Attempt to fix build again

seems if a cpu supports avx2 but not ssse3, the build will fail. but how
could that possible... :<

Attempt to fix build reported fails in containers

Fix possible int overflow on vars used in loop with size_t counter

Update author info

Optimized base64_encode/decode with SIMD instructions

year++

Provide zend_string wrappers for php_base64_(en|de)code

Also pull existing php_base64_decode() proxy out to an inline.
Bump PHP_API_VERSION for ABI change.

Update copyright headers to 2017

Use new param API in standard

Merge branch 'PHP-7.0' into PHP-7.1

* PHP-7.0: (48 commits)
Update NEWs
Unused label
Fixed bug #72853 (stream_set_blocking doesn't work)
fix test
Bug #72663 - p

Merge branch 'PHP-7.0' into PHP-7.1

* PHP-7.0: (48 commits)
Update NEWs
Unused label
Fixed bug #72853 (stream_set_blocking doesn't work)
fix test
Bug #72663 - part 3
Bug #72663 - part 2
Bug #72663 - part 1
Update NEWS
BLock test with memory leak
fix tests
Fix TSRM build
Fix bug #72850 - integer overflow in uuencode
Fixed bug #72849 - integer overflow in urlencode
Fix bug #72848 - integer overflow in quoted_printable_encode caused heap corruption
Fix bug #72838 - Integer overflow lead to heap corruption in sql_regcase
Fix bug #72837 - integer overflow in bzdecompress caused heap corruption
Fix bug #72836 - integer overflow in base64_decode caused heap corruption
Fix for bug #72807 - do not produce strings with negative length
Fix for bug #72790 and bug #72799
Fix bug #72730 - imagegammacorrect allows arbitrary write access
...

Conflicts:
ext/standard/var_unserializer.c

show more ...

Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6: (24 commits)
Update NEWS
BLock test with memory leak
fix tests
Fix TSRM build
Fix bug #72850 - integer overflow i

Merge branch 'PHP-5.6' into PHP-7.0

* PHP-5.6: (24 commits)
Update NEWS
BLock test with memory leak
fix tests
Fix TSRM build
Fix bug #72850 - integer overflow in uuencode
Fixed bug #72849 - integer overflow in urlencode
Fix bug #72848 - integer overflow in quoted_printable_encode caused heap corruption
Fix bug #72838 - Integer overflow lead to heap corruption in sql_regcase
Fix bug #72837 - integer overflow in bzdecompress caused heap corruption
Fix bug #72836 - integer overflow in base64_decode caused heap corruption
Fix for bug #72807 - do not produce strings with negative length
Fix for bug #72790 and bug #72799
Fix bug #72730 - imagegammacorrect allows arbitrary write access
Fix bug#72697 - select_colors write out-of-bounds
Fixed bug #72627: Memory Leakage In exif_process_IFD_in_TIFF
Fix bug #72750: wddx_deserialize null dereference
Fix bug #72771: ftps:// opendir wrapper is vulnerable to protocol downgrade attack
Improve fix for #72663
Fix bug #70436: Use After Free Vulnerability in unserialize()
Fix bug #72749: wddx_deserialize allows illegal memory access
...

Conflicts:
Zend/zend_API.h
ext/bz2/bz2.c
ext/curl/interface.c
ext/ereg/ereg.c
ext/exif/exif.c
ext/gd/gd.c
ext/gd/tests/imagetruecolortopalette_error3.phpt
ext/gd/tests/imagetruecolortopalette_error4.phpt
ext/session/session.c
ext/snmp/snmp.c
ext/standard/base64.c
ext/standard/ftp_fopen_wrapper.c
ext/standard/quot_print.c
ext/standard/url.c
ext/standard/uuencode.c
ext/standard/var.c
ext/standard/var_unserializer.c
ext/standard/var_unserializer.re
ext/wddx/tests/bug72790.phpt
ext/wddx/tests/bug72799.phpt
ext/wddx/wddx.c
sapi/cli/generate_mime_type_map.php

show more ...

Fix TSRM build

Fix bug #72836 - integer overflow in base64_decode caused heap corruption

Make base64_decode() in non-strict mode binary safe

base64_decode: Handle all invalid padding equally

base64_decode: strict: Fail on excessive padding

base64_decode: strict: Fail on truncated input

base64_decode: Avoid code duplication in failures

base64_decode: fix bug #72264 ('VV= =' shouldn't fail in strict mode)

base64_decode: remove redundant code

case 1 is already handled in the first lines of the for loop;
it would only be entered in the invalid case where the string
continues past the de

base64_decode: remove redundant code

case 1 is already handled in the first lines of the for loop;
it would only be entered in the invalid case where the string
continues past the defined length (ch != 0 but length-- == 0).

case 2 and case 3 are redundant, since k >= j and later the
string is truncated to j characters anyway.

show more ...

base64_decode: fix bug #72263 (skips char after padding)

base64_decode: fix bug #72152 (fail on NUL bytes in strict mode)

This added check is actually for NOT failing in NON-strict mode.
The ch == -2 check later causes the desired failure in s

base64_decode: fix bug #72152 (fail on NUL bytes in strict mode)

This added check is actually for NOT failing in NON-strict mode.
The ch == -2 check later causes the desired failure in strict mode.

show more ...

base64_decode: remove redundant check

If length == 0 || *current != '=' is false, the for loop will always
end up in this same point, until the if statement becomes true.
Thus, the i

base64_decode: remove redundant check

If length == 0 || *current != '=' is false, the for loop will always
end up in this same point, until the if statement becomes true.
Thus, the if statement is not needed.

show more ...

base64_decode: reorder to fix out of bounds read

base64_decode: fix bug #72264 ('VV= =' shouldn't fail in strict mode)