Re: [Mono-list] build once, run anywhere?
On 05/03/2011 07:08 PM, pshomov wrote: Hi guys, I am building mono from source on one machine and then I am zipping the binaries and copying them over to other machines and using them (saving on the build time here). Is this a good idea? I am getting some exceptions from mono on these other machines that are making me suspicious. The problem with doing this is that different distros (and versions of the same distro) bundle different versions of libraries and might put them in different locations, etc. You might also not realize that you are missing libraries (since you didn't get a build warning) on some machines. Also need to make sure all systems are the same architecture. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] undefined reference to `libiconv'
On 04/26/2011 10:54 PM, Makoto KISHIMOTO wrote: Hello, In my FreeBSD Box, mono trunk build failed. The following is build log. $ gmake gmake all-recursive gmake[1]: Entering directory `/export/home/ksmakoto/Mono/BUILD' (snip) gmake[2]: Leaving directory `/export/home/ksmakoto/Mono/BUILD/libgc' Making all in eglib gmake[2]: Entering directory `/export/home/ksmakoto/Mono/BUILD/eglib' gmake all-recursive gmake[3]: Entering directory `/export/home/ksmakoto/Mono/BUILD/eglib' Making all in m4 gmake[4]: Entering directory `/export/home/ksmakoto/Mono/BUILD/eglib/m4' gmake[4]: Nothing to be done for `all'. gmake[4]: Leaving directory `/export/home/ksmakoto/Mono/BUILD/eglib/m4' Making all in src gmake[4]: Entering directory `/export/home/ksmakoto/Mono/BUILD/eglib/src' gmake[4]: Nothing to be done for `all'. gmake[4]: Leaving directory `/export/home/ksmakoto/Mono/BUILD/eglib/src' Making all in test gmake[4]: Entering directory `/export/home/ksmakoto/Mono/BUILD/eglib/test' ./doltlibtool --tag=CC --mode=link gcc -Wall -D_FORTIFY_SOURCE=2 -I../../../mono/eglib/test/../src -I../src -DDRIVER_NAME=\EGlib\ -g -O2 -g -D_GNU_SOURCE -Wl,-export-dynamic -o test-eglib test_eglib-test.o test_eglib-driver.o test_eglib-hashtable.o test_eglib-string-util.o test_eglib-string.o test_eglib-slist.o test_eglib-sizes.o test_eglib-ptrarray.o test_eglib-list.o test_eglib-array.o test_eglib-fake.o test_eglib-path.o test_eglib-queue.o test_eglib-shell.o test_eglib-spawn.o test_eglib-timer.o test_eglib-file.o test_eglib-pattern.o test_eglib-dir.o test_eglib-markup.o test_eglib-unicode.o test_eglib-utf8.o test_eglib-endian.o test_eglib-module.o test_eglib-memory.o ../src/libeglib.la -lm libtool: link: gcc -Wall -D_FORTIFY_SOURCE=2 -I../../../mono/eglib/test/../src -I../src -DDRIVER_NAME=\EGlib\ -g -O2 -g -D_GNU_SOURCE -Wl,-export-dynamic -o test-eglib test_eglib-test.o test_eglib-driver.o test_eglib-hashtable.o test_eglib-string-util.o test_eglib-string.o test_eglib-slist.o test_eglib-sizes.o test_eglib-ptrarray.o test_eglib-list.o test_eglib-array.o test_eglib-fake.o test_eglib-path.o test_eglib-queue.o test_eglib-shell.o test_eglib-spawn.o test_eglib-timer.o test_eglib-file.o test_eglib-pattern.o test_eglib-dir.o test_eglib-markup.o test_eglib-unicode.o test_eglib-utf8.o test_eglib-endian.o test_eglib-module.o test_eglib-memory.o ../src/.libs/libeglib.a -lm ./src/.libs/libeglib.a(libeglib_la-gunicode.o)(.text+0x5b6): In function `monoeg_g_get_charset': ./../../mono/eglib/src/gunicode.c:362: undefined reference to `locale_charset' Looks like configure detected that your system had localcharset.h which should have prototyped a function called locale_charset() which we use to get the user's locale charset. You might be able to fix this by commenting out the #define HAVE_LOCALCHARSET_H in config.h Alternatively, it looks like it might be a part of libiconv (solution below). ./src/.libs/libeglib.a(libeglib_la-giconv.o)(.text+0x4c): In function `monoeg_g_iconv': ./../../mono/eglib/src/giconv.c:177: undefined reference to `libiconv' ./src/.libs/libeglib.a(libeglib_la-giconv.o)(.text+0x18f): In function `monoeg_g_iconv_close': ./../../mono/eglib/src/giconv.c:158: undefined reference to `libiconv_close' ./src/.libs/libeglib.a(libeglib_la-giconv.o)(.text+0x8c6): In function `monoeg_g_iconv_open': ./../../mono/eglib/src/giconv.c:132: undefined reference to `libiconv_open' These errors suggest that configure detected that you had libiconv installed, so it tried to use it but apparently thought it was part of your libc because I do not see a -liconv in the link line. I fixed an issue similar to this yesterday, maybe your version of mono from git is older than my fix (commit = 1deef5a760b37cf1e42a60110981d0f752d84d9d ). You can either try updating from git or add $(LIBICONV) right after each ($GLIB_LIBS) in the Makefile.am files and then rebuild. Hope that helps, Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] gunicode.c:234: error: `ENOTSUP' undeclared (first use in this function)
On 04/24/2011 02:39 AM, Rod wrote: While attempting to make mono from GIT I got this error: gunicode.c:234: error: `ENOTSUP' undeclared (first use in this function) I Google and nothing about gunicode.c and ENOTSUP. I just fixed this, try pulling from master again. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] [Mono-patches] [mono/moon] caef081d: Implemented an iconv-like API for unicode conversion
On 04/20/2011 12:00 AM, Rodrigo Kumpera wrote:
Jeff, shouldn't this be part of eglib as a build time option?
I suppose it could be, but I think the plan right now is to simply
remove g_iconv() from eglib since eglib is supposed to remain a
minimalist version of glib as per discussion with Miguel.
It's also probably not very useful for most people who want to use
g_iconv() to *just* have UTF-16/32BE/LE to UTF-8 conversion.
Jeff
On Apr 19, 2011 10:06 PM, Jeffrey Stedfast (f...@gnome.org
mailto:f...@gnome.org) mono-patc...@lists.ximian.com
mailto:mono-patc...@lists.ximian.com wrote:
Branch: refs/heads/master
Home: https://github.com/mono/moon
Commit: caef081dc43c17327c37932f65264deec19d8022
Author: Jeffrey Stedfast f...@gnome.org mailto:f...@gnome.org
Date: 04/19/2011 21:05:59
URL:
https://github.com/mono/moon/commit/caef081dc43c17327c37932f65264deec19d8022
Implemented an iconv-like API for unicode conversion
Changed paths:
M src/Makefile.am
Added paths:
A src/miconv.cpp
A src/miconv.h
Modified: src/Makefile.am
===
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -78,6 +78,7 @@ libmoon_include_headers = \
medialog.h \
mediaplayer.h \
messaging.h \
+ miconv.h \
moon-curves.h \
moonlightconfiguration.h\
moon-path.h \
@@ -234,6 +235,7 @@ dist_libmoon_la_SOURCES = \
medialog.cpp \
mediaplayer.cpp \
messaging.cpp \
+ miconv.cpp \
moon-curves.c \
moonlightconfiguration.cpp \
moon-path.c \
Added: src/miconv.cpp
===
--- /dev/null
+++ b/src/miconv.cpp
@@ -0,0 +1,344 @@
+/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: t;
c-basic-offset: 8 -*- */
+/*
+ * miconv.cpp:
+ *
+ * Contact:
+ * Moonlight List (moonlight-l...@lists.ximian.com
mailto:moonlight-l...@lists.ximian.com)
+ *
+ * Copyright 2011 Novell, Inc. (http://www.novell.com)
+ *
+ * See the LICENSE file included with the distribution for details.
+ */
+
+#include config.h
+
+#include glib.h
+#include string.h
+#include errno.h
+
+#include miconv.h
+
+namespace Moonlight {
+
+enum Endian {
+ LittleEndian,
+ BigEndian
+};
+
+typedef int (* Decoder) (Endian endian, char **inbytes, size_t
*inbytesleft, gunichar *outchar);
+typedef int (* Encoder) (gunichar c, char **outbytes, size_t
*outbytesleft);
+
+static int decode_utf32 (Endian endian, char **inbytes, size_t
*inbytesleft, gunichar *outchar);
+//static int encode_utf32 (gunichar c, char **outbytes, size_t
*outbytesleft);
+
+static int decode_utf16 (Endian endian, char **inbytes, size_t
*inbytesleft, gunichar *outchar);
+//static int encode_utf16 (gunichar c, char **outbytes, size_t
*outbytesleft);
+
+static int decode_utf8 (Endian endian, char **inbytes, size_t
*inbytesleft, gunichar *outchar);
+static int encode_utf8 (gunichar c, char **outbytes, size_t
*outbytesleft);
+
+static struct {
+ const char *name;
+ Decoder decoder;
+ Encoder encoder;
+ Endian endian;
+} charsets[] = {
+ { UTF-32BE, decode_utf32, NULL, BigEndian },
+ { UTF-32LE, decode_utf32, NULL, LittleEndian },
+ { UTF-16BE, decode_utf16, NULL, BigEndian },
+ { UTF-16LE, decode_utf16, NULL, LittleEndian },
+ { UTF-8, decode_utf8, encode_utf8, LittleEndian },
+};
+
+struct _miconv_t {
+ Decoder decode;
+ Encoder encode;
+ Endian endian;
+ gunichar c;
+};
+
+
+miconv_t
+miconv_open (const char *to, const char *from)
+{
+ Decoder decoder = NULL;
+ Encoder encoder = NULL;
+ Endian endian;
+ miconv_t cd;
+ guint i;
+
+ if (!to || !from)
+ return (miconv_t) -1;
+
+ for (i = 0; i G_N_ELEMENTS (charsets); i++) {
+ if (!strcmp (charsets[i].name, from)) {
+ decoder = charsets[i].decoder;
+ endian = charsets[i].endian;
+ }
+
+ if (!strcmp (charsets[i].name, to))
+ encoder = charsets[i].encoder;
+ }
+
+ if (encoder == NULL || decoder == NULL)
+ return (miconv_t) -1;
+
+ cd = (miconv_t) g_malloc (sizeof (*cd));
+ cd-decode = decoder;
+ cd-encode = encoder;
+ cd-endian = endian;
+ cd-c = -1;
+
+ return cd;
+}
+
+int
+miconv_close (miconv_t cd)
+{
+ g_free (cd);
+ return 0;
+}
+
+int
+miconv (miconv_t cd, char **inbytes, size_t *inbytesleft,
+ char **outbytes, size_t *outbytesleft)
+{
+ size_t inleft, outleft;
+ char *inptr, *outptr;
+ gunichar c;
+ int rc = 0;
+
+ if (outbytes == NULL || outbytesleft == NULL) {
+ /* reset converter */
+ cd-c = -1;
+ return 0;
+ }
+
+ inleft = inbytesleft ? *inbytesleft : 0;
+ inptr = inbytes ? *inbytes : NULL;
+ outleft = *outbytesleft;
+ outptr = *outbytes;
+ c = cd-c;
+
+ while (inleft = 0 outleft 0) {
+ if (c == (gunichar) -1 cd-decode (cd-endian, inptr, inleft,
c) == -1) {
+ rc = -1;
+ break;
+ }
+
+ if (cd-encode (c, outptr, outleft) == -1) {
+ rc = -1;
+ break;
+ }
+
+ c = -1;
+ }
+
+ if (inbytesleft)
+ *inbytesleft = inleft;
+
+ if (inbytes)
+ *inbytes = inptr
Re: [Mono-dev] Possible QuickSort optimizations for Array.Sort()
On 04/09/2011 04:39 PM, Jeff Stedfast wrote:
On 04/06/2011 03:18 AM, Marek Safar wrote:
Hello,
public class SortT {
public delegate int Comparison (T v0, T v1);
You don't need yet another delegate, just use standard Func
static readonly int INSERTIONSORT_THRESHOLD = 7;
Why not to used const int ?
