Bruce Momjian wrote:
>
>
>> This broke the buildfarm and finally explains the following kluge which
>> has been puzzling me for four years:
>>
>> /*
>> * for some reason MinGW (and MSVC) outputs an extra newline, so
>> this
>> * suppresses it
>> */
>> #ifndef WIN32
>> fputc('\n', fout);
>> #endif
>>
>> I have removed the kluge (and yes, I tested it).
>>
>
> Oh, that kluge. Why did the isatty() addition fix this? Was the pager
> being used on Win32 for the regression tests and somehow eating a line
> or something?
>
It apparently produced an extra line which we had compensated for with
the kluge (without really understanding why we had to).
Anyway, all is good now, as the buildfarm shows.
cheers
andrew