[Mono-dev] How to use cross compiler arm-linux-gcc to build mono for ARM?
Hi, I've build my own board using Samsung S3C2410 ARM9 CPU, and it can run Linux 2.6.14 now. I have only the cross-compiler arm-linux-gcc that running in i386, and I found there is someone has ported mono to Nokia 770, I want to know How to use cross compiler arm-linux-gcc to build mono for ARM? Thank you! ZhangZQ ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] gtkSharpNotification and PKG_CONFIG_PATH help
This is my first posting to the list, so hello, and hope I'm not out of line with anything. I'm in the midst of putting together an OSX/Win application and am new to Mono/Gtk/Cocoa development (although not new to c/++/# or windows development). I'm receiving the following error when trying to run mono TrayTest.exe from the following source (after downloading both and compiling). http://www.mono-project.com/GtkSharpNotificationIcon unhandled exception : system.dllnotfoundexception:gdk-x11-2.0 in Egg.TrayIcon:gdk_x11_display_get_display 1) I have mono version 1.1.14 installed. 2) I'm using cygwin, have all -devel packages installed and can see gdk-x11-2.0.pc in /lib/pkgconfig. 3) PKG_CONFIG_PATH is set and looks like this : /usr/X11R6/lib/pkgconfig:/lib/pkgconfig:/bin/Mono-1.1.14/lib/pkgconfig (mono is installed in c:/cygwin/bin) 4) If I run pkg-config --libs gdk-x11-2.0, I get "Package gdk-x11-2.0 was not found in the pkg-config search path." 5) I turned on PKG_CONFIG_DEBUG_SPEW, and I get the following as a dump : "Cannot open directory '/usr/X11R6/lib/pkgconfig:/lib/pkgconfig:/bin/Mono-1.1.14/lib/pkgconfig' in package search path: No such file or directory. I've scoured the web and read through the blogs of the authors and the only thing I've found is to have the correct libraries installed. I found a few things in gnome to install both the gtk+ devel and glib -devel libraries, but I have (I believe), all the .pc files installed (from the error messages it just looks like pkg-config isn't looking in the right spot). Or if anyone has another idea as to how to get a notification area up and running on win/osX, I'd be much obliged. Any help or advice would be greatly appreciated. Thanks! ...Dave ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Cairo.ImageSurface
On 4/15/06, Philipp Baer <[EMAIL PROTECTED]> wrote: John Luke wrote:[...]>> I must admit that I don't exactly know how the content of a string>> is passed to the non-managed context. Because of the fact that a>> string is composed of chars (which represent 2-byte unicode characters), >> the raw string buffer presumably is a char array. IMO a byte array>> would be much better suited (it simply is more generic).>>> Yes. If you can resend it as a patch, something like that could probably > be committed.Ok, here it is.This is ok to commit in my opinion, however I don't have a mono setup right now to do it. ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] remoting after ip change
I have problem that when I change the IP of the system doing remoting to itself some parts of remoting fail. The error I get is at the bottom. What is really strange is that part of the remoting works, as far as I can tell I can retrieve the remoting object using Activator.GetObject and I can make a method call which returns an array of Marshalbyref objects. The error doesn't happen until I start looping those Objects and printing out values from them. If I kill the code and run it again then it works just fine, this only happens when the IP changes while the code is running. Any ideas? System.Net.WebException: Error: ConnectFailure in <0x000e1> System.Net.HttpWebRequest:EndGetRequestStream (IAsyncResult asyncResult) in <0x00077> System.Net.HttpWebRequest:GetRequestStream () in <0x002ab> System.Runtime.Remoting.Channels.Http.HttpClientTransportSink:CreateWebRequest (System.String url, ITransportHeaders requestHeaders, System.IO.Stream requestStream) in <0x000d0> System.Runtime.Remoting.Channels.Http.HttpClientTransportSink:ProcessMessage (IMessage msg, ITransportHeaders requestHeaders, System.IO.Stream requestStream, ITransportHeaders responseHeaders, System.IO.Stream responseStream) in <0x00068> System.Runtime.Remoting.Channels.SoapClientFormatterSink:SyncProcessMessage (IMessage msg) in <0x002e8> System.Runtime.Remoting.Proxies.RemotingProxy:Invoke (IMessage request) in <0x002e5> System.Runtime.Remoting.Proxies.RealProxy:PrivateInvoke (System.Runtime.Remoting.Proxies.RealProxy rp, IMessage msg, System.Exception exc, System.Object[] out_args) ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Feedback for Windows Forms in 1.1.15