More importantly what is performance of sorting array of 10-20
elements called 100 times ?
Attached is another test program with 2 newer/faster implementations
of QuickSort, one of which is Non-Recursive.
You can run this program like so:
mono ./qsort.exe [-random, -sorted, -reversed] [array size] [# loops]
These 2 new implementations manage to reduce the number of comparisons
done by merging the Median() function from the previous set into the
main body of the QuickSort function and swapping those 3 elements into
ascending order as it compares to find the optimal pivot. This has the
advantage of allowing our loops to start at low+1 and high-1 (thus
reducing our compares by 2 for each partition).
My previous implementation was able to reduce the compares by 1 by
starting at low+1 by moving the pivot into low, but this has the
negative impact of destroying performance on reverse-sorted arrays. My
new implementation doesn't have this problem (and therefore, we don't
need to fall back to that insertion-sort if we detect that
pathological case).
Interestingly, the NonRecursiveQuickSort() implementation does not
noticeably improve performance over the recursive version and in fact
is slower than even our current implementation for sorting small
arrays (made evident when looped millions of times).
How do you feel about replacing corlib's qsort() implementation with
my FasterQuickSort() implementation?
Jeff
Just to be clear, I meant how do you feel about *me* doing the work to
replace the current QuickSort impl in corlib with mine? :-)
I just want to make sure before I create a patch that people are interested.
That said, I forgot to include benchmark results:
[fejj@serenity sorting]$ mono ./qsort.exe -random 10 1000
BasicQuickSort comparisons needed: 43000
FasterQuickSort comparisons needed:25000
BasicQuickSort finished in:00:00:18.8535031s
FasterQuickSort finished in: 00:00:16.4478594s
[fejj@serenity sorting]$ mono ./qsort.exe -sorted 10 1000
BasicQuickSort comparisons needed: 33000
FasterQuickSort comparisons needed:18000
BasicQuickSort finished in:00:00:16.9532771s
FasterQuickSort finished in: 00:00:14.1202212s
[fejj@serenity sorting]$ mono ./qsort.exe -reversed 10 1000
BasicQuickSort comparisons needed: 36000
FasterQuickSort comparisons needed:18000
BasicQuickSort finished in:00:00:17.4450895s
FasterQuickSort finished in: 00:00:15.1590930s
Jeff
___
Mono-devel-list mailing list
Mono-devel-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Possible QuickSort optimizations for Array.Sort()
On 04/06/2011 03:18 AM, Marek Safar wrote:
Hello,
public class SortT {
public delegate int Comparison (T v0, T v1);
You don't need yet another delegate, just use standard Func
static readonly int INSERTIONSORT_THRESHOLD = 7;
Why not to used const int ?
More importantly what is performance of sorting array of 10-20
elements called 100 times ?
Attached is another test program with 2 newer/faster implementations of
QuickSort, one of which is Non-Recursive.
You can run this program like so:
mono ./qsort.exe [-random, -sorted, -reversed] [array size] [# loops]
These 2 new implementations manage to reduce the number of comparisons
done by merging the Median() function from the previous set into the
main body of the QuickSort function and swapping those 3 elements into
ascending order as it compares to find the optimal pivot. This has the
advantage of allowing our loops to start at low+1 and high-1 (thus
reducing our compares by 2 for each partition).
My previous implementation was able to reduce the compares by 1 by
starting at low+1 by moving the pivot into low, but this has the
negative impact of destroying performance on reverse-sorted arrays. My
new implementation doesn't have this problem (and therefore, we don't
need to fall back to that insertion-sort if we detect that pathological
case).
Interestingly, the NonRecursiveQuickSort() implementation does not
noticeably improve performance over the recursive version and in fact is
slower than even our current implementation for sorting small arrays
(made evident when looped millions of times).
How do you feel about replacing corlib's qsort() implementation with my
FasterQuickSort() implementation?
Jeff
using System;
using System.Diagnostics;
namespace SortingTest {
public class SortT {
public delegate int Comparison (T v0, T v1);
static void swap (T [] array, int i, int j)
{
T tmp = array[i];
array[i] = array[j];
array[j] = tmp;
}
static void QuickSortBasicR (T [] array, int low0, int high0, Comparison comparison)
{
int low = low0;
int high = high0;
// Be careful with overflows
int mid = low + ((high - low) / 2);
T keyPivot = array [mid];
while (true) {
if (keyPivot == null) {
while (low high0 array [low] == null)
++low;
while (high low0 array [high] != null)
--high;
} else {
// Move the walls in
while (low high0 comparison (array [low], keyPivot) 0)
++low;
while (high low0 comparison (keyPivot, array [high]) 0)
--high;
}
if (low = high) {
swap (array, low, high);
++low;
--high;
} else
break;
}
if (low0 high)
QuickSortBasicR (array, low0, high, comparison);
if (low high0)
QuickSortBasicR (array, low, high0, comparison);
}
public static void QuickSortBasic (T [] array, int index, int length, Comparison comparison)
{
QuickSortBasicR (array, index, index + length - 1, comparison);
}
const int INSERTIONSORT_THRESHOLD = 7;
static void FasterQuickSortR (T [] array, int low, int high, Comparison comparison)
{
int mid, i, k;
T pval;
if ((low + INSERTIONSORT_THRESHOLD) high) {
// switch to insertion sort
for (i = low + 1; i = high; i++)
for (k = i; k low comparison (array[k], array[k-1]) 0; k--)
swap (array, k - 1, k);
return;
}
// calculate the middle element
mid = low + ((high - low) / 2);
// once we re-order the lo, mid, and hi elements to be in
// ascending order, we'll use mid as our pivot.
if (comparison (array[mid], array[low]) 0)
swap (array, mid, low);
if (comparison (array[high], array[mid]) 0) {
swap (array, mid, high);
if (comparison (array[mid], array[low]) 0)
swap (array, mid, low);
}
/* since we've already guaranteed that lo = mid and mid = hi,
* we can skip comparing them again */
pval = array[mid];
k = high - 1;
i = low + 1;
do {
if (pval == null) {
while (i k array[i] == null)
i++;
while (k = i array[k] != null)
k--;
} else {
// find the first element with a value pivot value
while (i k comparison (array[i], pval) = 0)
i++;
// find the last element with a value = pivot value
while (k = i comparison (pval, array[k]) 0)
k--;
}
if (k = i)
break;
swap (array, i, k);
// make sure we keep track of our pivot element
if (mid == i)
mid = k;
else if (mid == k)
mid = i;
i++;
k--;
} while (true);
if (k != mid) {
// swap the pivot with the last element in the first partition
swap (array, mid, k);
}
/* recursively sort each partition */
if ((k + 1) high) FasterQuickSortR (array, k + 1, high, comparison);
if ((k - 1) low) FasterQuickSortR (array, low, k - 1,
Re: [Mono-dev] Possible QuickSort optimizations for Array.Sort()
On 04/06/2011 03:18 AM, Marek Safar wrote:
Hello,
public class SortT {
public delegate int Comparison (T v0, T v1);
You don't need yet another delegate, just use standard Func
It was just a quick hack to get things going, if we decide it's worth
having corlib move to my QuickSort implementation, it would use the same
standard func corlib is using now.
static readonly int INSERTIONSORT_THRESHOLD = 7;
Why not to used const int ?
I never learned when to use const vs static/readonly in C# :-(
I assume you suggest const because it is more appropriate, so I will :-)
More importantly what is performance of sorting array of 10-20
elements called 100 times ?
I've attached a sample program to test this in case you'd like to play
with it some more, but here are the results on my system for 20 elements
@ 1 million iterations:
[fejj@serenity sorting]$ mono ./qsort-small.exe -random 100
Timings for sorting 20 random items 100 times.
Basic QuickSort finished in: 00:00:03.0227324s
QuickSort+Insertion finished in: 00:00:02.5357660s
[fejj@serenity sorting]$ mono ./qsort-small.exe -sorted 100
Timings for sorting 20 sorted items 100 times.
Basic QuickSort finished in: 00:00:02.3003898s
QuickSort+Insertion finished in: 00:00:01.9493073s
[fejj@serenity sorting]$ mono ./qsort-small.exe -reversed 100
Timings for sorting 20 reversed items 100 times.
Basic QuickSort finished in: 00:00:02.4001482s
QuickSort+Insertion finished in: 00:00:02.3277809s
And here are the results for 10 elements @ 1 million iterations:
[fejj@serenity sorting]$ mono ./qsort-small.exe -random 100
Timings for sorting 10 random items 100 times.
Basic QuickSort finished in: 00:00:01.8165901s
QuickSort+Insertion finished in: 00:00:01.7363410s
[fejj@serenity sorting]$ mono ./qsort-small.exe -sorted 100
Timings for sorting 10 sorted items 100 times.
Basic QuickSort finished in: 00:00:01.7398261s
QuickSort+Insertion finished in: 00:00:01.5270049s
[fejj@serenity sorting]$ mono ./qsort-small.exe -reversed 100
Timings for sorting 10 reversed items 100 times.
