[Bug math/28327] New: Wrong return type for {,u}fromfp{,x} functions

[Bug math/28327] New: Wrong return type for {,u}fromfp{,x} functions

[schwab@linux-m68k.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28327

            Bug ID: 28327
           Summary: Wrong return type for {,u}fromfp{,x} functions
           Product: glibc
           Version: 2.34
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: schwab@linux-m68k.org
  Target Milestone: ---

The {,u}fromfp{,x} functions are declared to return {,u}intmax_t, but that
doesn't make sense.  Even though the synopsis in 7.12.9.10/11 say they return
{,u}intmax_t, the descriptions actually talks about returning integer-valued
floating point values or NaN or the first argument.  Section B.11 actually
lists the correct return types.  It looks like N2548 has only been partially
applied to N2596.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug math/28327] Wrong return type for {,u}fromfp{,x} functions

[joseph at codesourcery dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28327

--- Comment #1 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
The types currently follow TS 18661-1.  As I said in the commit message 
for commit 858045ad1c5ac1682288bbcb3676632b97a21ddf, my inclination is to 
treat these like the change of totalorder type (new symbol versions etc. 
for the ABI change; old versions become compat symbols and are no longer 
supported as an API).  That will probably mean completely separate 
implementations for the new return types, rather than one implementation 
calling the other, as well as completely separate testcases (though with 
some overlap in the test inputs).

I reported the incomplete application of N2548 as issue 78 in the C 
standard GitLab in May.

-- 
You are receiving this mail because:
You are on the CC list for the bug.