fixed date not allocating enough buffer, bug 7963

Fix date('r') overflow
@Fix date('r') overflow

Namespace protect parse_date interface and implementation

RFC822 does specify two digit years, but practically every decent mail
client I could check is using 4-digit years these days.

some changes, and added 'r' flag which generates an RFC822 compatible date string.

@- Fixed a bug in checkdate() which caused < 1 years to be valid (Jani)
# Bug #6254

Fixed bug #6409.
# And propably numerous other..? Either I am wrong or the documentation is.

Cleaning up some mess

protect datetime() against all zero inputs ...

Use dashes and short day name in cookies since some browsers seem picky
about this
@ Use dashes and short day name in cookies since some browsers seem picky
@ about this (Rasmus)

(PHP date) Added new 'O' format modifier for printing out the GMT Offset in
the same manner as emails have in their Date: header. The format
is similar to [+-]HHMM of o

(PHP date) Added new 'O' format modifier for printing out the GMT Offset in
the same manner as emails have in their Date: header. The format
is similar to [+-]HHMM of offset.
@- Added new 'O' format modifier that will output the GMT offset as "[+-]HHMM"
@ (eg: Pacific time is -0700). This is useful for things such as Date: mail
@ headers.
# Um, this should have existed LONG time ago... much better than that gmt
# offset in seconds modifier :)

show more ...

Now its a todo.

Use hardcoded 3600 seconds offset instead of altzone which may not be present
on all systems.

@- Updated strtotime() to handle many more formats. It now has complete
@ feature parity with GNU date command. (Andrei)

return FALSE from strftime on negative timestamp values
they do not make sense anyway and seem to crash win2k

Refine checkdate() fix to accept numeric strings and floating point as well

Require the year arg to checkdate to be an integer - fix for bug #5034
@Require the year arg to checkdate to be an integer (Rasmus)

@- Fixed 'Z' flag in date() to adjust for daylight savings time. (Andrei)

- ARG_COUNT(ht) -> ZEND_NUM_ARGS() mega patch

a second (cleaner?) try on warnings about unsupported functions

- I'm not sure if PHP_NOT_IN_THIS_BUILD is the way to handle these
- functions but in any case I have changed it to PHP_NOT_IN_THIS_BUILD()

if a module provides a function it should also do so if the function
is not functional due to configure findings
the function should offer a meaningful warning if it is not supported
inst

if a module provides a function it should also do so if the function
is not functional due to configure findings
the function should offer a meaningful warning if it is not supported
instead of just beeing undefined
i had already changed this in 'gd', now this is doing it for 'standard'

show more ...

Update the license with the new clause 6

Change reentrancy API to always use the php prefix.

Check for the declaration of reentrant functions, so that we can use
them in non-ZTS mode on all platforms.

Sterling, only one line for the short descritption, even it is massiv.