Basic QuickSort finished in: 00:00:01.7166351s
QuickSort+Insertion finished in: 00:00:01.6798535s
Hope that helps,
Jeff
using System;
using System.Diagnostics;
namespace SortingTest {
public class SortT {
public delegate int Comparison (T v0, T v1);
static void swap (T [] array, int i, int j)
{
T tmp = array[i];
array[i] = array[j];
array[j] = tmp;
}
static void QuickSortBasicR (T [] array, int low0, int high0, Comparison comparison)
{
int low = low0;
int high = high0;
// Be careful with overflows
int mid = low + ((high - low) / 2);
T keyPivot = array [mid];
while (true) {
if (keyPivot == null) {
while (low high0 array [low] == null)
++low;
while (high low0 array [high] != null)
--high;
} else {
// Move the walls in
while (low high0 comparison (array [low], keyPivot) 0)
++low;
while (high low0 comparison (keyPivot, array [high]) 0)
--high;
}
if (low = high) {
swap (array, low, high);
++low;
--high;
} else
break;
}
if (low0 high)
QuickSortBasicR (array, low0, high, comparison);
if (low high0)
QuickSortBasicR (array, low, high0, comparison);
}
public static void QuickSortBasic (T [] array, int index, int length, Comparison comparison)
{
QuickSortBasicR (array, index, index + length - 1, comparison);
}
static int Median (T [] array, int a, int b, int c, Comparison comparison)
{
T keyA = array[a];
T keyB = array[b];
T keyC = array[c];
if (comparison (keyA, keyB) 0) {
/* a b c */
if (comparison (keyB, keyC) 0)
return b;
/* a c b */
if (comparison (keyA, keyC) 0)
return c;
/* c a b */
return a;
} else {
/* b a c */
if (comparison (keyA, keyC) 0)
return a;
/* b c a */
if (comparison (keyB, keyC) 0)
return c;
/* c b a */
return b;
}
}
static void QuickSortMedianR (T [] array, int low, int high, Comparison comparison)
{
int pivot, mid, i, k;
T pval;
// determine which element contains the median value
mid = low + ((high - low) / 2);
pivot = Median (array, low, mid, high, comparison);
if (pivot != low) {
// swap pivot value into first element (so the location stays constant)
swap (array, low, pivot);
pivot = low;
}
pval = array[pivot];
i = low + 1;
k = high;
do {
if (pval == null) {
while (i k array[i] == null)
i++;
while (k = i array[k] != null)
k--;
} else {
// find the first element with a value pivot value
while (i k comparison (array[i], pval) = 0)
i++;
// find the last element with a value = pivot value
Re: [Mono-dev] Possible QuickSort optimizations for Array.Sort()
On 04/06/2011 10:40 AM, Robert Jordan wrote: Hi, On 06.04.2011 15:48, Jeffrey Stedfast wrote: static readonly int INSERTIONSORT_THRESHOLD = 7; Why not to used const int ? I never learned when to use const vs static/readonly in C# :-( Consts will be expanded inline by the compiler. pros: fast cons: versioning trouble when the const is public and it is consumed by another assembly. Aha, thanks. So yea, in this case it makes sense to use const. Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] Possible QuickSort optimizations for Array.Sort()
Attached you'll find qsort.cs which includes 3 QuickSort implementations:
1. QuickSortBasic[R]: This is a basic QuickSort implementation and the
one currently in use as the sorting function in Array.Sort(). This
implementation always uses the middle element as the pivot.
2. QuickSortMedian[R]: This implementation is also a basic QuickSort
implementation which attempts choose a better pivot based on the first,
last, and middle elements. For sorted inputs and random inputs, this
seems to help performance somewhat as the array gets larger and larger.
However, for reverse-sorted inputs, it seems to kill performance.
Perhaps if I choose 3 random points to feed into Median() it will do better.
3. QuickSortInsertion[R]: This implementation includes the above perf
trick and also includes an Insertion Sort fallback when partitions are
broken down smaller than some threshold # of elements (7 in the attached
code). It also falls back to Insertion Sort when it encounters a
worst-case scenario. As always, the performance improvement with this is
greater as the number of elements increases and/or as the comparison
function becomes more complex.
You'll notice that the attached test program will print out the number
of comparisons that each sort routine uses in order to complete its job.
It's important to keep in mind that the more complex the comparison
function is, the more important it is that the number of comparisons is
kept to a minimum.
Here are some example runs:
[fejj@warpig sorting]$ mono ./qsort.exe -random 1
Basic QuickSort comparisons needed: 3807986012
QuickSort+Median comparisons needed:3077915654
QuickSort+Insertion comparisons needed: 3021355043
Basic QuickSort finished in: 00:00:43.8178721s
QuickSort+Median finished in:00:00:37.3047443s
QuickSort+Insertion finished in: 00:00:36.6122681s
[fejj@warpig sorting]$ mono ./qsort.exe -sorted 1
Basic QuickSort comparisons needed: 2632227866
QuickSort+Median comparisons needed:2632227867
QuickSort+Insertion comparisons needed: 2483222808
Basic QuickSort finished in: 00:00:21.8213618s
QuickSort+Median finished in:00:00:21.2622527s
QuickSort+Insertion finished in: 00:00:19.0837352s
(Note: I reduced the array size here due to wanting to head home for dinner)
[fejj@warpig sorting]$ mono ./qsort.exe -reversed 1
Basic QuickSort comparisons needed: 129546
QuickSort+Median comparisons needed:12522499
QuickSort+Insertion comparisons needed: 39993
Basic QuickSort finished in: 00:00:00.0016905s
QuickSort+Median finished in:00:00:00.0911385s
QuickSort+Insertion finished in: 00:00:00.0008721s
Jeff
using System;
using System.Diagnostics;
namespace SortingTest {
public class SortT {
public delegate int Comparison (T v0, T v1);
static void swap (T [] array, int i, int j)
{
T tmp = array[i];
array[i] = array[j];
array[j] = tmp;
}
static void QuickSortBasicR (T [] array, int low0, int high0, Comparison comparison)
{
int low = low0;
int high = high0;
// Be careful with overflows
int mid = low + ((high - low) / 2);
T keyPivot = array [mid];
while (true) {
if (keyPivot == null) {
while (low high0 array [low] == null)
++low;
while (high low0 array [high] != null)
--high;
} else {
// Move the walls in
while (low high0 comparison (array [low], keyPivot) 0)
++low;
while (high low0 comparison (keyPivot, array [high]) 0)
--high;
}
if (low = high) {
swap (array, low, high);
++low;
--high;
} else
break;
}
if (low0 high)
QuickSortBasicR (array, low0, high, comparison);
if (low high0)
QuickSortBasicR (array, low, high0, comparison);
}
public static void QuickSortBasic (T [] array, int index, int length, Comparison comparison)
{
QuickSortBasicR (array, index, index + length - 1, comparison);
}
static int Median (T [] array, int a, int b, int c, Comparison comparison)
{
T keyA = array[a];
T keyB = array[b];
T keyC = array[c];
if (comparison (keyA, keyB) 0) {
/* a b c */
if (comparison (keyB, keyC) 0)
return b;
/* a c b */
if (comparison (keyA, keyC) 0)
return c;
/* c a b */
return a;
} else {
/* b a c */
if (comparison (keyA, keyC) 0)
return a;
/* b c a */
if (comparison (keyB, keyC) 0)
return c;
/* c b a */
return b;
}
}
static void QuickSortMedianR (T [] array, int low, int high, Comparison comparison)
{
int pivot, mid, i, k;
T pval;
// determine which element contains the median value
mid = low + ((high - low) / 2);
pivot = Median (array, low, mid, high, comparison);
if (pivot != low) {
// swap pivot value into first element (so the location stays constant)
swap (array, low, pivot);
pivot = low;
}
pval =
Re: [Mono-dev] Sale of Novell
On 11/23/2010 08:50 AM, Paul F. Johnson wrote: Hi, Now that Novell has been sold and along with it 882 patents either directly or indirectly to MS, can anyone give a definitive statement on what will happen with Mono and if the licences will remain the same? I don't want to see the sale mean the death of Mono, but with the future of SUSE unclear now, we could do with some sort of re-assurance. TTFN Paul As Miguel stated yesterday, the only thing that will change wrt Mono is where we, the developers, will be getting our paychecks. They'll be coming from Attachmate instead of Novell in the near future (assuming the deal goes through). There are also no plans to change the licensing of Mono. The same holds true for SUSE. Hope that helps, Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] Ubuntu
On 08/11/2010 09:24 PM, Daniel Hughes wrote: Ubuntu does not believe it is its responsibility to update mono between OS releases. It's their responsibility to update *all* of their packages between OS releases. If they don't, then they presumably had a reason not to. It's kinda hard for them not to take responsibility for the packages they ship ;-) Mono does not believe it is its responsibility to provide ubuntu packages for new mono releases. Users fall into a gap between the two. And must compile from source or use unsupported third party PPA's if and when they are available. Jo (aka Directhex) is the guy that packages Mono for Ubuntu proper, so his PPA isn't really third party. Hope that helps, Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] mono performance, 20x differential with Java (what am i doing wrong)
On 02/01/2010 08:16 PM, Diego Frata wrote: Jon, Not going to happen? Heck, few years back I wouldn't say Mono would happen. Implementing the CLR, porting lots of .NET libraries without even looking at implementation details? No f*ing way. :) It just requires work and time (= money) and passion, as everything else on earth. Agreed ;-) FWIW, the JIT guys are looking into this and have some ideas. I'm confidant that it'll happen. Hope that sometime between this year and the next one I can start contributing instead of just using it. And maybe you too! I think this is certainly possible! Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Mono 2.6 for Ubuntu
On 01/11/2010 03:42 PM, James Mansion wrote: B.R. wrote: It's not really up to Novell here, it's up to the Debian/Ubuntu maintainers to decide when new versions get supported. Part of the reason for this is that Debian, and by extension Ubuntu, have very specific policies on how things should be packaged (the Debian Free Software Guidelines are very strict, and the Mono project does violate them in a few places), and Novell simply haven't the resources to invest in hiring someone specifically to do Debian (and by extension Ubuntu) packaging, so it's left up to the maintainers who do a pretty good job with the time and resources they have. The reason binaries are shipped for Windows with every release is because there's no package specification to follow, so the binary packages are easy to build (build the libs, put them in the right places, package them up in an installer and you're good to go). I think that's a bit lame - it makes a big assumption about the mono team shipping a package for a start. Binary compatibility for apps (if not drivers) is at least passable with current Linux systems. Why should it be hard to ship a universal binary when Adobe and Sun (in particular) have managed it for a long time? We used to make binary packages but they aren't upgradeable via the distribution's package system. You really really want it to be upgradeable via the package system and have the package system be aware of the Mono installation, otherwise the system quickly becomes a mess. I doubt you have the binary install of Java or OpenOffice on your machine. I'm sure you've gotten them via the package manager from your distro. There are people working on making Ubuntu packages (that also make the official ubuntu/debian packages). I can see that Novell need SLES to be the best platform for Mono for commercial, but I think it needs to be ubiquitous first and then given value add on the Novell product - and that means better support for other Linuxen, and OpenSolaris, and FreeBSD. We are always accepting patches to make it work better on other platforms - there's no plot to make Mono work best on SLE (it's the same Mono as on every other platform). Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Sites Using Mono
On 11/20/2009 07:01 AM, sirmak wrote: Hi, I sent a similar message before and couldn't get an answer. Web Projects Using Mono section in http://www.mono-project.com/Companies_Using_Mono Companies_Using_Mono page is old, I checked every site there with alexa and there is no commerical high load site in the list. Wikipedia only uses it for search with lucene.net and Mozilla is only using deki. Isn't there ANY high load site reference in mono community that constructed with mono ? Cheers, Wouldn't you consider SecondLife to be high-load? Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Embeding Mono. Memory management
On 11/20/2009 08:25 AM, Alexander Smirnov wrote:
But how GC knows when to free such objects? When it frees such objects?
Could GC free the object when I still need it?
for example
void func() {
MonoString* str=mono_string_new (domain, some text);
/*
could GC free str before using str, if no, why?
*/
// using str
}
thanks
You might find this article informative:
http://en.wikipedia.org/wiki/Garbage_collection_%28computer_science%29
The gist of it is that Mono's GC (which is just Boehm's GC) scans for
references to memory that it has allocated, so as long as your program
has pointers to these objects, the GC will not collect them.
