[string.cons] Replace postcondition tables by ordinary Postcondition elements #1876
Conversation
|
I think you lost the former table 61. Shouldn't there be a "Postcondition:" after (what is now) 24.3.2.2/25? |
|
On the other hand, maybe table 61 was just in the wrong place before (in N4713). |
|
@mclow: The old Table 61 is now paragraph 23. |
source/strings.tex
Outdated
| \tcode{data()} points at the first element of an allocated copy | ||
| of the array whose first element is pointed at by \tcode{s}, | ||
| \tcode{size()} is equal to \tcode{traits::length(s)}, and | ||
| \tcode{capacity()} is a value at least as large as \tcode{size()}. | ||
| \end{itemdescr} |
There was a problem hiding this comment.
Duplicate space after capacity()? (Not that it matters.)
|
Looks good to me. |
source/strings.tex
Outdated
| \postconditions | ||
| \tcode{data()} points at the first element of an allocated copy of the | ||
| array whose first element is pointed at by \tcode{str.data()}, | ||
| \tcode{size()} is equal to \tcode{str.size()}, and \tcode{capacity()} is |
There was a problem hiding this comment.
I'm not sure this is right. size() is equal to the value that str.size() had prior to the call, right? str.size() is unspecified for the second overload. (Likewise for data().) I think we're usually more explicit about whether we're talking about the before / after value.
There was a problem hiding this comment.
Would it help if I added "the original value of" in a few places?
There was a problem hiding this comment.
@zygoloid: "original value" is what the allocator versions already use. In fact, my attempt to add "original value" to the description here led me to independently invent identical wording!
source/strings.tex
Outdated
| \tcode{*this} | ||
| and \tcode{str} are the same object, | ||
| the member has no effect. | ||
| If \tcode{*this} and \tcode{str} are not the same object, modifies \tcode{*this}. |
There was a problem hiding this comment.
I think we still need to say "as described below" or something here. Otherwise, a reader is left thinking "ok, but how?" Maybe this should be merged into the postconditions instead, though, since the "points at [...] an allocated copy" postcondition is not correct in the this == &str case.
There was a problem hiding this comment.
I turned the two conditions around and made the second one say "see below".
source/strings.tex
Outdated
| \tcode{alloc}. In the second form, \tcode{str} is left in a valid state with an | ||
| \effects Constructs an object of class \tcode{basic_string}. | ||
| The stored allocator is constructed from \tcode{alloc}. | ||
| In the second form, \tcode{str} is left in a valid state with an |
There was a problem hiding this comment.
For the corresponding constructors without an allocator parameter, this sentence is a postcondition rather than an effect. We should be consistent.
source/strings.tex
Outdated
| \tcode{data()} points at the first element of an allocated copy | ||
| of the array whose first element is pointed at by \tcode{str.data()}, | ||
| \tcode{size()} is equal to \tcode{str.size()}, and | ||
| \tcode{capacity()} is a value at least as large as \tcode{size()}. |
There was a problem hiding this comment.
I think the "If *this and str are the same object" condition should be moved down here -- the postcondition should ideally be written such that it applies for all (valid) calls to the function.
There was a problem hiding this comment.
Moved or copied? Do you want to delete the Effects clause?
source/strings.tex
Outdated
| \end{libefftabvalue} | ||
| \pnum | ||
| \postconditions | ||
| \effects |
There was a problem hiding this comment.
Yikes, also stray pnum above. (Which prints very strangely.) Fixed.
Fixes #1875.