Further cherry-picks for the c++20 branch #4225
Conversation
…or-(move_iterator, move_iterator)
|
Further suggestions:
|
There was a problem hiding this comment.
I think I'm fine with all of these -- though I don't feel motivated to take some of them, I'm also not opposed. I'm mainly just worried about the sticker shock of people seeing hundreds of changes between DIS and IS, even if many of them are trivial, but we've probably already crossed that bridge.
source/classes.tex
Outdated
| @@ -728,7 +728,7 @@ | |||
| with the same access control\iref{class.access} and | |||
| non-zero size\iref{intro.object} | |||
| are allocated so that later | |||
| members have higher addresses within a class object. | |||
| members have higher addresses within a class object\iref{expr.rel}. | |||
There was a problem hiding this comment.
Hm, [expr.rel] says that pointers to them compare greater, it doesn't say anything about "higher addresses" or allocation order. I don't think we actually guarantee anything about field layout order. (A weird implementation could presumably make lower addresses compare greater.)
I'm not sure the footnote makes this worse, though it might make the problem more obvious, because we're pointing directly to where we don't say the thing in question. I'm not sure I see motivation to add this xref for C++20.
There was a problem hiding this comment.
Fair enough, I'll just remove the commit!
source/expressions.tex
Outdated
| postfix-expression \opt{\terminal{.} \terminal{template}} id-expression\br | ||
| postfix-expression \opt{\terminal{->} \terminal{template}} id-expression\br | ||
| postfix-expression \terminal{.} \opt{\terminal{template}} id-expression\br | ||
| postfix-expression \terminal{->} \opt{\terminal{template}} id-expression\br |
There was a problem hiding this comment.
I omitted this one because it should make no difference to the rendered output.
There was a problem hiding this comment.
Right, that was precisely a reason in favour for me in order to keep diffs down, but I realize that it makes the Editor's Report more awkward, so I'll remove this and the next commit.
source/lex.tex
Outdated
| single quote & \tcode{'} & \tcode{\textbackslash\tcode{'}} \\ | ||
| double quote & \tcode{"} & \tcode{\textbackslash\tcode{"}} \\ | ||
| single quote & \tcode{'} & \tcode{\textbackslash '} \\ | ||
| double quote & \tcode{"} & \tcode{\textbackslash "} \\ |
There was a problem hiding this comment.
I omitted this one because it should make no difference to the rendered output.
There was a problem hiding this comment.
Agreed, see above, removing.
|
As for what else from the list that should be added... These seem like harmless yet nice to have consistency improvements: Several others are changes to the library for which I don't have an |
tkoeppe
force-pushed
the
c++20-extras
branch
from
56b5fca to
8c1918f
Compare
tkoeppe
force-pushed
the
c++20-extras
branch
from
8c1918f to
ff568a3
Compare
|
I remove the unobservable commits and the one with the questionable xref, and I added the commits for the extraneous table hrules and the missing friends. If this is OK now, I'd like to merge this and then restart a PR with burblebee's suggestions. |
|
@CaseyCarter: could you please review the suggestions in #4225 (comment)? |
|
I assume Dawn meant to get Casey to review the rest of the list that affects the library, not the ones in #4225 (comment) Anyway, here's my review of the ones in that comment (green where I feel my opinion counts, i.e. library wording): -[tab:atomic.types.pointer.comp] Fix column captions. (#4137) ✔️ |
|
You're right, that's probably how it was meant! OK, I'll merge this now and reopen a PR for additional discussions. Thank you! |
Yes, I meant for Casey (or someone from lib) to review #4220 (comment). But given Richard's concerns, I say we stop after pr4227. |
These are changes that either have no effect but decrease the diff between branches, or that are improvements to the overall quality of the final document.
Please discuss.
See #4220 (comment) for a list of candidate commits.