ming-users — Fun with Ming- the Ming users mailing list

Re: [Ming-users] PHP module spams httpd error.log

[Stuart A. <and...@ne...>]

On Tue, 6 Dec 2011, Sandro Santilli wrote:

> On Tue, Dec 06, 2011 at 02:14:04PM -0500, Stuart Anderson wrote:
>> On Tue, 6 Dec 2011, Ed Hynan wrote:
>>
>>> BTW, bugs.libming.org hasn't sent a confirm email. It's been a few
>>> hours.  Maybe it's broken.
>>
>> Sigh. That's on me. I'll take a look.
>
> Stuart: time to move bugs to github ?

Whatever the group wants to do it OK w/ me. My only concern is wether people
will find the new interface easier to use, or wether they find bugzilla to
be a familiar interface which they are used to using.

In this particular situation, everything worked properly, there was just some
extended delay between the mail servers.


                                 Stuart

Stuart R. Anderson                               and...@ne...
Network & Software Engineering                   http://www.netsweng.com/
1024D/37A79149:                                  0791 D3B8 9A4C 2CDC A31F
                                                  BD03 0A62 E534 37A7 9149

Re: [Ming-users] PHP module spams httpd error.log

[Sandro S. <st...@ke...>]

On Tue, Dec 06, 2011 at 02:14:04PM -0500, Stuart Anderson wrote:
> On Tue, 6 Dec 2011, Ed Hynan wrote:
> 
> > BTW, bugs.libming.org hasn't sent a confirm email. It's been a few
> > hours.  Maybe it's broken.
> 
> Sigh. That's on me. I'll take a look.

Stuart: time to move bugs to github ?
Ed already found his way there: https://github.com/libming/libming/issues/3

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] PHP module spams httpd error.log

[Stuart A. <and...@ne...>]

On Tue, 6 Dec 2011, Ed Hynan wrote:

> BTW, bugs.libming.org hasn't sent a confirm email. It's been a few
> hours.  Maybe it's broken.

Sigh. That's on me. I'll take a look.


                                 Stuart

Stuart R. Anderson                               and...@ne...
Network & Software Engineering                   http://www.netsweng.com/
1024D/37A79149:                                  0791 D3B8 9A4C 2CDC A31F
                                                  BD03 0A62 E534 37A7 9149

Re: [Ming-users] PHP module spams httpd error.log

[Ed H. <ev...@ve...>]

On 12/06/2011 02:14 PM, Stuart Anderson wrote:
> On Tue, 6 Dec 2011, Ed Hynan wrote:
> 
>> BTW, bugs.libming.org hasn't sent a confirm email. It's been a few
>> hours.  Maybe it's broken.
> 
> Sigh. That's on me. I'll take a look.
> 

Got it, just a little while hefore your message.  Thanks.

- Ed

Re: [Ming-users] PHP module spams httpd error.log

[Ed H. <ev...@ve...>]

On 12/06/2011 11:14 AM, Sandro Santilli wrote:
> If you have an account on github filing a bug there would also work
> (possibly better).

<g> You'll find a new issue at github.  You'll also see I had
trouble posting a patch.  In the second comment I think I got
it right.

BTW, bugs.libming.org hasn't sent a confirm email. It's been a few
hours.  Maybe it's broken.

- Ed

Re: [Ming-users] PHP module spams httpd error.log

[Sandro S. <st...@ke...>]

On Tue, Dec 06, 2011 at 10:21:19AM -0500, Ed Hynan wrote:
> On 12/06/2011 06:30 AM, Sandro Santilli wrote:
> > 
> > Well, libming is clearly not completely thread safe, so let's not
> > bother about that right now. Rather, the apache case seems
> > interesting. That's simply a matter of calling Ming_init multiple
> > times from within the same process. I'd think it is the PHP module
> > responsibility to avoid that.
> 
> OK.  I've been using the ming module with a guard against calling
> Ming_init() again, and so far it's been fine.  I want to look at
> php_ext/ming.c closer before submitting a patch.
> 
> I'm creating an account with bugs.libming.org now.  (Actually
> the first bug I'll post there is for tests/actionscript
> failing on OpenBSD. A BSD vs. GNU make difference I think.)

If you have an account on github filing a bug there would also work
(possibly better).

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] PHP module spams httpd error.log

[Ed H. <ev...@ve...>]

On 12/06/2011 06:30 AM, Sandro Santilli wrote:
> 
> Well, libming is clearly not completely thread safe, so let's not
> bother about that right now. Rather, the apache case seems
> interesting. That's simply a matter of calling Ming_init multiple
> times from within the same process. I'd think it is the PHP module
> responsibility to avoid that.

OK.  I've been using the ming module with a guard against calling
Ming_init() again, and so far it's been fine.  I want to look at
php_ext/ming.c closer before submitting a patch.

I'm creating an account with bugs.libming.org now.  (Actually
the first bug I'll post there is for tests/actionscript
failing on OpenBSD. A BSD vs. GNU make difference I think.)

