diff options
Diffstat (limited to 'app/Main.hs')
| -rw-r--r-- | app/Main.hs | 39 |
1 files changed, 0 insertions, 39 deletions
diff --git a/app/Main.hs b/app/Main.hs index 5fea8dc..4491747 100644 --- a/app/Main.hs +++ b/app/Main.hs @@ -1,42 +1,3 @@ --- TODO Separate database issues and history issues --- --- History code re-uses the issue concept which leads to undesirable effects. --- --- (1) History issues don't hold up to the guarantee that provenance is --- computed over the history, because we compute issues with trivial provenance --- per commit, and then later merge those trivial provenance in an --- history-encompassing provenance. --- --- (2) Internal tags depend on provenance, and thus, have to re-computed after --- the final provenance has been generated. --- --- (3) We need to compare issues in history code, and several fields of issue --- prevent us from doing that cleanly. --- --- To remedy, history should introduce its own notion (data type) of issue, --- and operate on that. --- --- @supersedes fix-issue-comparison --- @supersedes refactor-non-issues --- --- @difficulty medium --- @priority high --- --- @assigned aforemny --- @status in-progress --- --- COMMENT I've attempted this, and have not been happy with the outcome. --- Short of separating provenance from issues, ie. maintaining a separate --- structure to track provenance, we cannot overcome (1) since it is part --- of the algorithm. --- --- I've overcome (2) by not memoizing `internalTags` and overcome (3) by --- capturing `rawText` and comparing that. --- --- I am as of now not convinced that separating provenance into an --- external structure is the way forward (nor am I convinced that is is --- not). - -- TODO Parse issues as markdown -- -- There are several issues related to the fact that we are not parsing |