[Privoxy-devel] how costly is --enable-strptime-sanity-checks ?

[Lee]

On 5/8/16, Fabian Keil <fk at fabiankeil.de> wrote:
> Lee <ler762 at gmail.com> wrote:
>
>> On 5/8/16, Fabian Keil <fk at fabiankeil.de> wrote:
>> > Lee <ler762 at gmail.com> wrote:
>> >
>> >> On 5/6/16, Fabian Keil <fk at fabiankeil.de> wrote:
>> >> > Fabian Keil <fk at fabiankeil.de> wrote:
>> >> >
>> >> >> Lee <ler762 at gmail.com> wrote:
>
>> >> >> > So why isn't there a 'year 2038' problem with gcc -mno-cygwin?
>> >> >>
>> >> >> Maybe there is a problem and the sanity checks just don't
>> >> >> detect it.
>> >> >
>> >> > Additionally most of the problematic time-related tests are
>> >> > level 17 tests and should thus be skipped on native windows
>> >> > builds due to:
>> >> >
>> >> > # Level 17 needs = feature status FEATURE_PTHREAD Yes
>> >> >
>> >> > Of course FEATURE_PTHREAD does not actually guarantee a 64
>> >> > bit time_t, so maybe we should add a FEATURE_64_BIT_TIME_T
>> >> > to make the test results easier to compare.
>> >>
>> >> That sounds good.  It looks like the standard way to check the size is
>> >>   AC_CHECK_SIZEOF(time_t)
>> >> which, on my machine, adds
>> >>
>> >> /* The size of `time_t', as computed by sizeof. */
>> >> #define SIZEOF_TIME_T 4
>> >>
>> >> to config.h
>> >> dunno how to get from there to FEATURE_64_BIT_TIME_T tho :(
>> >
>> > FEATURE_64_BIT_TIME_T is available in CVS now.
>>
>> I'll give it a try.  Expected difference in behavior now is what?
>
> As of regression-tests.action,v 1.60, level 17 tests should be
> skipped if FEATURE_64_BIT_TIME_T is unavailable.

What about this one?

2016-05-08 09:44:42: Ooops. Got: 'REMOVAL' while expecting: SOME CHANGE
2016-05-08 09:44:42: Failure for test 641. Header 'If-Modified-Since:
Wed, 31 Dec 1969 23:59:59 GMT' and tag 'hide-if-modified-since{+60}'
2016-05-08 09:45:06: Executed 822 regression tests. Skipped 83. 821
successes, 1 failures.

and the
Error: Failed to parse 'Thursday Jan 01 00:00:00 123456789 GMT' using
'%A %b %d %H:%M:%S %Y'. Moving on.

Thanks,
Lee