Andreas, Thanks for the testing. I'd be great if you could log bugs for these, that way we don't loose track of them. Also, if you could provide links where to get the sample apps you used, that would help in reproducing the problems. What were the problems with the real-world apps? Cheers, Peter -Original Message- From: "Andreas Nahr" <[EMAIL PROTECTED]> To: Date: 18 April, 2006 13:04 Subject: [Mono-dev] Feedback for Windows Forms in 1.1.15 >I tested some of my Winforms Apps on Mono 1.1.15 running on WinXP x64. >Most of the real-world apps didn't start, but some small sample ones did. >Here are some notes about Problems with the ones that did: > >Toolwindow: >Wrong clipping region/ Window size calculation? See screenshot > >Propertygrid: >* Dividing line at wrong position -> should be at 50% of width of control. >* Readonly Elements not grayed out -> Gray out >* Readonly Elements show designers (if designers are used -> crash) -> >Don't >display designers for readonly elements >* Category Names not determined correctly if inherited -> Use inherited >CategoryName Attributes >* Description default is currently: "TitleThe long important >Description" -> should be "" >* Default: All Entries expanded -> Should be: All entries NOT expanded >* Setting values does not work >* ToolbarVisible/HelpVisible ignored sometimes -> See screenshot > >DirectorySearcher: >Crashes with any selection > >Happy hacking! >Andreas > ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] "safe" way about Marshal.UnsafeAddrOfPinnedArrayElement
Hello, On Tue, 2006-04-18 at 12:27 +0800, GaoXianchao wrote: > hi all, > I'm wrapping epoll api on linux. > To pass address of managed struct array to unmanaged code, I use > Marshal.UnsafeAddrOfPinnedArrayElement . But the method is "unsafe". > Is there a "safe" way to do what the Marshal.UnsafeAddrOfPinnedArrayElement > do? Safe and unsafe are used for different meanings between languages (e.g. C#) and the framework base class library. In this case, like many times it's used in the FX BCL, the Unsafe prefix means that special security requirements for this method makes it dangerous to use with non fully-trusted code (see MSDN documentation for more details). So unless you are using "mono --security" to activate the (partial and unsupported) security manager this *doesn't* affect you at all (as all your code is *always* executed under FullTrust). -- Sebastien Pouliot <[EMAIL PROTECTED]> Blog: http://pages.infinit.net/ctech/ ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] "safe" way about Marshal.UnsafeAddrOfPinnedArrayElement
Hi, Unsafe code usually means code that uses pointers. But that unsafe code has to use the fixed statement when accessing objects that can be moved. fixed statement pins the object so that GC will not move it within the block of fixed statement. (As far as I know Mono still uses non-moving GC but that will change in the future and MS.NET uses moving GC.) Marshal.UnsafeAddrOfPinnedArrayElement is unsafe in addition to pointer operations because it assumes that the array is already pinned so it won't (an in fact can't as it has no scope) pin the object so the GC can move it that can result in invalidation of the returned pointer so has to be used carefully. (When the array is moved by the GC you can overwrite some other object that can have unpredictable results.) Marshal.UnsafeAddrOfPinnedArrayElement is intended to be used with a pinned GCHandle. (Or alternatively on an array that is already pinned using fixed statement but in that case you have a pointer to an array element so you probably don't need this method.) If you have a non-pinned array you should use fixed statement instead as Jonathan explained it. So Marshal.UnsafeAddrOfPinnedArrayElement is even more unsafe than unsafe code when used on a non-pinned array. Kornél - Original Message - From: "Jonathan Pryor" <[EMAIL PROTECTED]> To: "GaoXianchao" <[EMAIL PROTECTED]> Cc: Sent: Tuesday, April 18, 2006 1:48 PM Subject: Re: [Mono-dev] "safe" way about Marshal.UnsafeAddrOfPinnedArrayElement On Tue, 2006-04-18 at 12:27 +0800, GaoXianchao wrote: hi all, I'm wrapping epoll api on linux. To pass address of managed struct array to unmanaged code, I use Marshal.UnsafeAddrOfPinnedArrayElement . But the method is "unsafe". Is there a "safe" way to do what the Marshal.UnsafeAddrOfPinnedArrayElement do? It's "unsafe" because you need to be careful. :-) An alternative is to use *real* unsafe code, e.g. FooStruct[] array = ...; unsafe { fixed (FooStruct* pArray = array) { FooStruct* element = &pArray [index]; } } I'm not entirely sure what your problem is though. "Unsafe" code (like the above `unsafe' block) just means that you're possibly executing without a net, which could result in memory corruption and abortion of the process. :-) Marshal.UnsafeAddrOfPinnedArrayElement isn't even in an unsafe C# context, it just has "Unsafe" in the name. Here's a hint of truth: ANY time you do a P/Invoke, you're potentially doing something unsafe. :-) (Especially considering that you're invoking external code which could really do anything, such as overwrite the entire GC heap, which would cause the process an ugly death in the future...) However, I'm not sure why you need Marshal.UnsafeAddrOfPinnedArrayElement in the first place. epoll_wait(2) takes a `struct epoll_event' array parameter, so unless you wanted to only send a subset of your array, the default marshaling rules which convert a managed array to an array pointer should be fine. - Jon ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] "safe" way about Marshal.UnsafeAddrOfPinnedArrayElement
On Tue, 2006-04-18 at 12:27 +0800, GaoXianchao wrote: > hi all, > I'm wrapping epoll api on linux. > To pass address of managed struct array to unmanaged code, I use > Marshal.UnsafeAddrOfPinnedArrayElement . But the method is "unsafe". > Is there a "safe" way to do what the Marshal.UnsafeAddrOfPinnedArrayElement > do? It's "unsafe" because you need to be careful. :-) An alternative is to use *real* unsafe code, e.g. FooStruct[] array = ...; unsafe { fixed (FooStruct* pArray = array) { FooStruct* element = &pArray [index]; } } I'm not entirely sure what your problem is though. "Unsafe" code (like the above `unsafe' block) just means that you're possibly executing without a net, which could result in memory corruption and abortion of the process. :-) Marshal.UnsafeAddrOfPinnedArrayElement isn't even in an unsafe C# context, it just has "Unsafe" in the name. Here's a hint of truth: ANY time you do a P/Invoke, you're potentially doing something unsafe. :-) (Especially considering that you're invoking external code which could really do anything, such as overwrite the entire GC heap, which would cause the process an ugly death in the future...) However, I'm not sure why you need Marshal.UnsafeAddrOfPinnedArrayElement in the first place. epoll_wait(2) takes a `struct epoll_event' array parameter, so unless you wanted to only send a subset of your array, the default marshaling rules which convert a managed array to an array pointer should be fine. - Jon ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Re: mono frustration
Hi, > > I will wait a few more months I suppose till fedora core 5 binaries are > > released or the yum dependencies are fixed ... > > Maintaining package dependencies and stability of the bleeding-edge > versions of programs is a very hard work for the distributions' > employees and/or contributors. It's normal to have problems like these > ones when using the latest versions of packages, because unstability > increases. Well, that's not quite true. What usually happens is that when a rebuild is requested, not all of the other packages which depend on the rebuilt one are rebuilt. The reasons can be many fold. Normal users (those using core) don't see this very often, but it's something those using the developer branch (rawhide) have to put up with a lot. But hey, it's part of the fun! > And when I have problems, I try to report them as detailed as possible, > and to the correct sources (probably these concrete reports are also > welcome on Fedora mailing-lists, for example, although I don't know if > you have already sent them there). But I understand that this could be > very annoying/frustrating. I've not seen any reports (except for mine) on any of the FC mailing lists (extras/test/developers/core). TTFN Paul -- "Logic, my dear Zoe, is merely the ability to be wrong with authority" - Dr Who ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] mono frustration
