1
0
mirror of https://github.com/php/php-src.git synced 2026-04-19 22:11:12 +02:00
Commit Graph

261 Commits

Author SHA1 Message Date
Nikita Popov
59dfaa3f99 Fix type inference of SEND_UNPACK with empty array
An empty array will not be turned into an array of references.
This violated the invariant than an array has values iff it has
keys.
2019-05-28 16:40:56 +02:00
Christoph M. Becker
e6a191de1b Prevent test case failure
If opcache.log_verbosity_level is greater than 1, opcache will raise
warnings, which will be written to stderr in the default case.  These
warnings are actually to be expected, but would break the test, so we
make sure that the log_verbosity_level is 1 when running this test.
2019-05-21 11:17:51 +02:00
Holly Li (WIPRO LIMITED)
202e6936d6 use {TMP} placeholder in phpt tests 2019-05-21 09:02:37 +02:00
Christoph M. Becker
64ec9f4dbe Fix test wrt. opcache.error_log
Since the test expects log messages to be written to stderr, we have to
make sure that opcache.error_log is set correctly when running the
test.
2019-05-20 11:09:44 +02:00
Christoph M. Becker
e359812ce3 Fix log_verbosity_bug.phpt
This test may fail on Windows due to the file cache fallback.  We
ensure that this will not happen.
2019-05-09 12:15:00 +02:00
Xinchen Hui
e072537e7f Simplfy test case
(I was working on the same bug while nikic committed the similar fix
2019-03-18 19:28:23 +08:00
Nikita Popov
54bf8c820f Fixed bug #77743 2019-03-18 10:49:53 +01:00
Nikita Popov
3f00c9367d Fixed bug #77691
We cannot replace an op1_def opcode with an ASSIGN, if it also has
a used res_def. Usually this doesn't happen because the res_def use
can be eliminated first. The example is a case where operand replacement
on the res_def use fails.
2019-03-04 13:11:12 +01:00
Nikita Popov
2cfb09caa7 Fix inference warning about missing key type 2019-02-22 11:55:16 +01:00
Nikita Popov
17f8b9fb36 Fixed bug #77215
Remove invalid assertion: A block can have multiple switch frees,
so if we don't do live range block splitting, it is not necessarily
true that the free is located at the start of a block.
2018-11-28 20:08:39 +01:00
Nikita Popov
4daa413898 Fixed bug #77092
Weird that this worked for so long, probably because nearly all
ext/standard functions use fast ZPP rather than ordinary ZPP.
2018-11-02 14:29:03 +01:00
Nikita Popov
902ec36710 Merge branch 'PHP-7.1' into PHP-7.2 2018-10-25 16:43:36 +02:00
Nikita Popov
f1ceec5533 Fixed bug #77058
Account for the fact that undef must be interpreted as null for
the purposes of INC/DEC inference.
2018-10-25 16:37:41 +02:00
Anatol Belski
6a5254bc96 Merge branch 'PHP-7.1' into PHP-7.2
* PHP-7.1:
  Fix test when it's run on another drive
2018-10-18 00:44:43 +02:00
Anatol Belski
fda06127fa Fix test when it's run on another drive 2018-10-18 00:43:24 +02:00
Peter Kokot
f1d7e3ca0b Sync leading and final newlines in *.phpt sections
This patch adds missing newlines, trims multiple redundant final
newlines into a single one, and trims redundant leading newlines in all
*.phpt sections.

According to POSIX, a line is a sequence of zero or more non-' <newline>'
characters plus a terminating '<newline>' character. [1] Files should
normally have at least one final newline character.

C89 [2] and later standards [3] mention a final newline:
"A source file that is not empty shall end in a new-line character,
which shall not be immediately preceded by a backslash character."

Although it is not mandatory for all files to have a final newline
fixed, a more consistent and homogeneous approach brings less of commit
differences issues and a better development experience in certain text
editors and IDEs.

[1] http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_206
[2] https://port70.net/~nsz/c/c89/c89-draft.html#2.1.1.2
[3] https://port70.net/~nsz/c/c99/n1256.html#5.1.1.2
2018-10-15 04:31:31 +02:00
Peter Kokot
113213f027 Sync leading and final newlines in *.phpt sections
This patch adds missing newlines, trims multiple redundant final
newlines into a single one, and trims redundant leading newlines in all
*.phpt sections.