Hope that helps,
Jeff
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] Should we replace MemoryStream?
Miguel de Icaza wrote: Hello folks, I just blogged about a memory fragmentation issue here: http://tirania.org/blog/archive/2009/Nov-09.html And I am wondering: since MemoryStream is one of these sources of problems, we could replace this implementation with MindTouch's ChunkedStream. I'm really liking the idea of a chunked stream like this. Even once sgen is complete and deployed, a chunked stream will still be more efficient - not just in avoiding fragmentation, but also because reallocating a new, larger buffer, has overhead associated with copying the content whereas allocating a new chunk is a cheaper operation. Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] Will there be an easy way to install MonoDevelop 2.2 on Ubuntu
Jo Shields (Mono* maintainer for Ubuntu) will likely build MonoDevelop 2.2 packages when it comes out. I'm not sure what his repository url is, however. Jeff Mike Christensen wrote: Hi - When MonoDevelop 2.2 is released, will there be an easy way to install this on Ubuntu, or will the only way be to download the sources and build? I tried building from the trunk, however I ran into an endless maze of dependencies and gave up after about 2 hours. Thanks! Mike ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Trying to muster up some community interest
Have you considered using GMimeSharp as your message parser/decoder system? I notice it looks like you were planning on targeting Gtk so it might be a good fit. In any event, good luck with your project. Jeff Timothy Roberts wrote: Hey Mono-list, I've been working on a C#/Mono project lately, a NNTP protocol suite with library, CLI, and GUI applications. I've made some significant headway with a full NNTP specification library, which supports all common NNTP commands and yEnc encoding and decoding, with CRC32 support. However, I need members of the community to review the code, find bugs, send in patches, etc. That's where you come in. :-) If you have the time, please head over to http://nntpsharp.codeplex.com and download the latest source and binaries release of NNTPSharpLib. Test it out in your own projects. It's almost feature-complete, with yEnc header and trailer parsing still needing to be implemented. --Timothy Roberts(tr.parac...@gmail.com mailto:tr.parac...@gmail.com) ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
[Mono-dev] IOMAP fix for CreateFile()
In mono/io-layer/io.c's CreateFile() function, if the first call to
_wapi_open() fails due to the path being a directory, it is retried w/o
WRITE permissions. As far as I can tell, the fallback open() call should
really be a _wapi_open() so that IOMAP magic is applied just as it was
in the first attempt.
Attached is a patch to fix that (assuming I'm correct in what it should
be doing).
Jeff
Index: ChangeLog
===
--- ChangeLog (revision 140963)
+++ ChangeLog (working copy)
@@ -1,3 +1,8 @@
+2009-09-10 Jeffrey Stedfast f...@novell.com
+
+ * io.c (CreateFile): In the directory fallback case, use
+ _wapi_open() again so that IOMAP magic happens.
+
2009-08-18 Christian Hergert ch...@dronelabs.com
* wthreads.c: Roll back change for thread-id as it embraces
Index: io.c
===
--- io.c (revision 140963)
+++ io.c (working copy)
@@ -1547,7 +1547,7 @@
if (fd == -1 errno == EISDIR)
{
/* Try again but don't try to make it writable */
- fd = open(filename, flags ~(O_RDWR|O_WRONLY), perms);
+ fd = _wapi_open (filename, flags ~(O_RDWR|O_WRONLY), perms);
}
if (fd == -1) {
___
Mono-devel-list mailing list
Mono-devel-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Mono 2.4 - glib requirement issues
John M wrote: Thanks for the answer Miguel. We installed a private copy of glib 2.3.4 and are using that. FWIW, glib 2.3.4 is a development version (i.e. unstable), you might want to use glib = 2.4.0 (where the middle number is even rather than odd). Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] BinaryFormatter serialization accross .NET and Mono
Chris Howie wrote: On Fri, Feb 6, 2009 at 5:19 PM, schmmd mich...@schmitztech.com wrote: Hi, I am asking to learn if it is a goal of the Mono project to have consistent serialization across Mono and .Net with respect to the BinaryFormatter. Serialization is the same in some cases. For example, serialization of a byte array is the same in .NET and Mono. However, serialization of a struct containing a byte array is not. [snip] Is this a bug or is consistency between .NET and Mono not a goal? I do not believe it is a very high-priority goal. You might try the SoapFormatter class. It at least produces more readable output so if there is an issue it's easier to spot, and hence easier for the developers to fix. Things might be able to be serialized slightly differently and yet achieve the same results. However, I would suggest that, if things are not working using the binary formatter between .NET and Mono, then you should submit a bug. It's important that Mono and .NET are able to communicate properly. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Learning Mono just like .NET?
stevemcc wrote: So I am a college student interested in programming. I've done some programming, just a little bit in java, python, VB, and php. I noticed theres a lot of demand for .NET developers and I use linux most of the time. Is mono close enough .net that learning mono would prepare me for an eventually job that might require me to develop using .NET and windows. Or are they close, but still different beasts? As Michael Hutchinson has already said, they should be close enough to not matter. As a possible added bonus to learning Mono, more and more .NET development houses may begin to favor new recruits with Linux/Mono experience because Linux is becoming an important platform to support (not to mention Mac). We've already seen evidence of companies porting their .NET software over to Linux to run on Mono to either save themselves a lot of money (e.g. custom web apps), or to reach a wider market (selling their .NET apps for Mac/Linux). Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Isn't silverlight already crossbrowser ?
On Sat, 2008-12-06 at 19:04 -0800, joelafrite wrote: Hi. I don't much about Silverlight but I understand it's advertised as cross browser/platform by Microsoft. Is that by using Moonlight or does Silverlight come in several version itselfs ? Microsoft's Silverlight works on Windows and Mac only. Moonlight is the Linux port (that we'd like to make work on other platforms as well. Currently I think there's work being done on a Solaris port, for example). Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Is Mono fully compatible with .net framework?
On Mon, 2008-11-17 at 18:22 -0800, Katmai wrote: Is Mono fully compatible with .net framework now?If not,when will it be?(or will it be?) The runtime and compiler are, afaik, fully compat with Microsoft's. The implemented class libraries are compatible in so far as we know, however the .NET class libraries are extensive and always growing - if you want to know if your application will run using Mono's class libraries, you can run a tool called MoMA and it will provide you with any problematic areas. That should answer your question, although you were pretty vague so maybe not. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Parsing email
You could try GMime-Sharp (make sure it is version 2.4.x), it's a binding to the GMime C library which can parse MIME messages. I don't have any monodoc docs written for it, but it's pretty close to the C API which is documented at http://library.gnome.org/devel/gmime/stable/ You can download the source from http://ftp.gnome.org/pub/GNOME/sources/gmime/2.4/ Jeff On Mon, 2008-11-10 at 23:01 +, Jorge Bastos wrote: I'm sure they all work. In fact I've seen them already, but they are all commercial and that's not an option for me now. I have no problem to buy software, in fact I have a lot of it that I buyed, but right now it's not a good time to invest. -Original Message- From: Stéphane Zanoni [mailto:[EMAIL PROTECTED] Sent: segunda-feira, 10 de Novembro de 2008 17:58 To: Jorge Bastos; 'mono-list' Subject: Re: [Mono-list] Parsing email I believe the .msg files use the OLE 2 Compound Document format (proprietary format for MAPI?). A quick google for parse msg files or .net .msg format returned a couple results, nothing jumps out in the open/shared source realm. 100% managed (mono support) + purchasable source: http://www.independentsoft.de/msg/index.html http://www.aspose.com/categories/utility-components/aspose.network-for- .net/default.aspx java options (may be portable?) -- http://www.rgagnon.com/javadetails/java-0613.html If you do find something that works, let me know. Stphane Jorge Bastos [EMAIL PROTECTED] 11/10/2008 9:43 AM And about the .msg one's? -Original Message- From: Stphane Zanoni [mailto:[EMAIL PROTECTED] Sent: segunda-feira, 10 de Novembro de 2008 16:23 To: Jorge Bastos; 'mono-list' Subject: Re: [Mono-list] Parsing email eml are just MIME.822 formatted, the Lumisoft library has well developed Mime parsing capabilities. Jorge Bastos [EMAIL PROTECTED] 11/10/2008 7:29 AM Sorry not .eml, but .msg msoutlook 2003/2007 -Original Message- From: [EMAIL PROTECTED] [mailto:mono-list- [EMAIL PROTECTED] On Behalf Of Jorge Bastos Sent: segunda-feira, 10 de Novembro de 2008 14:26 To: 'mono-list' Subject: Re: [Mono-list] Parsing email Can this library parse the .eml messages from MSOutlook 2003/2007? -Original Message- From: [EMAIL PROTECTED] [mailto:mono-list- [EMAIL PROTECTED] On Behalf Of Damien Sent: segunda-feira, 10 de Novembro de 2008 14:22 To: Peter Hagen; mono-list@lists.ximian.com Subject: Re: [Mono-list] Parsing email Hi, The Lumisoft Net library : http://www.lumisoft.ee/lsWWW/Download/Downloads/Net/ Cheers, Damien 2008/11/10 Peter Hagen [EMAIL PROTECTED]: Hi all does anyone know if there is a existing library or function for parsing email messages? For example to easily get the attachments? cheers Peter Hagen ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Linux Equivalent to Visual Studio?
On Tue, 2008-10-28 at 18:46 -0700, Jerry Houston wrote: Please be assured that I'm not poking fun at mono, or Linux. I use Windows because I have to, and I use Linux because I want to. I'm a full-time .NET developer at work, and I really want mono to be to Linux what .NET is to Windows. However, I'm finding it really slow getting started with mono, because I haven't yet found anything I can use to create winform applications visually. Does such a thing exist for mono? I can build an application using Visual Studio, and then port it to mono without a lot of trouble, but every time I've tried to *start* an application using mono, I end up frustrated and quit. I just downloaded the latest version, thinking that I'd be pleasantly surprised this time, but monodevelop still seems like little more than a fairly smart code editor. Surely, there must be a mono equivalent to Visual Studio somewhere, but I haven't found it yet. Something with a visual designer, and Intellisense-type help, and so on. Can someone point me in the right direction? MonoDevelop has all of these things, including a visual designer, the only catch is that the designer is for Gtk# rather than Windows.Forms. There was some work a while back to implement a Windows.Forms designer for MonoDevelop as a Google Summer of Code project, but I don't remember what ever came of it. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Does Mono Support Partial Methods?