Hi, On Tue, 2006-04-18 at 12:24 +0300, Anton Andreev wrote: > Did anyone installed mono on fedora core 5 Yes. It's quite nice. Okay, it needs updating to 1.1.15, but this isn't the right list for discussion on that. > There is a problem with gtkhtml3 and gtk-sharp, pango ... Is there? > Mono frustrated me again,cause I wanted to try the new Fedora with new > Monodevelop with the new Glade designer. Monodevelop is not in FC-extras yet. It is currently awaiting approval. > I even reinstalled my fedora, replaced > the 64 bit edition with 32 bit edition, but guess what - I had the same > problem. And that doesn't surprise me. > I will wait a few more months I suppose till fedora core 5 binaries are > released or the yum dependencies are fixed ... Okay, quick fix (won't work on 64bit fedora core currently). Open a terminal window su yum -y install fedora-rpmdevtools exit fedora-buildrpmtree wget http://www.smmp.salford.ac.uk/packages/boo-0.7.5.2013-2.src.rpm wget http://www.smmp.salford.ac.uk/packages/gtksourceview-sharp-2.0-3.src.rpm wget http://www.smmp.salford.ac.uk/packages/ikvm-0.22-4.src.rpm wget http://www.smmp.salford.ac.uk/packages/mod_mono-1.1.14-1.src.rpm wget http://www.smmp.salford.ac.uk/packages/mono-debugger-0.12-1.src.rpm wget http://www.smmp.salford.ac.uk/packages/monodevelop-0.10-4.src.rpm wget http://www.smmp.salford.ac.uk/packages/monodoc-1.1.13-3.src.rpm wget http://www.smmp.salford.ac.uk/packages/xsp-1.1.13-1.src.rpm rpm -ihv *.src.rpm rpmbuild -ba rpmbuild/SPECS/boo.spec rpmbuild -ba rpmbuild/SPECS/gtksourceview-sharp.spec su rpm -ihv rpmbuild/noarch/* rpm -ihv rpmbuild/i386/* exit rpmbuild -ba rpmbuild/SPECS/ikvm.spec su rpm -ihv rpmbuild/i386/ikvm* exit (you'll need to do this after each package being built - change ikvm for the name of the package) rpmbuild -ba rpmbuild/SPECS/monodoc.spec rpmbuild -ba rpmbuild/SPECS/mono-debugger.spec rpmbuild -ba rpmbuild/SPECS/monodevelop.spec If you want to server ASP.NET files via Apache rpmbuild -ba rpmbuild/SPECS/xsp.spec rpmbuild -ba rpmbuild/SPECS/mod_mono.spec If the build process stops and says you need a particular package, yum -y install That lot may look a bit hairy, but if you follow my instructions, it will be quite painless. I've not put the rpms up on the Salford server as they're not required for Fedora Extras approval. You now have just about all of the rpms available installed on your machine. Any problems, please email me directly (or report them here). I am pushing quite hard for these to be included into Fedora Extras, but things are a tad slow with the chaps currently. It may be worth joining the fedora-extras list as well and poking there. TTFN Paul -- "Logic, my dear Zoe, is merely the ability to be wrong with authority" - Dr Who ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] Re: mono frustration
Anton Andreev escribió: Did anyone installed mono on fedora core 5 There is a problem with gtkhtml3 and gtk-sharp, pango ... Mono frustrated me again,cause I wanted to try the new Fedora with new Monodevelop with the new Glade designer. I could not install the monodevelop. It never ever installed with "yum install monodevelop". Have in mind that I am an ordinary Linux user, .NET developer and I ve been intrested in Mono for a long time. I even reinstalled my fedora, replaced the 64 bit edition with 32 bit edition, but guess what - I had the same problem. I will wait a few more months I suppose till fedora core 5 binaries are released or the yum dependencies are fixed ... Maintaining package dependencies and stability of the bleeding-edge versions of programs is a very hard work for the distributions' employees and/or contributors. It's normal to have problems like these ones when using the latest versions of packages, because unstability increases. In my