On Wed, Nov 2, 2016 at 1:27 AM, William A Rowe Jr <[email protected]> wrote:
> On Tue, Nov 1, 2016 at 10:24 AM, Petr Gajdos <[email protected]> wrote: > >> On Tue, Nov 01, 2016 at 10:12:31AM -0500, William A Rowe Jr wrote: >> > On Tue, Nov 1, 2016 at 8:22 AM, Petr Gajdos <[email protected]> wrote: >> > $ php -r 'var_dump(rawurlencode("~"));' >> > string(1) "~" >> > $ >> > >> > http://php.net/manual/en/function.rawurlencode.php >> > >> > Following patch is perhaps needed: >> > https://build.opensuse.org/package/view_file/Apache:Test/ap >> ache-test/ >> > apache-test-php-rawurlencode-tilde.patch >> > >> > >> > Silly follow-up question, why %7E out of urlencode() a few lines >> further down? >> > >> > The canonical form of '~' is '~', not '%7E', as the character is not >> reserved. >> >> Not that I can answer this question, but urlencode() seem to convert >> it so: >> >> http://php.net/manual/en/function.urlencode.php >> >> $ php -r 'var_dump(urlencode("~"));' >> string(3) "%7E" >> $ php --version >> PHP 7.0.12 (cli) ( NTS ) >> [..] >> > > Fair enough, thanks for confirming. > > Wondering if we don't toggle both and set one as a TODO, but because > all of the tests are evaluated only once, we can't be that selective in > this > particular case. > > Further wondering which of the PHP tests are Apache httpd tests, and > which are actually core PHP tests that should be part of its own suite, > and dropped from our suite entirely. "Fixing" this will lead to further > failures when PHP adopts the current conventions in that second test, > and immediately fail anyone who is still on PHP 3.2. Should we simply > drop this test altogether? > Committed all three patches in r1767800, thanks Petr