According to POSIX, a line is a sequence of zero or more non-' <newline>'
characters plus a terminating '<newline>' character. [1] Files should
normally have at least one final newline character.

C89 [2] and later standards [3] mention a final newline:
"A source file that is not empty shall end in a new-line character,
which shall not be immediately preceded by a backslash character."

Although it is not mandatory for all files to have a final newline
fixed, a more consistent and homogeneous approach brings less of commit
differences issues and a better development experience in certain text
editors and IDEs.

[1] http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_206
[2] https://port70.net/~nsz/c/c89/c89-draft.html#2.1.1.2
[3] https://port70.net/~nsz/c/c99/n1256.html#5.1.1.2
2018-10-15 04:29:24 +02:00
Peter Kokot
17ccbeec32 Trim trailing whitespace in *.phpt 2018-10-14 19:44:14 +02:00
Peter Kokot
7af945e271 Trim trailing whitespace in *.phpt 2018-10-14 19:43:12 +02:00
Peter Kokot
60a69daec6 Sync leading and final newlines in source code files
This patch adds missing newlines, trims multiple redundant final
newlines into a single one, and trims redundant leading newlines.

According to POSIX, a line is a sequence of zero or more non-' <newline>'
characters plus a terminating '<newline>' character. [1] Files should
normally have at least one final newline character.

C89 [2] and later standards [3] mention a final newline:
"A source file that is not empty shall end in a new-line character,
which shall not be immediately preceded by a backslash character."

Although it is not mandatory for all files to have a final newline
fixed, a more consistent and homogeneous approach brings less of commit
differences issues and a better development experience in certain text
editors and IDEs.

[1] http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_206
[2] https://port70.net/~nsz/c/c89/c89-draft.html#2.1.1.2
[3] https://port70.net/~nsz/c/c99/n1256.html#5.1.1.2
2018-10-14 12:54:08 +02:00
Peter Kokot
03f3b8479b Sync leading and final newlines in source code files
This patch adds missing newlines, trims multiple redundant final
newlines into a single one, and trims redundant leading newlines.

According to POSIX, a line is a sequence of zero or more non-' <newline>'
characters plus a terminating '<newline>' character. [1] Files should
normally have at least one final newline character.

C89 [2] and later standards [3] mention a final newline:
"A source file that is not empty shall end in a new-line character,
which shall not be immediately preceded by a backslash character."

Although it is not mandatory for all files to have a final newline
fixed, a more consistent and homogeneous approach brings less of commit
differences issues and a better development experience in certain text
editors and IDEs.

