Is the Ironpath running successfully without any problems in any web environment? I am running on some issues.
The first issue, I'm not really running any IronPython-specific scripts, I'm implementing the library so that I can get server side syntax highlighting. The library is approximately 20+ files.
In addition to the fact that the latest Ironpathion release can not be compiled in the DLL (due to), I am successfully running it by copying all the files and dependencies in my bin folder.
The problem is, I saw how my w3wp.exe process is doing while highlighting and have seen some shortstart problems:
-
Even here Under Cassini, in a completely original, blank web site, select the code from the
select * table. Every time I execute it using SQL lexer (refreshes the page) ... I'm explicitly shutting down the engine and in a function call
Lightweight scope . It starts about 30 MB and with 150 refresh it is 150 MB or more. In my real web application, my app pool increases by using SQL lexer (the same code) about 200 MB / s (literally, when I reach 1 GB, Until it kills it) until it crashes either WWW or slows down its crawl on your PC. This empty test site is not there and there is no problem with the console app with the same exact code. Other Leaksource like C # is not the cause of memory leakage, but every time you call a function, memory enhancement has the same effect.
This makes me believe the web-specific problem, there is no problem considering the console app (due to an increase of 20 MB in memory, the runtime for immediate).
I release the 2.7 IPI release and release of Python 1.4.
I do not have the exact code at the moment but it looks like this:
var option = something; Option ["LightWetsk"] = ScriptRuntimeHelpers.True; // any other SO post, 'true' var engine = Python. CreteNine (option) did not seem to work; // // Set up the search path here ... // dynamic field = whatever; ScriptSource source = engine.CreateScriptSourceFromFile ("myscript.py"); // executed? Compile? It populates the scope at this point ... source.Compile (scope); // execute (code, lexer name, style) // This is a python function which highlights the piggs code I have = scope.generate_html ("select from table", "sql", "monokai"); Engine.Shutdown (); shed light on; Like I said, I have copied this code to: a) a console app, b) a new empty web app, c) my original web app console app memory. Did not leak, but used web applications.
I have also executed the function with both the original Python (
python myscript.py ) and the IPI (
). Ipy myscript.py ) and there was no memory leak.
Is there some best practice which is missing me for the runtime? I plan to run it in a shared environment, so the potential task of starting an engine in a separate app pool probably will not work for me (a big leak with the 200 / S is like the showstopper).
Currently, unless a miracle does a miracle, I am planning to scrap my code and just the Javascript syntax highlighter is really unfortunate because the runner is great ...
This is a long story (TLD, sorry) It is happening. NET, you are seeing the possibility of stack fragmentation rather than memory exhaustion.
If you are actually leaking memory then you are holding contexts for a very long time, review your IDisposables (especially written information, especially like links to things, especially There may be injuries like linky things).
I once made an application as fast as 2x, when it is tedious / to deal with a simple one. The largest number of televisions () I needed to run a bigger process
There are good memory profiles for .NET, but I am currently only a mono-propeller (bad memory). A simple Google can help you find a professor you can use on your environment: It should tell you exactly how the what objects are allocated where your heap is breaking
< P> PS Since I saw you pointing to finger pointing on web applications on the second scanning of your question, I will add: Check what is going on indirectly from session state (application, session).
Comments
Post a Comment