Hi Tom,

 

Thank you for this updated draft.

 

It would be useful to have a column in the table that shows the proposed conforming behaviour.

I presume you mean in the table in the "Existing practice" section. I didn't add a column there because 1) the table is already fairly wide, 2) the proposed behavior isn't described until a later section, and 3) the additional column would exactly duplicate the gcc and MSVC columns.

Would it help if I updated the "Proposal" section as follows?

Standardize the behavior exhibited by gcc with fmt and MSVC as demonstrated in the "Existing practice" section:

If not, is there another way I can satisfy what you are looking for?

Tom.

 

Thanks,

 

                 Peter

 

From: SG16 <sg16-bounces@lists.isocpp.org> On Behalf Of Tom Honermann via SG16
Sent: 25 May 2022 05:23
To: SG16 <sg16@lists.isocpp.org>
Cc: Tom Honermann <tom@honermann.net>
Subject: [SG16] Updated draft of P2572R0: std::format() fill character allowances

 

EXTERNAL MAIL

An updated draft of P2572R0 (std::format() fill character allowances) is now available.

The following changes have been made relative to the previously reviewed draft.

• The proposal was modified to align with the outcome of our last review. It now seeks to standardize the behavior exhibited by gcc and MSVC.
• Wording was updated to align with the proposal direction.
• The character examples were extended to include the delete character and a right-to-left directional character (Hebrew letter tav).
• The "Estimated display width restrictions" section was expanded to discuss the option of producing an error only when the requested field alignment cannot be achieved.
• The "Existing practice" section was updated to include character names, example behavior for the delete and Hebrew letter tav characters, and an additional note about handling of right-to-left characters.
• The "Existing practice" section was updated to note that the previously reported buffer overflow that occurs with MSVC 19.31 has been fixed for the next release and to include an example of the diagnostic that is now issued.
• A new "Future considerations and ABI" section has been added.
• The "Implementation experience" section has been updated to note that, now that the paper seeks to standardize the current gcc and MSVC behavior, that the proposal has been implemented (with the exception of some possible ABI break mitigations that may or may not be misguided).

Tom.