[1] http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_206
[2] https://port70.net/~nsz/c/c89/c89-draft.html#2.1.1.2
[3] https://port70.net/~nsz/c/c99/n1256.html#5.1.1.2
2018-10-14 12:51:01 +02:00
Nikita Popov
cfdd828163 Fixed bug #76796 2018-09-07 14:02:03 +02:00
Xinchen Hui
f31ba7cb53 Fixed bug #76477 (Opcache causes empty return value) 2018-06-15 16:31:31 +08:00
Anatol Belski
f7b431fb30 Merge branch 'PHP-7.1' into PHP-7.2
* PHP-7.1:
  Fix Opcache test fails regarding to AppVeyor image update
2018-05-18 18:37:53 +02:00
Anatol Belski
6043a2d6f7 Fix Opcache test fails regarding to AppVeyor image update 2018-05-18 18:36:39 +02:00
Nikita Popov
4c083e7a66 Merge branch 'PHP-7.1' into PHP-7.2 2018-04-27 22:23:55 +02:00
Nikita Popov
6738d19fb8 Fix bug #76281
Make sure we keep the smart-branch inhibiting NOP even if there
are multiple NOPs in sequence.
2018-04-27 22:21:59 +02:00
Nikita Popov
2c602be7c8 Merge branch 'PHP-7.1' into PHP-7.2 2018-04-27 17:02:01 +02:00
Nikita Popov
279ba58edb Fixed bug #76275
* Adjust IS_SERIALIZED() check to account for potential empty
  allocations at the end of the memory region.
* Deallocate empty allocation if all try/catch elements have been
  removed in the block pass (similar to what we do if all live
  ranges have been eliminated).
2018-04-27 17:01:35 +02:00
Nikita Popov
3306577797 Followup fix for bug #75729
The case of overloaded substr() with 3 arguments was not handled.
2018-03-24 13:15:59 +01:00
Xinchen Hui
4758164ac5 Fixed bug #76094 (Access violation when using opcache) 2018-03-15 11:17:00 +08:00
Bob Weinand
cfd0979be5 Merge branch 'PHP-7.1' into PHP-7.2 2018-03-10 15:21:16 +01:00
Bob Weinand
9c6df8a238 Fix bug #76074 (opcache corrupts variable in for-loop) 2018-03-10 15:20:45 +01:00
Nikita Popov
39974dd65d Disable negative range inference
The negative range inference implementation does not work correctly,
and it's not clear right now how it can be fixed. As such, disable
it entirely for now.
2018-02-18 15:33:28 +01:00
Nikita Popov
0674d677bc Merge branch 'PHP-7.1' into PHP-7.2 2018-02-16 20:30:49 +01:00
Nikita Popov
372bf8a923 Fixed bug #75969
Move NOP stripping out of zend_optimize_block: NOP stripping may
move instructions, which may invalidate a Tsource shared across
an extended basic block.
2018-02-16 20:30:03 +01:00
Anatol Belski
2235da59af Extend skipif check 2018-02-13 14:04:51 +01:00
Nikita Popov
7aac61ce76 Fixed bug #75938
New modulus range inference implementation has been verified using
https://gist.github.com/nikic/67947ff92cf0e1f7e931f2f0d4cf817f.

The computed bounds are not tight, but it seems to be very hard to
compute tight bounds on modulus operations.
2018-02-09 15:10:23 +01:00
Nikita Popov
1391a0fa0b Fixed bug #75893
It is not sufficient to just add the additional types for aliased
variables at the end of type inference, because types of derived
variables may depend on them. Make sure the additional types are
always added whenever the type of an aliased variable is updated.
2018-02-05 21:41:44 +01:00
Nikita Popov
d9e71169e6 Fixed bug #75729
Don't try to evaluate substr() and strpos() at compile-time if
mbstring function overloading is enabled. I'm reusing the
COMPILE_NO_BUILTIN_STRLEN for the detection of mbstring function
overloading here, even if it's technically only about strlen()...
2018-02-05 21:08:18 +01:00
Nikita Popov
ee45104dba Fixed bug #75698 2017-12-16 22:04:18 +01:00
Nikita Popov
19d0c5dc9f Add test for bug #75681
The issue itself has been fixed by 8a4532319d.
2017-12-16 21:24:26 +01:00
Xinchen Hui
a20c9bd3fe Fixed bug #75687 (var 8 (TMP) has array key type but not value type) 2017-12-16 23:59:10 +08:00
Dmitry Stogov
3503dc7482 Merge branch 'PHP-7.1' into PHP-7.2
* PHP-7.1:
  Fixed bug #75608 ("Narrowing occurred during type inference" error)
2017-12-04 17:22:28 +03:00
Dmitry Stogov
5934bff913 Fixed bug #75608 ("Narrowing occurred during type inference" error) 2017-12-04 17:22:06 +03:00
Xinchen Hui
e02c2d2dfb Merge branch 'PHP-7.1' into PHP-7.2
* PHP-7.1:
  Update NEWs (added bug entry)
2017-11-28 13:16:22 +08:00
Xinchen Hui
ee9e32caca Update NEWs (added bug entry) 2017-11-28 13:15:54 +08:00
Dmitry Stogov
a4b2ae69e9 Merge branch 'PHP-7.1' into PHP-7.2
* PHP-7.1:
  Fixed bug in type inference
2017-11-27 10:16:46 +03:00
Dmitry Stogov
776bef8447 Fixed bug in type inference 2017-11-27 10:16:16 +03:00
Xinchen Hui
c412b8b384 Fixed bug #75556 (Invalid opcode 138/1/1) 2017-11-25 00:54:15 +08:00