On Mon, 2008-10-20 at 20:08 -0700, used2could wrote: I am trying to use some classes that contain partial methods (not to be confused with partial classes) and i am getting a compilation Error: Feature 'partial methods' cannot be used because it is not part of the C# 2.0 language spec... Sounds like you might be using mcs instead of gmcs? Make sure you are using the gmcs compiler, mcs is for .NET 1.1 Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Packages for armv4t (OpenMoko)
On Fri, 2008-09-26 at 12:43 +0200, [EMAIL PROTECTED] wrote: Hi all, i tryed to make the NDesk.DBus package with the existing mono 1.2.5. i get the following error. are there newer packages as 1.2.5 avail or existing NDesk.Dbus for armv4t? I dont know how to(and dont want to) build mono from source. There's a 1.2.6 if you really want to stick to a 1.2.x release or you can try the latest 2.0 release candidate. Ideally you'll probably want 2.0 once it's released. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Name of the CSharp Shell
On Sun, 2008-09-21 at 17:39 -0400, Miguel de Icaza wrote: Hey folks, The name of the C# shell today is `csharp', but this is awkward to type, we have been considering renaming it, the options are: * csi - stands for C# interactive Pros: short, cute, easy to remember. Cons: popular TV show. I really like the name csi :) * imcs - interactive Mono C# compiler Pros: similar to `gmcs' and `mcs' Cons: hard to figure out it exists. eh :\ Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] Mono invoke the C++ method
The easiest solution is to wrap each of the c++ methods you need to call in a c-library (since Mono can P/Invoke C functions). Jeff On Fri, 2008-09-19 at 10:15 +0800, Taoye wrote: Hi , i am a C# programmer .Now we develop a project in Window , .NET Framework 2.0. We need to run it on mono.I know it can run well . But the project use a C++ DLL which develop in Window VC++6.0 . So the problems comes. When we run the project on mono , it cannt invoke the method in C++. Can you give me some ideas. Thanks for your help. Ye Tao 2008-09-19 ___ 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-list] Congrats!
Thanks! Your gratitude is appreciated by all of us ;-) Jeff On Thu, 2008-08-21 at 16:16 -0400, Abe Gillespie wrote: Hi Mono Monkeys, Just wanted to extend a huge congrats on the accomplishment with Linden Labs / Second Life. I know it's been a long time coming. Congratulations! -Abe ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] Moonlight 0.8 - Where are they Release Notes
On Mon, 2008-08-18 at 10:14 -0700, Daniel Morgan wrote: As Eno pointed out, where are the release notes for Moonlight 0.8? Is Moonlight good enough to watch video from http://www.nbcolympics.com/ on Linux now? Yea, we slipped up on updating the release notes file. Moonlight 0.8 only reliably works for Silverlight 1.0 sites so far. Silverlight 2.0 is still in Beta and so the API's are not yet frozen. Unfortunately, nbcolympics.com targets Silverlight 2.0 Beta and so Moonlight 0.8 will not work with it. Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] moonlight source in mono-2.0
The 2.1 assemblies are just stripped-down 2.0 assemblies and smcs is just gmcs built to target 2.1 None of the Silverlight classes are included inside Mono, so there shouldn't need to be any source-code removal at all. Also... why would the source code need to be removed at all in the source packages anyway? I'm totally lost. Jeff On Mon, 2008-08-18 at 20:15 +0100, Paul wrote: Hi, As part of the regular review of updated packages within fedora, we've found that the current 2.0 preview has the ability to build the moonlight assemblies (or not depending on the configure option). Moonlight is a forbidden item in Fedora, so I've built it with the configure script saying no thanks to moonlight. However, if the code is in the mono tarballs, it has to be removed when it's packaged up. Looking at the source, I can't see anything glaringly obvious, but from memory, its smcs, System.Net and anything 2.1 in gac. Can someone confirm this? Thanks TTFN Paul ___ 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] moonlight source in mono-2.0
On Mon, 2008-08-18 at 23:02 +0100, Paul wrote: Hi, Also... why would the source code need to be removed at all in the source packages anyway? I'm totally lost. Moonlight is a forbidden package by fedora (it is considered too dodgy due to the unholy pact between Novell and the borg - or something like that). This means that they will neither ship the source or the built binaries. Other examples : MP3, mpeg2 and anything else, which under US law, can land them in hot water. For that reason, anything to do with moonlight has to be removed from the sourceball, therefore the code to build smcs has to go. but the source to smcs is exactly the same as that of gmcs!?!? As Yoda might say: No sense does this logic make. TTFN Paul ___ 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] moonlight source in mono-2.0
Forgot to add that IMHO, ideally you should package (even if you don't ship by default) the 2.1 assemblies (and smcs?) so that it is easier for Fedora's end-users to get a working Moonlight should they choose to install it. W/o the 2.1 assemblies packaged, they would have to build their own Mono. Since Fedora deems Mono legally safe, there's no reason for Fedora not to at least package the 2.1 assemblies as there is absolutely nothing new in the code. Jeff On Mon, 2008-08-18 at 23:37 +0100, Paul wrote: Hi, For that reason, anything to do with moonlight has to be removed from the sourceball, therefore the code to build smcs has to go. but the source to smcs is exactly the same as that of gmcs!?!? As Yoda might say: No sense does this logic make. In that case, what does smcs really do that gmcs doesn't other than target the moonlight bits and pieces? Fedora are really picky over things, so I play it safe. Now, onto more pressing matters. How the hell do I get XIM to play nicely... TTFN Paul ___ 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] moonlight source in mono-2.0
On Mon, 2008-08-18 at 23:37 +0100, Paul wrote: Hi, For that reason, anything to do with moonlight has to be removed from the sourceball, therefore the code to build smcs has to go. but the source to smcs is exactly the same as that of gmcs!?!? As Yoda might say: No sense does this logic make. In that case, what does smcs really do that gmcs doesn't other than target the moonlight bits and pieces? Nothing. That's the entire difference. The difference between the 2.0 and the 2.1 assemblies is that the 2.1 assemblies are only a small subset of the 2.0 assemblies (wouldn't want Silverlight apps accessing the filesystem, or accessing many of the other System classes due to security concerns, so the 2.1 assemblies simply strip all of that out). Fedora are really picky over things, so I play it safe. Now, onto more pressing matters. How the hell do I get XIM to play nicely... TTFN Paul ___ 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] moonlight source in mono-2.0
On Mon, 2008-08-18 at 20:10 -0400, Geoff Norton wrote: Jeff, On Mon, 2008-08-18 at 19:58 -0400, Jeffrey Stedfast wrote: Forgot to add that IMHO, ideally you should package (even if you don't ship by default) the 2.1 assemblies (and smcs?) so that it is easier for Fedora's end-users to get a working Moonlight should they choose to install it. W/o the 2.1 assemblies packaged, they would have to build their own Mono. I'm fairly certain the plan here is for us to provide a .xpi that would run on a mono-less machine for SL2. In other words we're going to bundle the runtime bits we need in the .xpi. Aha, then disregard my previous mail ;-) Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] [Mono-dev] Moonlight 0.8 - Where are they Release Notes
On Mon, 2008-08-18 at 10:14 -0700, Daniel Morgan wrote: As Eno pointed out, where are the release notes for Moonlight 0.8? Is Moonlight good enough to watch video from http://www.nbcolympics.com/ on Linux now? Yea, we slipped up on updating the release notes file. Moonlight 0.8 only reliably works for Silverlight 1.0 sites so far. Silverlight 2.0 is still in Beta and so the API's are not yet frozen. Unfortunately, nbcolympics.com targets Silverlight 2.0 Beta and so Moonlight 0.8 will not work with it. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Moonlight 0.8 installation
It's fairly common for the following X.Org config option (in the Device section) to speed up Moonlight considerably: Option XaaNoOffscreenPixmaps true Jeff On Sat, 2008-08-16 at 21:15 -0400, Geoff Norton wrote: Filipe, From this message, figured libxcb-xlib was missing :) and went ahead and installed it. It's now working! Great However, i'm finding it extremely slow with the above example (haven't compared with running it with silverlight, though). If you're having performance problems please file a bug, and make sure to include the following information: Distro Architecture Video Card Video Card Driver Version XOrg version Xorg.conf The site / test in question thanks -g ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] Streaming with Moonlight/Silverlight vs Flash
On Fri, 2008-07-04 at 12:31 +0200, Federico Di Gregorio wrote: Hi everybody, I am about to choose the technology that will be used during the next two years to build the ESOF2010 web site (if you're curious about ESOF just check this year website, http://www.esof2008.org/). The project for 2010 includes a lot of streaming using preferably open protocols and open source software (mandatory on the servers, the client code should work at least on Windows and Linux and may use proprietary plugins). The tools to create the streams should be free software. We choosed to use feng as the streaming server because it supports RTP/RTSP and RTMP (now mostly free) with fallback on HTTP encapsulation and a lot of different non-proprietary codecs. Looking at MS website for Silverlight specifications it seems to me that streaming using the standard protocols cited above is not supported and that only MS codecs (that we don't like to use) can be used. We also investigated the possibility of writing a a managed RTP implementation but it seems that UDP is not supported in Silverlight. And even if we manage to have a working implementation of RTP it is not clear how you can decode the data and send it to the multimedia part of the framework. So it seems that while we would much better like to work with Mono/Moonlight and keep compatibility with Silverlight in the end will be forced to use Adobe Flash (RTMP + one of the non-free codecs: at least there are open source tools to do the authoring and the encoding conversions). I am writing to this list because I _hope_ we missed something and that Silverlight/Moonlight can be used to play streams using open protocols and codecs. I /believe/ that Silverlight/Moonlight can do RTSP and RSTPT (and obviously MMS), but I'm not 100% sure. I don't know too much about the protocols but at one point I was helping out with the media streaming code and noticed code to handle those 2 protocol schemes (although it's possible that the code just did immediate fallback to http or something, I can't remember). Hopefully one of the moonlight media people can answer this more accurately. However, in the end, if you aren't happy with the codecs that Silverlight supports, then even if we add support for your preferred codec into Moonlight, it wouldn't solve your codec problem. Out of curiosity, which codec do you plan on using? Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] Mono Licensing for Dedicated/Embedded Devices
There's probably a special email address for this but I don't know what it is off the top of my head. Emailing [EMAIL PROTECTED] and/or [EMAIL PROTECTED] would probably be enough. Jeff On Wed, 2008-06-18 at 12:43 -0700, Praerit wrote: hi, We are building a client/server solution using Mono and C#. Clients are going to be traditional applications installed on PC/Mac systems and eventually mobile devices also. Servers however may be installed on dedicated servers or embedded server devices. Both client servers are expected to be distributed widely. We do not expect or want our customers to be updating Mono bits independent of our solution - particularly for the servers. We will manage this update process for our customers to make sure the end-to-end solution doesn't break. From reading the FAQ, it seems like there are restrictions with using Mono when trying to build such a solution. Can someone please clarify or point me to the right person to talk to? thx! --pg ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Simple Console ReadLine problem
On Wed, 2008-03-26 at 12:13 -0700, radzfoto wrote:
I am a moderately experienced programmer especially in C and C++. I am just
learning C#.
I compiled the following under MS Visual Studio and it worked fine. However,
exact same code fails to read the string using ReadLine. Any ideas that
could help me out?
Here is the code:
using System;
namespace TestReadLine
{
class MainClass
{
public static void Main(string[] args)
{
string str;
Console.Write(Enter some text and press enter: );
str = Console.ReadLine();
Console.WriteLine(This is the text you entered: +
str);
Console.ReadLine
this line ^^^ needs to be changed to:
Console.ReadLine();
}
}
}
Thanks!
Radzfoto
before making the change, I was not able to compile your test program,
so I assume you must have had the (); in your original source.
