Skip to content
Snippets Groups Projects
Select Git revision
  • benchmark-tools
  • postgres-lambda
  • master default
  • REL9_4_25
  • REL9_5_20
  • REL9_6_16
  • REL_10_11
  • REL_11_6
  • REL_12_1
  • REL_12_0
  • REL_12_RC1
  • REL_12_BETA4
  • REL9_4_24
  • REL9_5_19
  • REL9_6_15
  • REL_10_10
  • REL_11_5
  • REL_12_BETA3
  • REL9_4_23
  • REL9_5_18
  • REL9_6_14
  • REL_10_9
  • REL_11_4
23 results

c-library.m4

Blame
    • Tom Lane's avatar
      b929614f
      Remove configure's check for nonstandard "long long" printf modifiers. · b929614f
      Tom Lane authored
      We used to claim to support platforms using 'q' or 'I64' as the printf
      length modifier for long long int, by dint of replacing snprintf with
      our own code which uses the C99 standard 'll' modifier.  But that is
      only adequate if we use INT64_MODIFIER only in snprintf-based calls,
      not directly with the platform's native printf or fprintf.  Which
      hasn't been the case for years.  We had not noticed, partially because
      of inadequate test coverage, and partially because the buildfarm is
      almost completely bare of machines that won't take 'll'.  The last
      one seems to have been frogmouth, which was adjusted recently so that
      it will take 'll'.  We might as well just give up on the pretense
      that anything else works, and save ourselves some configure cycles.
      
      Discussion: https://postgr.es/m/13103.1526749980@sss.pgh.pa.us
      Discussion: https://postgr.es/m/24769.1526772680@sss.pgh.pa.us
      b929614f
      History
      Remove configure's check for nonstandard "long long" printf modifiers.
      Tom Lane authored
      We used to claim to support platforms using 'q' or 'I64' as the printf
      length modifier for long long int, by dint of replacing snprintf with
      our own code which uses the C99 standard 'll' modifier.  But that is
      only adequate if we use INT64_MODIFIER only in snprintf-based calls,
      not directly with the platform's native printf or fprintf.  Which
      hasn't been the case for years.  We had not noticed, partially because
      of inadequate test coverage, and partially because the buildfarm is
      almost completely bare of machines that won't take 'll'.  The last
      one seems to have been frogmouth, which was adjusted recently so that
      it will take 'll'.  We might as well just give up on the pretense
      that anything else works, and save ourselves some configure cycles.
      
      Discussion: https://postgr.es/m/13103.1526749980@sss.pgh.pa.us
      Discussion: https://postgr.es/m/24769.1526772680@sss.pgh.pa.us