On Fri, Mar 28, 2025 at 9:14 AM David Anderson <dave...@gmail.com> wrote:
> On 3/28/25 08:30, David Blaikie wrote: > > I believe the intent is that version numbers are able to be compared > > numerically. > > > > In any case, they are numbers per https://dwarfstd.org/ > > issues/210419.1.html <https://dwarfstd.org/issues/210419.1.html> - "A > > DW_AT_language_version attribute may be specified whose constant value > > is an integer code indicating the version of the source language." > > > > My point is that attempting to use an integer is a bad fit to semantic > versioning and breaks down in real cases. > > I suppose the attraction of an integer is that the integer will be fewer > bytes than an ascii representation. > It also allows some amount of version comparisons without decomposing/parsing the version number, etc. This, at least, is how C++'s number is defined/works - makes it easier to check for "do I have at least C++17". Which is one of the benefits of the DWARFv6 codes in general - a consumer can check "is this any version of C++" and "is this at least version X of C++", etc. > > The chances of compiler versions violating the default > field size assumptions seem pretty small, I admit. > > Never mind. > DavidA >
-- Dwarf-discuss mailing list Dwarf-discuss@lists.dwarfstd.org https://lists.dwarfstd.org/mailman/listinfo/dwarf-discuss
