There's no reason to parse "http_status_line"; the PHP interpreter already
does this. If the line contains a valid status code, it's assigned to
"http_response_code".
This also fixes invalid status line handling, where the nxt_int_parse()
function returned -1; it was cast to unsigned, yielding response code 65535.
Now index is always initialized for create operations.
The changes in nxt_conf_op_compile() simplify adding upcoming support of
operations with arrays.
No functional changes.
According to CGI/1.1 RFC 3875:
The server MUST set this variable; if the Script-URI does not include a
query component, the QUERY_STRING MUST be defined as an empty string ("").
Python's PEP 333(3) allows omitting it in WSGI interface; PHP docs force no
requirements; PSGI and Rack specifications require it even if empty.
When nginx proxies requests over FastCGI, it always provides QUERY_STRING.
and some PHP apps have been observed to fail if it is missing (see issue
#201 on GitHub).
A drawback of this change (besides a small overhead) is that there will be
no easy way to tell a missing query string from an empty one (i.e. requests
with or without the "?" character); yet, it's negligible compared to the
possible benefits of wider application compatibility.
This closes#226 issue on GitHub.
While it looks nicer without zero 3-rd version number, this should
improve interoperability. Version string can be parsed or used for
sorting. And it is easier to handle and less confusing when there
is constant number of version parts.
Moreover, NPM also expects version format with 3 parts.
So ".0" has already been used in Node.js module version.
Previously, the nxt_router_prepare_msg() function expected server host among
other headers unmodified. It's not true anymore since normalization of the
Host header has been introduced in 77aad2c142a0.
The nxt_unit_split_host() function was removed. It didn't work correctly with
IPv6 literals. Anyway, after 77aad2c142a0 the port splitting is done in router
while Host header processing.
In order to reduce number of operations over rb-tree and process them in
batches simultaneously, all the timers changes are temporary stored in array.
While processing of these changes, the same memory is also used for storing
pointers to postpone timers adding.
As the same block of memory has been referenced by two different types of
pointers (nxt_timer_change_t * and nxt_timer_t **), some compilers may reorder
operations with these pointers and produce broken code. See ticket #221 on
GitHub for a particular case.
Now the same "nxt_timer_change_t" structure is used in both cases.
Also, reverted the -fno-strict-aliasing flag, which has been introduced in
ef76227ec159 as a workaround for this issue.
It doesn't do anything useful, among creating a JSON message and logging it
to debug log. Besides that it causes segmentation fault if the RPC handler
is triggered with an empty buffer due to exiting of the main process.
PyErr_Print() writes traceback to "sys.stderr", which is a file object that
can buffer the output. If the process exits immediately, the buffer can be
destroyed before flushing to the log. As a result, the user doesn't see
the traceback.
Now Py_Finalize() is also called in case of any errors during initialization.
It finalizes the interpreter and flushes all data.
By default "npm install" switches to non-privileged user to run package scripts
if it is invoked by root. As a result it may prevent node-gyp from writing to
package directory and break installation of the module.
To disable this switching the --unsafe-perm flag is added.
In most cases it is not needed because Ruby libraries are in the default path.
At the same time, rpath pointing to the default path is prohibited by rpmbuild
on Fedora.
This is related to issue #87 on GitHub.
Since PHP 7, a zend_signal_startup() call is required if the interpreter
was built with ZEND_SIGNALS defined; such a call was added in 3fd76e4ce70a.
However, the zend_signal_startup() export is missing from the PHP library;
as the result, dlopen() fails with the 'Undefined symbol "zend_signal_startup"'
error while loading the PHP module.
Meanwhile, if PHP is built without ZTS, the zend_signal_startup() call can
be omitted; otherwise, the missing call causes segmentation fault.
The PHP fix already was committed to upstream, but we still have to deal
with numerous unpatched versions remaining at large.
See the related PHP bug: https://bugs.php.net/bug.php?id=71041
The nxt_murmur_hash2() generated 4-byte hash that was stored in uintptr_t,
which was 8 bytes long on 64-bit systems. At each iteration, it took the
previous key and hashed it again.
The problem was that it took only the first 4 bytes of the key, and these
4 bytes were always zero on 64-bit big-endian system. That resulted in
equal keys at each iteration.
The bug was discovered on IBM/S390x.
The nxt_conf_map_object() function used nxt_int_t for NXT_CONF_MAP_INT, which
was 8 bytes long on 64-bit systems.
But the nxt_port_main_start_worker_handler() used it to map into the int field
of the nxt_common_app_conf_t structure, which was 4 bytes. As the result, on
a 64-bit big-endian system all the meaningful module type numbers were assigned
into the gap above the "type" field.
The bug was discovered on IBM/S390x.
The issue was caused by misplacement of allocated blocks in rbtree due
to broken comparison function if the distance between two allocations
did not fit into intptr_t. As the result, nxt_mp_free() could have
failed to find the allocation.
In particular, it was mostly observed when Unit was compiled with
musl C library on 32-bits systems.
This closes#118 issue on GitHub.
Timers that don't require maximum precision (most of them, actually) can be
triggered earlier or later within the bias interval.
To reduce wakeups by timers, the expire function now triggers not only all
timers that fall within the elapsed time, but also those whose bias falls
within this interval.
Delete/insert operation complexity for a red-black tree is O(log n),
where n is the total number of tree elements.
If all delete operations are performed before all insert operations,
the average number of tree elements during an operation will be
lower than in the mixed-operations case.
To optimize rbtree operations, all changes are stored in array and later
processed in batches.
The previous implementation of this mechanics had a number of design flaws.
Each change was saved in a new array entry; until the changes were applied,
the timer remained in an intermediate state (NXT_TIMER_CHANGING).
This intermediate state didn't allow to identify if time was going to be
disabled or enabled. However, the nxt_conn_io_read() function relied on
this information; as a result, in some cases the read timeout wasn't set.
Also, the nxt_timer_delete() function did not reliably track whether a timer
was added to the work queue. It checked the NXT_TIMER_ENQUEUED state of
a timer, but this state could be reset to NXT_TIMER_DISABLED by a
nxt_timer_disable() call or another nxt_timer_delete() call.
Now, instead of keeping the whole history of the timer's changes, the new
implementation updates the timer state immediately, and only one operation
is added to the array to add or delete timer in the rbtree according
to its final state.
Previously, passing 0 resulted in reading the whole body and all negative
values raised an exception.
Now the behaviour is in consistentance with io.RawIOBase.read() interface,
and passing 0 returns empty (byte) string, while -1 results in reading the
whole body.
Overriding LANG might not work, since it has less precedence than LC_* settings.
LC_ALL has the highest precedence.
This should resolve issue #121 on GitHub.
It has the following structure and default values:
{
"http": {
"header_read_timeout": 30,
"body_read_timeout": 30,
"send_timeout": 30,
"idle_timeout": 180,
"max_body_size": 8388608
}
}
The implementation of module was based on the assumption that PHP reads request
body and headers in the particular order. For the POST request the body goes
before headers and vice versa for all other requests.
But as it appeared later, this order is unspecified and depends on many factors,
including the particular code of PHP application. Among other factors those
can affect ordering:
- presence of "Content-Type" header;
- "variables_order" php.ini setting;
- "enable_post_data_reading" php.ini setting;
- reading php://input by application;
and this list can be incomplete.
As a temporary workaround, request body now is always put before headers and it
is gracefully skipped whenever PHP wants to get headers.
This closes#144 issue on GitHub.
Valentin Bartenev