On 7/13/26 09:06, Rune Torgersen via tz wrote:
I do still get the overflow warings, not sure if those are anything to worry about ot not.
I think you're OK, as ancient GCC's preprocessor incorrectly warned about overflow but computed the correct value anyway.
For what it's worth, I reproduced the diagnostic bug on 32-bit SPARC with GCC 3.4.3 (csl-sol210-3_4-branch+sol_rpath), which is still sort-of-supported by Oracle, and verified that the generated code is OK. So it seems it's merely a false alarm, and we don't need to change tzcode because of this issue.
PS. I assume you know that old system will stop working in January 2038 when 32-bit signed time_t values overflow. (Perhaps that is why Oracle is promising support for Solaris only through 2037.)