That said, I wasn't able to reproduce your problem:
[EMAIL PROTECTED] src]$ gmcs readln.cs
[EMAIL PROTECTED] src]$ mono ./readln.exe
Enter some text and press enter: sdffasfasfasfasf
This is the text you entered: sdffasfasfasfasf
In case you were maybe using mcs instead of gmcs, I tried again:
[EMAIL PROTECTED] src]$ mcs readln.cs
[EMAIL PROTECTED] src]$ mono ./readln.exe
Enter some text and press enter: asfasdasdasdas
This is the text you entered: asfasdasdasdas
Can you explain your problem a bit more?
Jeff
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Hosting Silverlight applications on Apache2
The only problem I've come across hosting Silverlight apps over lighttpd (did feel like configuring Apache for local testing) was filename case sensitivity. IIS seems to be case insensitive while lighttpd (and likely Apache) are not. Also, fwiw, I didn't have to add any mime types. Jeff On Wed, 2008-01-30 at 18:16 +0100, Timothy Parez wrote: Hi, While this is not purely mono related, I think some of you might be able to help. I had a sample application which worked fine when hosted with IIS, but doesn't work when hosted on Apache2. http://www.itcrowd.be/silverlight/netrissilver.html The exception I get it this: The Page at http://www.itcrowd.be says: Silverlight error message ErrorCode: 1001 ErrorType: ParserError Message: AG_E_UNKNOWN_ERROR XamlFile: netris.xaml Line: 11 Position: 9 If you check out a sample by someone else, also hosted on Apache: http://www.zendurl.com/t/timheuer/v1_1/Default.html You get this The page at http://www.zendurl.com says: Silverlight error message Errorcode: 1001 ErrorType: ParserError Message: AG_E_UNKNOWN_ERROR XamleFile: Page.xaml Line: 9 Position: 2 Quite similar, so I'm guessing it's not my code. I've added the following Mime types to the Apache configuration, but it still doesn't work: AddType application/xaml+xml xaml AddType application/x-ms-application application AddType application/x-ms-xbap xbap AddType application/manifest manifest AddType application/octet-stream deploy AddType application/vnd.ms-xpsdocument xps AddType application/x-msdownload dll Since the Mono team is working on Moonlight, I assume they also know how to host these applications on Linux? It's all client side isn't it? So it shouldn't require Mono/ASP.NET right? (Even though it's an ASP.NET site, so even then it should still work) (Mono is installed on my server. It's version 1.2.6 --with-preview=yes --with-moonlight=yes) Thank you for any suggestions. Timothy. ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] [Mono-list] trying to compil moon
Do you have (the right version?) of (lib)expat-devel on your system? Jeff On Sun, 2007-12-30 at 13:20 +0100, Petit Eric wrote: hey folk somebody have the same error and know how to fix it ? make[2]: entrant dans le répertoire « /root/mono/moon/src » /bin/sh ../libtool --tag=CXX --mode=link g++ -g -O2 -o demo demo.o -lmoon -pthread -L/usr/lib/pkgconfig/../../lib -lgtk-x11-2.0 -lgdk-x11-2.0 -latk-1.0 -lgdk_pixbuf-2.0 -lpangocairo-1.0 -lcairo -lpango-1.0 -lgobject-2.0 -lgmodule-2.0 -lasound -lfreetype -lfontconfig -lXrandr -lmono -lpthread -lgthread-2.0 -lrt -lglib-2.0 -lavformat -lavcodec -lz -la52 -lamrnb -lamrwb -lfaac -lfaad -lmp3lame -lm -lnut -lx264 -lxvidcore -ldl -lX11 -lXext -ldirac_encoder -ldirac_decoder -lavutil g++ -g -O2 -o .libs/demo demo.o -pthread /root/mono/moon/src/.libs/libmoon.so /usr/lib/libdirectfb.so /usr/lib/libfusion.so /usr/lib/libdirect.so /usr/lib/libglitz.so -L/usr/lib/pkgconfig/../../lib /usr/lib/libgtk-x11-2.0.so /usr/lib/libgdk-x11-2.0.so /usr/lib/libatk-1.0.so /usr/lib/libgdk_pixbuf-2.0.so /usr/lib/libpangocairo-1.0.so /usr/lib/libpangoft2-1.0.so /usr/lib/libcairo.so -lpng12 /usr/lib/libpixman-1.so /usr/lib/libpango-1.0.so /usr/lib/libgobject-2.0.so /usr/lib/libgmodule-2.0.so /usr/lib/libasound.so /usr/lib/libfontconfig.so /usr/lib/libfreetype.so /usr/lib/libxml2.so /usr/lib/libXrandr.so /usr/lib/libXrender.so /usr/lib/libmono.so /usr/lib/libgthread-2.0.so -lpthread -lrt /usr/lib/libglib-2.0.so -lavformat -lavcodec -lz /usr/lib/liba52.so -lamrnb -lamrwb /usr/lib/libfaac.so /usr/lib/libfaad.so /usr/lib/libmp3lame.so -lm -lnut -lx264 -lxvidcore /usr/lib/libXext.so /usr/lib/libX11.so /usr/lib/libXdmcp.so -ldl /usr/lib/libXau.so /usr/lib/libdirac_encoder.so /usr/lib/libdirac_decoder.so -lavutil /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ErrorString' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetUserData' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_GetCurrentLineNumber' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetNamespaceDeclHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ParserFree' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetElementHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_GetCurrentColumnNumber' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetDoctypeDeclHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetCharacterDataHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ParserCreateNS' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_GetErrorCode' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_StopParser' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ParserCreate' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_Parse' collect2: ld returned 1 exit status make[2]: *** [demo] Erreur 1 make[2]: quittant le répertoire « /root/mono/moon/src » make[1]: *** [all-recursive] Erreur 1 make[1]: quittant le répertoire « /root/mono/moon » make: *** [all] Erreur 2 [EMAIL PROTECTED] moon]# ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] trying to compil moon
Do you have (the right version?) of (lib)expat-devel on your system? Jeff On Sun, 2007-12-30 at 13:20 +0100, Petit Eric wrote: hey folk somebody have the same error and know how to fix it ? make[2]: entrant dans le répertoire « /root/mono/moon/src » /bin/sh ../libtool --tag=CXX --mode=link g++ -g -O2 -o demo demo.o -lmoon -pthread -L/usr/lib/pkgconfig/../../lib -lgtk-x11-2.0 -lgdk-x11-2.0 -latk-1.0 -lgdk_pixbuf-2.0 -lpangocairo-1.0 -lcairo -lpango-1.0 -lgobject-2.0 -lgmodule-2.0 -lasound -lfreetype -lfontconfig -lXrandr -lmono -lpthread -lgthread-2.0 -lrt -lglib-2.0 -lavformat -lavcodec -lz -la52 -lamrnb -lamrwb -lfaac -lfaad -lmp3lame -lm -lnut -lx264 -lxvidcore -ldl -lX11 -lXext -ldirac_encoder -ldirac_decoder -lavutil g++ -g -O2 -o .libs/demo demo.o -pthread /root/mono/moon/src/.libs/libmoon.so /usr/lib/libdirectfb.so /usr/lib/libfusion.so /usr/lib/libdirect.so /usr/lib/libglitz.so -L/usr/lib/pkgconfig/../../lib /usr/lib/libgtk-x11-2.0.so /usr/lib/libgdk-x11-2.0.so /usr/lib/libatk-1.0.so /usr/lib/libgdk_pixbuf-2.0.so /usr/lib/libpangocairo-1.0.so /usr/lib/libpangoft2-1.0.so /usr/lib/libcairo.so -lpng12 /usr/lib/libpixman-1.so /usr/lib/libpango-1.0.so /usr/lib/libgobject-2.0.so /usr/lib/libgmodule-2.0.so /usr/lib/libasound.so /usr/lib/libfontconfig.so /usr/lib/libfreetype.so /usr/lib/libxml2.so /usr/lib/libXrandr.so /usr/lib/libXrender.so /usr/lib/libmono.so /usr/lib/libgthread-2.0.so -lpthread -lrt /usr/lib/libglib-2.0.so -lavformat -lavcodec -lz /usr/lib/liba52.so -lamrnb -lamrwb /usr/lib/libfaac.so /usr/lib/libfaad.so /usr/lib/libmp3lame.so -lm -lnut -lx264 -lxvidcore /usr/lib/libXext.so /usr/lib/libX11.so /usr/lib/libXdmcp.so -ldl /usr/lib/libXau.so /usr/lib/libdirac_encoder.so /usr/lib/libdirac_decoder.so -lavutil /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ErrorString' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetUserData' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_GetCurrentLineNumber' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetNamespaceDeclHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ParserFree' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetElementHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_GetCurrentColumnNumber' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetDoctypeDeclHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_SetCharacterDataHandler' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ParserCreateNS' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_GetErrorCode' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_StopParser' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_ParserCreate' /root/mono/moon/src/.libs/libmoon.so: undefined reference to `XML_Parse' collect2: ld returned 1 exit status make[2]: *** [demo] Erreur 1 make[2]: quittant le répertoire « /root/mono/moon/src » make[1]: *** [all-recursive] Erreur 1 make[1]: quittant le répertoire « /root/mono/moon » make: *** [all] Erreur 2 [EMAIL PROTECTED] moon]# ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Process.Exitcode is always less than 255
Programs under Linux always return a value between 0 and 255 (inclusively), they cannot return anything outside of that range. Jeff On Tue, 2007-12-11 at 22:35 -0800, sumesh0710 wrote: Hello I have a dotnet program which is compiled in .net 2003 in windows. I run this program in linux. It calls a c program in process.start() method from .net. Upon some error conditions the c program returns the error code to .net program. The problem is here. Suppose i return an error code 2053, and when i read this code from process.exitcode in .net, what i get is 5. on experiment i found that, what ever i return from c program, .net resets it after 255. In above case its like 256*8=2048. 2048+5=2053. and i get 5 in .net. In windows it works correctly.How to solve this problem?any idea? ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Mono C# versus Java 6 -server performance
I could be completely wrong, but my thinking is that the slowness is is the Mono implementations of various things (VM, classes, etc). There has been significant performance improvements done since 1.2.3, so you might try that and see how things have changed (Mono still may be slower than java6). Jeff On Thu, 2007-11-29 at 10:34 +, Robert Hulme wrote: Hi, I have recently been part of a discussion with colleagues about the merits of C# versus Java. I am of the opinion that C# is a superior language, but my colleagues questioned whether the performance is comparable on Linux (which is what we use here at EBI). I had a look on the Language Shootout, comparing Mono 1.2.5 with Java 6 -server. http://shootout.alioth.debian.org/gp4/benchmark.php?test=alllang=csharp In almost all the benchmarks Mono C# was significantly slower than Java 6. I downloaded some of the benchmarks to see if I could reproduce the results on my machine (Mono JIT compiler version 1.2.3.1 / Java 6). The results were the same. Twiddling with mono / mcs flags didn't seem to make a difference. I was wondering if anyone had any insights in to why this is, whether it is a CIL/C# issue or is specific to the Mono implementation, and whether there are any plans to bring Mono performance in line with Java performance. Many thanks, Robert Hulme ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] OpenGL/TAO
