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
Managing a list of identifiers that will be globally renamed in all instances can solve these limitations:
unclean macros (macros containing references to identifiers other than their own arguments) break
macro names are not shortened
user-defined type names are not shortened (# 290 "Minifying structs ...")
Global renaming will be subject to some restrictions:
Any variable or function whose name is a globally renamed identifier must be excluded from inlining, reordering, and dead code removal.
Global renaming short names may not be reused for normal scoped renaming.
Managing a list of struct member names and their renamings can solve #290 "...and their members". Consistent struct member renaming avoids the issue of needing to know an expression's type.
Please leave a message before implementing a significant change, so that we can agree on the solution.
The implementations for each of these mechanisms are similar. In each case: it falls somewhat short of optimal renaming analysis, but the implementation is far simpler than for a fully general solution. Is this reasonable?
Exposing --global-renaming-list option to the user may be usable for workarounds in some situations.
The text was updated successfully, but these errors were encountered:
Managing a list of identifiers that will be globally renamed in all instances can solve these limitations:
Global renaming will be subject to some restrictions:
Managing a list of struct member names and their renamings can solve #290 "...and their members". Consistent struct member renaming avoids the issue of needing to know an expression's type.
The implementations for each of these mechanisms are similar. In each case: it falls somewhat short of optimal renaming analysis, but the implementation is far simpler than for a fully general solution. Is this reasonable?
Exposing
--global-renaming-list
option to the user may be usable for workarounds in some situations.The text was updated successfully, but these errors were encountered: