dxgi: Add DXGI_OUTDUPL_FLAG enum annotation to DuplicateOutput1() - #1981
dxgi: Add DXGI_OUTDUPL_FLAG enum annotation to DuplicateOutput1()#1981Marijn Suijten (MarijnS95) wants to merge 1 commit into
Conversation
| <IncludeRoot>/um/dxprogrammablecapture.h | ||
| --namespace | ||
| Windows.Win32.Graphics.Dxgi | ||
| --with-attribute |
There was a problem hiding this comment.
Inversely, could we also have all the =AssociatedEnum in here? It seems much nicer to me to have configuration in the namespace-specific settings.rsp than in one humongous global file, if possible?
There was a problem hiding this comment.
No. The scraper and the emitter have different roles and supporting toolchains.
There was a problem hiding this comment.
Is there any complete docs on what configures the scraper (probably the namespace-specific settings.rsp) and what configures the emitter (probably the global emitter.settings.rsp... 🙂), and when to use either?
There was a problem hiding this comment.
There was a problem hiding this comment.
Mike Battista (@mikebattista) unfortunately that doesn't explain directly which flags are available, nor what precedence I should stick to. Searching further in the repo shows that --enumMakeFlags is a feature of ClangSharpSourceToWinmd so it makes sense that that only works via emitter settings (and likewise, --with-attribute seems to be a feature of the scraper).
The question remains: do we prefer to set the Flags attribute on the emitter via --enumMakeFlags, or do we prefer to set it earlier via --with-attribute on the scraper as done before this PR?
There was a problem hiding this comment.
Mike Battista (@mikebattista) are you able to help with the above question, so that we can further this PR?
There was a problem hiding this comment.
Resolving this thread as it seems like a tangent and unrelated to the PR.
There was a problem hiding this comment.
Rafael Rivera (@riverar) it's a valid question to ask whether to configure partition-specific types inside the... partition, or in the global file (and if either method has repercussions). That affects how this PR is structured (i.e. related) which now mostly hacks more types into the global file.
Sure we can merge this without that discussion since there are many more DXGI annotations to move if partition-specific configuration is preferred, but we should still have that discussion (and resulting documentation improvements and .rsp changes somewhere). Should I click "reference in new issue"?
26da534 to
c4cc88f
Compare
c4cc88f to
4e0588c
Compare
4e0588c to
de8ad04
Compare
de8ad04 to
04e7edd
Compare
04e7edd to
e091ffd
Compare
On the side, alphabetically sort some flags and use `--enumMakeFlags` in `emitter.settings.rsp` instead of `--with-attribute XXX=Flags` in `Namespace/settings.rsp`.
e091ffd to
25ad66f
Compare
3 participants
Separated from #1911 (comment).
On the side, alphabetically sort some flags and use
--enumMakeFlagsinemitter.settings.rspinstead of--with-attribute XXX=FlagsinNamespace/settings.rsp.