humble and personal opinion, if you are very interested in the Mono platform and like to try the latest releases of the Mono tools, try the SUSE distribution because it's the closest to this software. Since I use SUSE I don't have any problems with package dependencies or stability of the latest versions of the Mono packages. And when I have problems, I try to report them as detailed as possible, and to the correct sources (probably these concrete reports are also welcome on Fedora mailing-lists, for example, although I don't know if you have already sent them there). But I understand that this could be very annoying/frustrating. Regards, Andrés [ knocte ] -- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
RE: [Mono-dev] System.Web.UI.WebControls/Menu
Yes, possibly. All I wanted is to let the default ("") work. So if someone creates , it does not throw for him. I'll store it in the ViewState. Regards, Konstantin Triger -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Lluis Sanchez Sent: Tuesday, April 18, 2006 11:07 AM To: Konstantin Triger Cc: mono-devel Subject: Re: [Mono-dev] System.Web.UI.WebControls/Menu Shouldn't SkipLinkText be stored in the view state? El lun, 17-04-2006 a las 07:41 -0700, Konstantin Triger escribió: > Hello, > > > > The attached patch handles the following: > > 1. Enables DataBinding by not throwing NotImplementedException in > OnDataBound event. > 2. Provides basic CreateChildControls implementation by ensuring > the control is bound. > 3. Ensures the child controls are created when the postback event > is raised. > 4. Provides default implementation for SkipLinkText to let the > default functionality to work. > > > > Please review. If no one objects I'll commit. > > > > Regards, > > Konstantin Triger > > > > > ___ > Mono-devel-list mailing list > Mono-devel-list@lists.ximian.com > http://lists.ximian.com/mailman/listinfo/mono-devel-list ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] mono frustration
Did anyone installed mono on fedora core 5 There is a problem with gtkhtml3 and gtk-sharp, pango ... Mono frustrated me again,cause I wanted to try the new Fedora with new Monodevelop with the new Glade designer. I could not install the monodevelop. It never ever installed with "yum install monodevelop". Have in mind that I am an ordinary Linux user, .NET developer and I ve been intrested in Mono for a long time. I even reinstalled my fedora, replaced the 64 bit edition with 32 bit edition, but guess what - I had the same problem. I will wait a few more months I suppose till fedora core 5 binaries are released or the yum dependencies are fixed ... ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] System.Web.UI.WebControls/Menu
Shouldn't SkipLinkText be stored in the view state? El lun, 17-04-2006 a las 07:41 -0700, Konstantin Triger escribió: > Hello, > > > > The attached patch handles the following: > > 1. Enables DataBinding by not throwing NotImplementedException in > OnDataBound event. > 2. Provides basic CreateChildControls implementation by ensuring > the control is bound. > 3. Ensures the child controls are created when the postback event > is raised. > 4. Provides default implementation for SkipLinkText to let the > default functionality to work. > > > > Please review. If no one objects I’ll commit. > > > > Regards, > > Konstantin Triger > > > > > ___ > Mono-devel-list mailing list > Mono-devel-list@lists.ximian.com > http://lists.ximian.com/mailman/listinfo/mono-devel-list ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
RE: [Mono-dev] System.Web.UI.WebControls/FormView
Hello, Attached a reworked patch, which keeps the old logic if FormView.AllowPaging is true. In addition, EnsureDataBound() and CreateControlsStyle() fallback to base. If no one objects I’ll commit. Regards, Konstantin Triger From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Konstantin Triger Sent: Sunday, April 16, 2006 4:04 PM To: mono-devel Subject: [Mono-dev] System.Web.UI.WebControls/FormView Hello, The attached patch ensures rebinding when FormView.PageIndex is called. A relevant test is inside. If no one objects I’ll commit. Regards, Konstantin Triger fv.patch Description: fv.patch ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list