[Mono-osx] OSX Reality Check
rmann at digitech.com
Tue Oct 25 15:44:27 EDT 2005
This is very much what we would like to do.
I would be very interested in your thoughts about bridging
strategies and real world experience.
Our 'shared code' will not employ anything really demanding, ... but
to write/debug this code once for two platforms is a big win for us.
I expect Apple will have something to say about Mono/C# at the next
Glad to hear performance is acceptable.
On Oct 25, 2005, at 1:25 PM, Allan Hsu wrote:
> I'm currently working on porting a windows client application
> written in C# (available at http://www.imeem.com) to OS X. We've
> peeled away the Windows-specific code and are running the managed
> code within an embedded Mono instance. We're writing the interface
> in native Cocoa/Objective-C because native look-and-feel is
> important to us. We wrote a bridging layer between the Cocoa AppKit/
> Objective-C and the Mono embedded runtime that makes this pretty
> easy. We plan on eventually contributing the bridging layer back to
> Mono, but this may not happen until some time opens up in our
> development schedule.
> We had to do minimal alterations to our non-Windows-specific code
> to get it to run properly under OS X. Mono really does help in that
> regard. One of the biggest issues we have yet to confront is
> figuring out how to package Mono and its dependencies with our
> application (for various reasons, relying on users to maintain
> their own Mono installations is undesirable for us).
> Also, there are a few outstanding bugs with Mono on OS X that may
> or may not be important to you. The biggest bug for us is that
> asynchronous TCP socket connects don't seem to work at all. This
> bug seems to be the reason why the MySQL.NET connector doesn't work
> under OS X.
> OS X doesn't seem to get as much developer attention as the more
> popular ports, but most of the important bits are there.
> In my experience, Mono on OS X performs roughly on par with Mono on
> x86. The only in-depth benchmarking I've done is with the AES/
> Rijndael implementation; here are my numbers:
> The big difference between Mono and the MS.NET performance is
> partially due to unoptimized implementation and partially due to
> worse JIT optimization. AES is a severe case; most of the time,
> Mono is plenty fast.
> Prime Time? I don't know. It's entirely possible to ship a real
> product using Mono on OS X, but it occasionally requires you to get
> out and push...
> Allan Hsu <allan at counterpop dot net>
> 1E64 E20F 34D9 CBA7 1300 1457 AC37 CBBB 0E92 C779
> On Oct 25, 2005, at 8:41 AM, Roger Mann wrote:
>> .... oops missing subject... sorry for sending this twice
>> I'm new to the list.
>> We are evaluating Mono for Mac OSX...in hopes of being able to
>> share code between Mac & Windows( .Net development).
>> I would love hear from any developer's 'real-world' experiences.
>> Is Mac Mono ready for Prime Time?...or will be ready by say Spring
>> Any Mac developers working on shipping Mac products ....embedding
>> Any real numbers on performance?
>> I spoke to Erik Dasque, and played with the 'Embedded
>> Samples'...so far so good.
>> Any feedback would be greatly appreciated.
>> Harman Music Group
>> Mono-osx mailing list
>> Mono-osx at lists.ximian.com
More information about the Mono-osx