Details
- Reviewers
rsmith bruno - Commits
- rG14a3f77ba103: [Implicit Modules] Add -cc1 option -fmodules-strict-context-hash which includes…
rL375322: [Implicit Modules] Add -cc1 option -fmodules-strict-context-hash which includes…
rC375322: [Implicit Modules] Add -cc1 option -fmodules-strict-context-hash which includes…
Diff Detail
- Repository
- rC Clang
Event Timeline
Hi Michael, thanks for working on this!
include/clang/Lex/HeaderSearchOptions.h | ||
---|---|---|
206 | *whether | |
209 | What else do you plan to add in the future as part of "all the things that could impact"? It seems to me that by default this should always be the case, but header search related things are special because one might want to handwave on correctness to have smaller caches (default behavior right now). I wonder if we should instead have fmodules-strict-header-seach and later on add a more generic thing that group such cases? WDYT? |
include/clang/Lex/HeaderSearchOptions.h | ||
---|---|---|
209 | This also includes diagnostic options. I'm not sure it's useful for users to pick each individual thing they care about for the hash. The intent here is just to capture every possible difference we find. |
LGTM with one minor change. Can you add an entry in the modules docs for this flag and mention that using it can lead to more PCMs in an implicit build?
*using implicit modules in a build where compiler flags in different invocations aren't homogeneous, or something along those lines.
While adding the documentation I realized that a better name for this option would be -fmodules-strict-context-hash to make it clear which hash it's referring to.
While adding the documentation I realized that a better name for this option would be -fmodules-strict-context-hash to make it clear which hash it's referring to.
-fmodules-strict-context-hash SGTM!
Looks like this fails on win: http://45.33.8.238/win/841/step_6.txt
Ptal!
Maybe just cat'ing all files instead of echoing the first and piping into cat works?
*whether