--/ :: sideline :: :: ~& "in merge commit" =/ sideline=tako:clay i.reverse-ancestors =/ mainline=tako:clay i.t.reverse-ancestors =/ mergebases .^((list tako:clay) %cs /(scot %p our)/base/(scot cas)/base-tako/(scot %uv mainline)/(scot %uv sideline)) =/ next-mergebase=(unit tako:clay) ?~(mergebases ~ (some i.mergebases)) :: take the first valid mergebase (by convention) if exists, else none =/ sideline-commit=yaki:clay .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv sideline)) =/ mainline-commit=yaki:clay .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv mainline)) =/ msg=(unit cord) (msg-from-commit current-commit tale) :: :: 1 - process current commit :: 2 - recur and queue processing on all commits on the sideline :: 3 - recur and queue processing on all commits on the mainline :: :: Because mainline messages are cons'd to result last, they are :: (by definition) towards the less recent side of the (resulting, flopped) list :: =. state [result=?~(msg result.state [u.msg result.state]) mergebase=next-mergebase] :: 1 =. state commit-loop(current-commit sideline-commit) :: 2 =. state commit-loop(current-commit mainline-commit) :: 3 state :: [tako:clay tako:clay tako:clay *] ~& "in 3+ ancestor commit" :: take the first two (sideline/mainline). calculate that mergebase :: recur through sideline (state-b). but instead of state-c line for mainline, =/ sideline=tako:clay i.reverse-ancestors =/ nowline=tako:clay i.t.reverse-ancestors =/ mergebases .^((list tako:clay) %cs /(scot %p our)/base/(scot cas)/base-tako/(scot %uv nowline)/(scot %uv sideline)) =/ next-mergebase=(unit tako:clay) ?~(mergebases ~ (some i.mergebases)) :: take the first valid mergebase (by convention) if exists, else none =/ sideline-commit=yaki:clay .^(yaki:clay %cs /(scot %p our)/base/(scot cas)/yaki/(scot %uv sideline)) =. mergebase.state next-mergebase =. state commit-loop(current-commit sideline-commit) =. state ancestor-loop(reverse-ancestors t.reverse-ancestors) state == --