You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
What does this option actually do? It's not being used to resolve Int to Int32/Int64, because we still see Ints in the symbolgraph. Maybe the only important part of the triple is the OS? Or maybe it doesn't do anything at all to the symbolgraph, and only affects the other outputs like the dylib? But if we don't pass this option when extracting a symbolgraph from a precompiled module, we get an error.
Try passing a bunch of different targets, varying CPU, OS (iOS vs macOS vs simulator), and OS version, and diff the json symbolgraphs.
Once we know what effect it has on the output, we can improve the API. If the OS is the only bit that matters, replace the target triple config option with an OS enum. If it has no effect, it should be removed. If it has some other important effect, we should consider generating symbolgraphs for all targets (or perhaps a user selected subset) and merging the symbolgraphs.
The text was updated successfully, but these errors were encountered:
What does this option actually do? It's not being used to resolve
Int
toInt32
/Int64
, because we still seeInt
s in the symbolgraph. Maybe the only important part of the triple is the OS? Or maybe it doesn't do anything at all to the symbolgraph, and only affects the other outputs like the dylib? But if we don't pass this option when extracting a symbolgraph from a precompiled module, we get an error.Try passing a bunch of different targets, varying CPU, OS (iOS vs macOS vs simulator), and OS version, and diff the json symbolgraphs.
Once we know what effect it has on the output, we can improve the API. If the OS is the only bit that matters, replace the target triple config option with an OS enum. If it has no effect, it should be removed. If it has some other important effect, we should consider generating symbolgraphs for all targets (or perhaps a user selected subset) and merging the symbolgraphs.
The text was updated successfully, but these errors were encountered: