| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Profiling was done using Firebug in Firefox 31 on Linux x86_64. Test was
performed as follows:
- rrd generated using collectd with default rrdtool settings.
- CPU utilization was visualized with 8 lines (idle, nice, etc.).
- The data has gaps between 20% and 40% and 55% and 85% (were no data
was collected).
- The measured performance was from advancing the data 40% to the right
(there are no gaps before that).
The total running time is 18859.094ms (old), 5151.822ms (new). In the
below profiles, common functions with similar running times and call
counts were omitted. Columns in the profile table: function, call count,
time spent in the function, average time spent per call in *total*
(including other functions called by the function) and percentage of the
total running time. The RrdGraph.data_calc function was not called much,
but seems to be heavy enough to include in the trace.
Profile for old implementation:
- getEndianByteAt 246959 12200.42ms 0.063ms 64.69%
- getDoubleAt 30306 3183.393ms 0.612ms 16.88%
- RRDRRA.getEl 30306 1177.33ms 0.665ms 6.24%
- RRDRRA.calc_idx 30306 449.201ms 0.015ms 2.38%
- (RrdDataFile.build 47 431.986ms total=21189.434ms 2.29%)
- getByteAt 254290 327.369ms 0.013ms 1.74%
- (2 functions omitted)
- getCStringAt 752 103.308ms 0.271ms 0.55%
- RrdRpn.calc 5504 98.55ms 0.018ms 0.52%
- (1 function omitted)
- (RrdGraph.data_calc 1 75.417ms total=174.535ms 0.4%)
- getLongAt 1128 65.796ms 0.212ms 0.35%
Profile for new implementation:
- RRDRRA.getEl 32280 1202.446ms 0.065ms 23.34%
- RrdRpn.calc 43968 848.693ms 0.019ms 16.47%
(probably includes the getByteAt call which is not visible in trace)
- (RrdGraph.data calc 2 616.648ms total=1466.432ms 11.97%)
- getDoubleAt 32280 460.894ms 0.014ms 8.95%
- (RrdDataFile.build 48 438.1ms total=2719.07ms 8.5%)
- (RRDRRA.calc_idx 32280 436.375ms total=436.375ms 8.47%)
- (10 functions omitted)
- getLongAt 1152 16.273ms 0.014ms 0.32%
- (2 functions omitted)
- getCStringAt 768 13.651ms 0.018ms 0.26%
A second test was performed on a graph with no visible data points. I
dragged it to the left to get in the future with only NaN values. Then
the profile started. It turns out that the byte readings are not
dominating here, 99% of the time was spent in RrdRpn.calc and
RrdGraph.data_calc:
- (old) RrdRpn.calc 3.14k 58493s 0.019ms
- (new) RrdRpn.calc 3.10k 56912s 0.018ms
- RrdGraph.data_calc (n=2) took about 46 seconds for both.
- getByteAt (n=12096): 0.014ms (old), invisible/inlined (new)
- getEndianByteAt (n=4608): 0.040ms (old only)
- getCStringAt (n=768): 0.285ms (old), 0.018ms (new)
- getLongAt (n=1152): 0.233ms (old), 0.014ms (new)
|
|
|
|
|
|
|
|
|
|
|
|
| |
Rename switch_endian to littleEndian to be more explicit (the meaning
is now inversed too). Make getEndianByteAt private since it is an
implementation detail.
Drop getStringAt as this is not used in this library. Move getCharAt
above getCStringAt since they are more related.
Move the methods definition for the current implementation to a new
function.
|
|
|
|
|
| |
These parameters are not used in this library. Should it ever be
necessary to process a slice of data, then the caller should do so.
|
|
|
|
|
| |
Use conditional compilation for the antique browsers that provide
VBScript support.
|
|
|
|
|
|
|
|
|
|
|
| |
commit 2a74a333ff143499b465234f0395a4aad7bdaa78 ("Upgrade to
javascriptrrd ver 1.1.1") converted getByteAt calls to a getEndianByteAt
call, but this also introduced a branch operation. Since endianess is
dependent on the file, we can move the switch_endian check outside the
function.
Performance improved from 0.148ms (n=336k) to .039ms (n=384k) which
translates to about 30 seconds!
|
|
|
|
|
|
|
|
|
|
|
| |
Added jshint comment on top, added "use strict". Added missing
semi-colons, converted to '===' (after validation).
Added comment for `typeof strData == "unknown"`. Break one long line
over two lines.
There are no other changes except for whitespace and line ending
changes. Check with `git diff -w`.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Many fall in the category missing semicolon, but there are legitimate
bugs (like throwing an error with an undefined variable, using
isInfinite instead of !isFinite or fabs instead of Math.abs).
At some places, I moved the variable declarations to avoid duplicate
definition warnings.
Redundant breaks have been removed (after return / throw).