LibGL is the OpenGL API implementation, it does all the low-level rendering. Glut is the GL Utility Toolkit which is very simple and sits on top of LibGL as far as a programming API goes. Does basic window creation, keyboard/mouse events, etc. It's also a portable API. GLX is an X11-specific windowing/keyboard/mouse/ library similar to Glut but more powerful and lets you hook into other X11 API. Not sure how exactly to explain Xgl and AIGLX, they are basically OpenGL accelerated X11 variants. Maybe someone else can explain it better. Indirect rendering means that rendering is done in software. Direct rendering means that rendering is done by the hardware, and is usually faster than Indirect (assuming the drivers don't suck). Anyways, this is all basic (other than Xgl and AIGLX) knowledge you can find in just about any books on OpenGL. The Red Book goes into LibGL, Glut, GLX and the win32/mac versions. Jeff On Tue, 2007-10-30 at 08:00 -0500, Adrien Dessemond wrote: Hello all! Little Question : how is considered Tao (usable/incomplete/...) ? (its last update is 2 years old). I am also (desesperately) looking looking for a clear schema/paper/reference about how words GLX/LibGL/Glut/AIGLX/Xgl/(Indirect) Direct Rendering are glued together. I have a fuzzy image in head but too far to catch a precise (high-level) idea of what is going on behind the scenes when an OpenGL application like glxgears is running. Regards, Adrien ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] Silverlight/Moonlight questions
On Mon, 2007-10-22 at 17:12 +0200, Andrés G. Aragoneses [ knocte ] wrote: Miguel de Icaza escribió: Is MS Silverlight also a native library like Moonlight? Doesn't it have a managed API to be consumed by managed programs in order to use it outside of a browser? It is a COM component, and Microsoft has stated that they do not plan on supporting that API and reserve the right to break it as needed. Ok! All is clear now. I confess that I couldn't have thought of using COM nowadays... Then, if one wants to make a cross-platform gadget, the only way is depending on the browser component, right? correct Or creating a compatibility layer library which would call to COM in Windows and deal with Moonlight API library in Linux.. But that seems too complex for me. Would it be feasible to create it? I'm personally not sure about this... But now I am thinking of another solution: Silverlight is a subset of WPF right? yes and no. The API is a stripped down and slightly modified version of WPF, it's not actually compatible w/ WPF. Then, in order to make moonlight work, this small subset of libraries must be already developed on olive, right? They are in Olive, but not as part of the WPF implementation. Then, I suppose that if one creates a .NET application that uses WPF3.0 libraries without accessing non-Silverlight parts, this application should be crossplatform I guess? I don't think so, no. Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] Monodevelop doesn't start :(
rm -rf $prefix/lib/monodevelop and then do another make install afaik, this means that you have old data still in that directory and md gets confused. Jeff On Wed, 2007-07-25 at 14:22 +0300, =?UTF-8?Q?Szentp=E1li_J=E1nos?= wrote: When I try to start monodevelop I get the following error (of course the app quits instantly): --- START --- Unhandled Exception: Mono.Addins.Setup.InstallException: Application not found: IDE at MonoDevelop.Core.ApplicationService.StartApplication (System.String appId, System.String[] parameters) [0x0] at MonoDevelop.Startup.SharpDevelopMain.Main (System.String[] args) [0x0] --- END --- My questions: - does anyone know what is going on? - how can I make it not complain about MOZILLA_FIVE_HOME... for I have no Mozilla installed and I don't intend to. I currently set it to '/' János ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Bug in Environment.TickCount
from what I can tell reading the man page... you want to do this:
guint32 GetTickCount (void)
{
struct tms tms;
clock_t ticks;
if ((ticks = times (tms)) == -1)
return 0;
#ifdef LINUX_KERNEL_2_6_0_OR_LATER
ticks -= (UINT_MAX - 300) * CLOCKS_PER_SEC;
#endif
#define MSEC_PER_SEC 1000
#define CLOCKS_PER_MSEC (CLOCKS_PER_SEC / MSEC_PER_SEC)
return (guint32) (ticks / CLOCKS_PER_MSEC);
}
On Thu, 2007-07-19 at 22:38 -0300, Thiago M. Sayão wrote:
Hello!
Enrironment.TickCount is supposed to return the miliseconds since the
boot. For my surprise it was returning a negative number which should
only happen on the 47th day of uptime.
I did i fix (attached) but it doesn't seem 100% correct and i am not
sure if it would work after some days.
Since i'm coding a msn library and msn p2p protocol uses GetTickCount(),
it's messing things for me!
Can anyone take a look?
Thanks!
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Bug in Environment.TickCount
n/m, that's not quite right... CLOCKS_PER_SEC is incorrect...
try this:
#include sys/times.h
#include inttypes.h
#include limits.h
#include unistd.h
static uint32_t initial_ticks;
static uint32_t ticks_per_msec;
static uint32_t ticks_per_sec;
static void
TickCountInit (void)
{
ticks_per_sec = sysconf (_SC_CLK_TCK);
ticks_per_msec = ticks_per_sec 1000 ? ticks_per_sec / 1000 : 1;
#if defined (__linux__) /* FIXME: kernel 2.6.0 */
initial_ticks = (UINT_MAX - 300) * ticks_per_sec;
#else
initial_ticks = 0;
#endif
}
static uint32_t
GetTickCount (void)
{
struct tms tms;
clock_t ticks;
if ((ticks = times (tms)) == (clock_t) -1)
return 0;
ticks -= initial_ticks;
return (uint32_t) ticks / ticks_per_msec;
}
this gives 1/100th of a second accuracy +/-, so not quite millisecond
accuracy but... eh.
Jeff
On Fri, 2007-07-20 at 00:23 -0400, Jeffrey Stedfast wrote:
from what I can tell reading the man page... you want to do this:
guint32 GetTickCount (void)
{
struct tms tms;
clock_t ticks;
if ((ticks = times (tms)) == -1)
return 0;
#ifdef LINUX_KERNEL_2_6_0_OR_LATER
ticks -= (UINT_MAX - 300) * CLOCKS_PER_SEC;
#endif
#define MSEC_PER_SEC 1000
#define CLOCKS_PER_MSEC (CLOCKS_PER_SEC / MSEC_PER_SEC)
return (guint32) (ticks / CLOCKS_PER_MSEC);
}
On Thu, 2007-07-19 at 22:38 -0300, Thiago M. Sayão wrote:
Hello!
Enrironment.TickCount is supposed to return the miliseconds since the
boot. For my surprise it was returning a negative number which should
only happen on the 47th day of uptime.
I did i fix (attached) but it doesn't seem 100% correct and i am not
sure if it would work after some days.
Since i'm coding a msn library and msn p2p protocol uses GetTickCount(),
it's messing things for me!
Can anyone take a look?
Thanks!
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
___
Mono-list maillist - Mono-list@lists.ximian.com
http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] SmtpClient only supports ESMTP?
On Wed, 2007-05-30 at 12:12 +0200, Mads Bondo Dydensborg wrote: Hi there Regrettable I crashed into a smtp server today, that does not support ESMTP, but only SMTP (some old kind of exchange server, I think). It seems that System.Net.Mail.SmtpClient only supports ESMTP - is this true? Any plans for supporting SMTP also? I've just committed a fix to make it work for non-ESMTP servers (as well as a few fixes for other low-hanging fruit that I spotted while browsing thru the code). Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] test message
test to see if servers are back up ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Character encoding issues on Linux
As far as I know, Mono depends on libiconv (as either a standalone package or built into glibc - which is the case on Linux systems afaik) for charset conversion. The charset names used by libiconv however, may be named differently - so I'm sure Mono has a table to try and do the proper mapping (I've had to write tables like this in the past working on other projects). I think CodePage 1256 probably maps to Windows-CP1256, which on my particular Linux system w/ whatever libiconv I have could be mapped to either CP1256 or WINDOWS-1256 (iconv --list will give you a listing of all supported charset names) Perhaps I'll take a look later and see if I can't see something obvious (I haven't ever worked on the runtime so I dunno for sure how it does charset conversion, I could be way off) Jeff On Thu, 2007-05-17 at 11:08 +0100, Martin Brown wrote: Hi, We've got an application running on Mono on Linux. For the most part it runs very well; however every so often we get an exception: CodePage 1256 not supported The code page number varies depending on the encoding present in the file. From the MS page here http://msdn.microsoft.com/library/shared/deeptree/asp/rightframe.asp?dtcfg=/library/deeptreeconfig.xmlurl=/library/en-us/cpref/html/frlrfsystemtextencodingclassgetencodingtopic1.asp?frame=truehidetoc=false it looks like the MS .Net implementation relies on the OS for character encoding services. Is this the case for Mono too? If so, what needs to be available on Linux to perform this encoding? Thanks Martin Filtered by 3BClean from http://www.3bview.com ___ 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-list] Benchmark Results - Mono vs .Net
I've been playing around with this benchmark and have discovered that some of the reasons mono performs slower than MS is that (g)mcs doesn't optimize things like: r += i++; very well. If you instead write the code as: r += i; i++; the performance of the arithmetic loops under mono matches that of the MS .Net vm. As far as the other tests... I dunno yet :) Jeff On Fri, 2007-04-13 at 09:46 +0200, Henk Tiggelaar wrote: Below are the results of benchmarks under Mono 1.2.3.1 and Microsoft.NET 2.0. Benchmark code used can be found at http://www.tommti-systems.de/main-Dateien/reviews/languages/benchmarks.html Mono 1.2.3.1 results Int arithmetic elapsed time: 8094 ms Double arithmetic elapsed time: 12141 ms long arithmetic elapsed time: 26406 ms Trig elapsed time: 2749 ms IO elapsed time: 3204 ms Array elapsed time: 406 ms Exception elapsed time: 3719 ms HashMap elapsed time: 234 ms HashMaps elapsed time: 5265 ms HeapSort elapsed time: 609 ms Vector elapsed time: 9953 ms Matrix Multiply elapsed time: 71562 ms Nested Loop elapsed time: 10359 ms String Concat. (fixed) elapsed time: 578 ms Total C# benchmark time: 155279 ms .NET 2.0 results Int arithmetic elapsed time: 5812 ms Double arithmetic elapsed time: 7249 ms long arithmetic elapsed time: 16265 ms Trig elapsed time: 2281 ms IO elapsed time: 2499 ms Array elapsed time: 203 ms Exception elapsed time: 26687 ms HashMap elapsed time: 124 ms HashMaps elapsed time: 3999 ms HeapSort elapsed time: 531 ms Vector elapsed time: 9890 ms Matrix Multiply elapsed time: 33687 ms Nested Loop elapsed time: 22265 ms String Concat. (fixed) elapsed time: 359 ms Total C# benchmark time: 131851 ms ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] [mono-list]DLL Import errors
