Coherence: terminology, rationale, alternatives considered #624
Conversation
josh11b
changed the title
| Furthermore, this means that the behavior of a file can depend on an import even | ||
| if nothing from that package is referenced explicitly. | ||
|
|
||
| ## Rejected alternative: dynamic implementation binding |
There was a problem hiding this comment.
It's not clear to me from the title of the section what this is an alternative to...
It feels like we're rejecting all forms of incoherence, including an incoherent implementation that uses this alternative's strategy for resolving the challenges. Is my understanding correct?
If so, maybe sink this down a level of the TOC? And clarify above that all of the incoherent alternatives are rejected in some way?
Or maybe I've misunderstood the point a bit? Maybe there is another way to structure/name the sections?
(To be clear, this isn't really an issue with the content, just the structure / headings)
There was a problem hiding this comment.
I've redone the section headings. Is this what you were looking for?
| body of the `SomethingWeirdHappens` function. | ||
|
|
||
| This idea is discussed briefly in section 5.4 on separate compilation of | ||
| [this proposal for implementing "Indiana" C++0x concepts proposal](https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.86.9526&rep=rep1&type=pdf). |
There was a problem hiding this comment.
There is a WG21 link for this paper as well: https://wg21.link/n1848
Unsure which link is best really...
There was a problem hiding this comment.
Changed to include both.
|
We triage inactive PRs and issues in order to make it easier to find active work. If this PR should remain active, please comment or remove the |
github-actions
bot
added
the
inactive
|
We triage inactive PRs and issues in order to make it easier to find active work. If this PR should remain active or becomes active again, please reopen it. |
|
I am still interested in feedback here -- I believe I wrote extra because I wasn't sure what the real alternatives were. It would be great to cut it down to just what is needed. |
josh11b
removed
the
inactive
| - It requires more data space at runtime because we need to store a pointer to | ||
| the witness table representing the implementation with the object, since it | ||
| varies instead of being known statically. | ||
| - It is slower to execute from dynamic dispatch and the inability to inline. |
There was a problem hiding this comment.
... and dynamic dispatch may not even be feasible in all cases.
There was a problem hiding this comment.
What cases are you thinking of?
There was a problem hiding this comment.
Is this things like we need to know the size of the type?
There was a problem hiding this comment.
Yes, things like that to do with how we would generate code. I had in mind cases where, for example, a method in the interface returns an associated type, and we don't know the calling convention of the function without knowing some details about the type.
There was a problem hiding this comment.
I've added that example.
Co-authored-by: Richard Smith <[email protected]>
Add an appendix with the rationale for and alternatives to coherence, along with an entry in the terminology and updates to the coherence discussion in the goals. Co-authored-by: Chandler Carruth <[email protected]> Co-authored-by: Richard Smith <[email protected]>
Add an appendix with the rationale for and alternatives to coherence, along with an entry in the terminology and updates to the coherence discussion in the goals.