Global variables were implicitly defined in RrdDataFile (which caught my
attention) and Base64, these have been made local.
Also fixed some whitespace errors. Yay, the consistency. Not all (style)
issues are fixed.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
Also fix some typos and white-space issues. Reported at
https://github.com/pommi/CGP/issues/88
|
|
|
|
| |
Closes #92
|
| |
|
|
|
|
| |
Closes #89
|
|
|
| |
Fix a link when no class selected.
|
| |
|
|
|
|
|
|
|
| |
detail.php
In detail.php $_GET['x'] and $_GET['y'] are overridden. filter_input
takes the original values. This results in NULL values.
|
| |
|
|
|
|
|
|
|
|
|
| |
Report 400 Bad Request on query errors instead of reporting 200 OK
(which can be cached).
Add some additional validation for the 'x' and 'y' parameters, to catch
underflow (test with `x=-10` for example). Also fix a typo in the error
message and include more details (the actual error).
|
|
|
|
|
|
|
|
|
| |
`[\w\W]` matches EVERYTHING. Validation failed. Also remove `\d` as that
`\w` is a superset of it.
Also check that the host parameter is actually non-empty, otherwise
`collectd_plugins` will happily return results as the directory
`$CONFIG['datadir'] . '/' . ''` trivially exists...
|
|
|
|
|
| |
If a datadir would contain characters with a special meaning in regex
(e.g. `/tmp/logs (new)`), then the function would blow up.
|
|
|
|
| |
... for instance, when using an overly large "s" parameter.
|
|
|
|
| |
Also init info_array in case the output is empty.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously, a command is built by string concatenation. Here, the
distinction between a value and multiple params got lost. Solve this
by using an array for shell arguments. As the escaping is now removed
from the `rrd_gen_graph` function, the canvas style needs to manually
add those quotes to make the JS code still work. That only supports
double-quotes, so hopefully nobody creates a name with a double quote
as that would break the fragile JS command line parser.
Separate the rrdtool options from the rrdtool graph command to make the
`$graph_type == 'canvas'` option work (it would otherwise not understand
the `rrdtool graph - -a PNG` option).
Merge the SVG and PNG cases as they are the same except for the
Content-Type header.
Fix a missing html escape in a debug style.
|
|
|
|
|
|
|
| |
This is a security bug. The contents of the file are not
immediately disclosed, but it could disclose environment information
to the attacker (when display_errors=1 and an ancient PHP version
is used that does not check for '\0' in `file_exists()`).
|
|
|
|
|
|
|
|
|
| |
index.php expects that collectd_hosts() always return an array for
array_diff. Since an empty array evaluates to FALSE anyway, do not
change the type and just return the empty array.
In base.php, if no files are returned, the loop is never called and
three variables are not initialized. Do it now.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Make build_url return an URL, not HTML. This separates presentation
from data.
plugin_header's return value is unused, remove the unnecessary return.
At places where `printf("<tag att='%s'>", $x);` is used, it is now
converted to `printf("<tag att=\"%s\">", htmlentities($x));` since
the single quote is not escaped by default by htmlentities.
In case the canvas style is used, JS should use `textContent` instead
of `innerHTML` to avoid reading `"` instead of `"`. Nobody (should)
use(s) IE6 anymore, so it is a safe change.
While at it, use the standard charset attribute of meta to specify
the character set (UTF-8).
|
|
|
|
| |
Closes #86
|
|
|
|
| |
Closes #73
|
|
|
|
|
|
| |
The value of HRULE could be a reference to for example a VDEF variable.
Closes #85
|
| |
|
|
|
|
|
|
|
| |
A malicious user can enter javascript code in "h" GET parameter and this
code gets executed because this parameter is copied in the HTML output
without any escaping. This patch escape the output before it is checked
it is valid.
|
| |
|
| |
|
|
|
|
|
| |
Forgotten in commit:
v0.3-77-gbf56a64 [remove ajax js code]
|
|\ |
|
| | |
|
|/ |
|
|
|
|
|
|
|
| |
array_merge resets the index, while in case of $this->colors and
$this->legend, we'd like to keep the index. An irq plugin sample:
array(1 => 1, 'ERR' => 'ERR') must not become
array(0 => 1, 'ERR' => 'ERR') after an array_merge.
|
|\ |
|
| | |
|
|\ \ |
|
| |/ |
|
|/
|
|
|
|
|
| |
Set $CONFIG['percentile'] to a number and a XXth percentile line + legend
will be added to all network traffic (if_octets) graphs.
Closes #81
|
|
|
|
| |
Closes #78
|
|
|
|
|
|
|
| |
When type wasn't set in the plugin json, a "PHP Notice: Undefined index"
would pop up. Just default to 'default' when type isn't set.
Closes #76 #77
|