On Wed, 2007-04-18 at 10:28 -0600, Eric Morgan wrote: Frequently the problem is that the library is improperly linked; mono (effectively) loads libraries using dlopen(3) with RTLD_NOW set, so if your library depends on a symbol from another library but wasn't linked against that library, you will get errors loading it. You can also use http://www.jprl.com/dltest.c to test this scenario (see the comments for compile commands). If that still fails, you can reply to the mailing list with the Mono output attached. - Jon Still fails, and the output is generally the same from when I was using MONO_LOG_LEVEL=debug. Here's the output from dltest: [EMAIL PROTECTED] dltest]$ ll /home/rengeo/RenegadeGeophysics/SeismicStudioLinux64/lib/ total 992 -r-xr-xr-x 1 rengeo rengeo 339956 Apr 17 10:14 libSentinelKeys32.a -r-xr-xr-x 1 rengeo rengeo 315468 Feb 8 14:42 libSentinelKeys32.so -r-xr-xr-x 1 rengeo rengeo 328388 Apr 17 12:26 libSentinelKeysJDK.so [EMAIL PROTECTED] dltest]$ ./dltest /home/rengeo/RenegadeGeophysics/SeismicStudioLinux64/lib/libSentinelKeys32.so error loading library `/home/rengeo/RenegadeGeophysics/SeismicStudioLinux64/lib/libSentinelKeys32.so': /home/rengeo/RenegadeGeophysics/SeismicStudioLinux64/lib/libSentinelKeys32.so: cannot open shared object file: No such file or directory [EMAIL PROTECTED] dltest]$ The library is clearly there, but both mono and this dltest is saying that it doesn't exist. I've tried copying the library directly into the folder with dltest, and it still can't find it, even though is literally right there. Don't assume that it means that the file you gave it doesn't exist (I know, it is the most reasonable assumption to make). I suspect that Jonathan is correct, the problem is that your library depends on symbols not found in either the application nor in the library you are trying to load, and so dlopen() is failing - of course, this is just a guess :) The only way to know for sure is to read the source code for dlopen() and try to figure out what the problem is that way. Jeff ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Benchmark Results - Mono vs .Net
For the Mono test, was that run on Windows? Or Linux? I'm getting slower results (typically not by much except for the Double Arith benchmark) on Linux with an AMD Athlon XP 2600 (altho /proc/cpuinfo says I have a 3200?). Either way my processor is rated at 2200 GHz +/-. Maybe if I get some spare cycles I'll dig into the performance problems a bit (assuming someone else doesn't get there first). Jeff On Fri, 2007-04-13 at 09:46 +0200, Henk Tiggelaar wrote: Below are the results of benchmarks under Mono 1.2.3.1 and Microsoft.NET 2.0. Benchmark code used can be found at http://www.tommti-systems.de/main-Dateien/reviews/languages/benchmarks.html Mono 1.2.3.1 results Int arithmetic elapsed time: 8094 ms Double arithmetic elapsed time: 12141 ms long arithmetic elapsed time: 26406 ms Trig elapsed time: 2749 ms IO elapsed time: 3204 ms Array elapsed time: 406 ms Exception elapsed time: 3719 ms HashMap elapsed time: 234 ms HashMaps elapsed time: 5265 ms HeapSort elapsed time: 609 ms Vector elapsed time: 9953 ms Matrix Multiply elapsed time: 71562 ms Nested Loop elapsed time: 10359 ms String Concat. (fixed) elapsed time: 578 ms Total C# benchmark time: 155279 ms .NET 2.0 results Int arithmetic elapsed time: 5812 ms Double arithmetic elapsed time: 7249 ms long arithmetic elapsed time: 16265 ms Trig elapsed time: 2281 ms IO elapsed time: 2499 ms Array elapsed time: 203 ms Exception elapsed time: 26687 ms HashMap elapsed time: 124 ms HashMaps elapsed time: 3999 ms HeapSort elapsed time: 531 ms Vector elapsed time: 9890 ms Matrix Multiply elapsed time: 33687 ms Nested Loop elapsed time: 22265 ms String Concat. (fixed) elapsed time: 359 ms Total C# benchmark time: 131851 ms ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - Mono-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-dev] Suggestions abt layout designer
Looks like QLayout is similar to Gtk's layout widgets, no? Jeff On Wed, 2007-03-21 at 16:37 +0530, Sharique uddin Ahmed Farooqui wrote: On 3/21/07, Jensen Somers [EMAIL PROTECTED] wrote: Hi, On 3/20/07, Sharique uddin Ahmed Farooqui [EMAIL PROTECTED] wrote: Hi, Qt's layout designer is really good. I think we should include it in mono. MonoDevelop has a built in stetic designer, and there is still a tool called Glade (http://glade.gnome.org/). I also recall a tool named Gazpacho ( http://gazpacho.sicem.biz/). Both tools are GTK designers, so the created XML file can be used in C, C++ and C#. I think my questions is not clear. I'm talking abt Qt's layout controls not whole designer. These controls ease laying controls or grp of controls. This article will give more idea what exactly I'm saying. http://www.codeproject.com/csharp/QtLikeLayout.asp I think these tools are good enough. The only thing that might be useful is a tool like Linguist. Regards, Jensen ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list -- Sharique uddin Ahmed Farooqui (C++/C# Developer, IT Consultant) http://www.manageweb.info/sharique A revolution is about to begin. A world is about to change. And you and me are the initiator. ___ 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] building mono, monodevelop, etc. from SVN on Ubuntu Edgy
On Wed, 2007-03-14 at 16:31 -0400, joel reed wrote: Eric Miller wrote: I've been Googling for some advice on building mono and monodevelop from SVN on Ubuntu Edgy and haven't come up with much. http://www.mono-project.com/Compiling_Mono anything else you need to know? after installing [libgdiplus,] mono/mcs, you'll need to build gtk-sharp, gnome-sharp, gtkmozembed-sharp and gtksourceview-sharp from SVN in order to build monodevelop you'll probably find you'll need to install various gtk/gnome/etc -dev packages as you go, most of the time the configure script will help with that (or at worst check the config.log to see what is missing) for gtk-sharp, you'll want to use ./bootstrap-2.10 (I think Edgy has gtk+-2.10) and pass any --prefix, etc configure flags you want. Jeff ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-list] RE: Is it Mono safe?
On Thu, 2004-05-20 at 13:50 -0700, Rick Kitts wrote: On May 20, 2004, at 12:52 PM, Christopher McGinnis wrote: [...snip...] I would have to whole heartedly agree. Good technology is good technology no mater where it comes from. Would you have accepted a technology that, say, reduced processor heat by 2 orders of magnitude (whatever) from, say, Stalin's soviet union knowing what you know now about that mass murdering government? I think the question at hand is not, or at least should not be, about what is practical. The CLR (or .NET if you'd prefer) is pretty sweet no doubt about it. There is an ethical question regards adopting a MS technology that will help MS. MS is not friendly to my ability to choose. Enhancing their position in any way offers them increasing potential to diminish my ability to choose. you are confusing the issue. it's not the idea that can limit your freedom, it's the implementation that can. that said, your concerns becomes moot. Jeff -- Jeffrey Stedfast Evolution Hacker - Novell, Inc. [EMAIL PROTECTED] - www.novell.com ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] I give up
It falls under the LGPL Jeff On Thu, 2004-04-08 at 07:33, Timothy Parez wrote: What license does GTK fall under? ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
RE: [Mono-list] I give up
On Thu, 2004-04-08 at 09:33, Dan Winship wrote: If someone would actually take the time, Qt# can be used to write good applications. Then take the time. If you find any bugs in mono that keep Qt# from working, we'll gladly accept your patches. ...and bug reports Jeff -- Dan ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Use of mcs sources
you should just be able to get away with using the lex/yacc (or, I think mono uses jay) files. Jeff On Mon, 2003-11-10 at 08:09, [EMAIL PROTECTED] wrote: I want to write a C# editor tool as Visual Studio or SharpDevelop, but specialized on Mono and GTk# enviroment. I would need a C# parser and some .cs files of the mcs sources (/mcs directory) that are exactly that I'm looking for (using it, I could implement code analysis or autocompletion). I don't know if I could include some files of the sources on my own project (making clear where I find it), if I have to notify to the user to download the mcs sources to compile my project or if I cannot use it anyway r:-/ Do I have to change my project to GPL license? thanks! Manuel Castrillo -- Mensaje Original -- Subject: Re: [Mono-list] Use of mcs sources From: Jonathan Pryor [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Date: Mon, 10 Nov 2003 07:19:29 -0500 You didn't specify which parts of mcs source code you wanted to use; this could refer to either the mcs.exe (the C# compiler) sources, or to the class library sources (which are in the mcs CVS repository). Thus, I will clarify things. Licens s are a per-assembly construct, not a per-repository construct. So, in the mcs CVS repository, the mcs.exe assembly (sources in /mcs) is GPL. Many (but not all, IIRC) of the tools (sources in /tools) are also GPL, but some are MIT X11 (again, I RC). The class libraries (sources in /class) are MIT X11. In general, .exe's are GPL, and .dll's are MIT X11. The only exceptions to this rule are some .exe's (which are X11); all .dll's should be X11. Of course, this doesn't affect fair us of the source code. You should be able to use GPL source code to determine how it works, and use that knowledge in new code; literal copying generally isn't permitted. Fair use might allow some literal copying, though; this gets into a lot of g ay areas of copyright law, so it's probably best to just avoid it. Or, write GPL code and use the existing mcs.exe code as much as you want. - Jon On Mon, 2003-11-10 at 02:50, [EMAIL PROTECTED] wrote: Hi list! I'm a bit confused bout licenses... Can I use parts of the mcs source code (GPL License) in a BSD proyect? Can I simply use parts of mcs sources? Thanks! PRUEBA TISCALI ADSL HASTA 6 MESES SIN CUOTAS FIJAS, SIN COMPROMISO PONNOS A PRUEBA http://acceso.tiscali.es/pag-acceso-adsl-Prueba.jsp?did=adp-7530007 ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono- ist PRUEBA TISCALI ADSL HASTA 6 MESES SIN CUOTAS FIJAS, SIN COMPROMISO PONNOS A PRUEBA http://acceso.tiscali.es/pag-acceso-adsl-Prueba.jsp?did=adp-7530007 ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list -- Jeffrey Stedfast Evolution Hacker - Ximian, Inc. [EMAIL PROTECTED] - www.ximian.com ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] ThreadPool, why?
the overhead of launching a new thread can be quite substantial for some operating systems, others might not be quite so bad, but even so - re-using threads is always a lot less overhead than spawning news ones. Jeff On Wed, 2003-08-27 at 21:31 +0200, Sergio Blanco Cuaresma wrote: Hello! I don't understand the benefits of using a ThreadPool instead of launching simple Threads, I have read in several articles that its just more efficient, but I don't know why? What are the advantages of using ThreadPools ? Thanks! -- Jeffrey Stedfast Evolution Hacker - Ximian, Inc. [EMAIL PROTECTED] - www.ximian.com ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
Re: [Mono-list] Build fails on OS X
you probably need to install libiconv from ftp://ftp.gnu.org/pub/gnu/libiconv/ (I think I got the url path right, if not - it's somewhere on ftp.gnu.org) Jeff On Wed, 2003-03-12 at 16:39, James Fitzsimons wrote: Hi all, I am trying to build mono on my OS X (10.2) machine at home. Unfortunately I ran into problems before I even got to Mono! I got the following error trying to build glibc: snip checking for indent... indent checking for perl... /usr/bin/perl checking for iconv_open... no checking for libiconv_open in -liconv... no checking for iconv_open in -liconv... no configure: error: *** No iconv() implementation found in C library or libiconv bash-2.05a$ /snip I was wondering if anyone else had struck this problem on OS X and could suggest a work around? Thanks very much, James Fitzsimons ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list -- Jeffrey Stedfast Evolution Hacker - Ximian, Inc. [EMAIL PROTECTED] - www.ximian.com ___ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