- Ed

Re: [Ming-users] PHP module spams httpd error.log

[Sandro S. <st...@ke...>]

On Mon, Dec 05, 2011 at 06:27:46PM -0500, Ed Hynan wrote:

> Running the script with the command line php is fine.  It's
> an issue with libming being reused, which happens in each Apache
> child.  It doesn't matter how many; sorry I caused confusion
> thinking out loud in my last message.
> 
> The ming module init zm_activate_ming() gets called for each
> php script (regardless of whether the script uses ming).  And
> each time zm_activate_ming() calls Ming_init() in the library.
> 
> Suppressing the warning in error.log is easy in either the
> ming php module, or in the ming library (move the static int
> in Ming_useSWFVersion() to global||file scope and have Ming_init()
> assign it zero).  I guess either would be safe if Apache or
> the like uses several distinct processes; but what if a server
> is using threads?  How much state is kept in globals?

Well, libming is clearly not completely thread safe, so let's not
bother about that right now. Rather, the apache case seems
interesting. That's simply a matter of calling Ming_init multiple
times from within the same process. I'd think it is the PHP module
responsibility to avoid that.

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] PHP module spams httpd error.log

[Ed H. <ev...@ve...>]

On 12/05/2011 09:31 AM, Sandro Santilli wrote:
> On Sat, Dec 03, 2011 at 12:26:32PM -0500, Ed Hynan wrote:
> 
>> EH: zm_activate_ming -- first init
>> EH: zm_activate_ming -- first init
>> EH: zm_activate_ming -- first init
>> EH: zm_activate_ming -- first init
>> EH: zm_activate_ming -- first init
> 
> I guess these are coming from the different apache childs.

You're right of course.

> In order to reduce your testing framework see if everything
> goes smooth outside of apache.
> 

Running the script with the command line php is fine.  It's
an issue with libming being reused, which happens in each Apache
child.  It doesn't matter how many; sorry I caused confusion
thinking out loud in my last message.

The ming module init zm_activate_ming() gets called for each
php script (regardless of whether the script uses ming).  And
each time zm_activate_ming() calls Ming_init() in the library.

Suppressing the warning in error.log is easy in either the
ming php module, or in the ming library (move the static int
in Ming_useSWFVersion() to global||file scope and have Ming_init()
assign it zero).  I guess either would be safe if Apache or
the like uses several distinct processes; but what if a server
is using threads?  How much state is kept in globals?

- Ed

Re: [Ming-users] PHP module spams httpd error.log

[Sandro S. <st...@ke...>]

On Sat, Dec 03, 2011 at 12:26:32PM -0500, Ed Hynan wrote:

> EH: zm_activate_ming -- first init
> EH: zm_activate_ming -- first init
> EH: zm_activate_ming -- first init
> EH: zm_activate_ming -- first init
> EH: zm_activate_ming -- first init

I guess these are coming from the different apache childs.
In order to reduce your testing framework see if everything
goes smooth outside of apache.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] PHP module spams httpd error.log

[Ed H. <ev...@ve...>]

On 12/03/2011 03:59 AM, Sandro Santilli wrote:
> On Fri, Dec 02, 2011 at 09:10:48PM -0500, Ed Hynan wrote:
>>
>> Using Ming PHP module with Apache on Ubuntu 10.10
>> (package version; Ming 0.4.3).
>>
>> This:
>>
>> WARNING: changing SWF target version during a run
>>          might result in malformed SWF output.
>>          You don't have to worry if you're careful about
>>          not mixing different version blocks in a movie.
>>
>>
>> is written again and again to Apache's error.log.  My
>> code is not changing the version anywhere; it's set
>> with SWFMovie() and ming_useswfversion(), same in both cases.
>>
>> I haven't found a way to stop it. Any ideas?
> 
> Idea for your case: don't use both calls, 1 is enough.

Well, I was being brief in that message. I didn't start
with both, but then tried anything I could think of.

> 
> Idea for improving Ming: send a patch checking for actual
> version and not raising the warning if they match...

I had looked at the 0.4.4 code before subscribing here,
but had to ask before going too far; maybe it was a known
issue. (And I have the same problem as everyone else: time.)

So I've poked around.  In php_ext/ming.c, zm_activate_ming()
(expands from PHP_RINIT_FUNCTION(ming)) eventually gets
called more than once, and each time it calls Ming_init()
in libming, which assigns SWF_versionNum the default 5, and
so the warning is triggered in libming's Ming_useSWFVersion(int).

Why does it get called more than once?  This is not my area.
Maybe Apache, or maybe PHP, load/unload an .so at first, then
keep it loaded after some amount of use? Not likely; more
likely something to do with Apache's pre-forked processes, but
then why is one initialized ming.so shared? I don't know.

Whatever the cause, the ming php module, and surely the other
language modules, will have to guard against re-initialization.

To explain, here's a small patch (_not_ a proposed patch):

--- ming.c-orig	2011-10-26 02:33:18.000000000 -0400
+++ ming.c	2011-12-03 11:13:13.000000000 -0500
@@ -6199,6 +6199,17 @@

 PHP_RINIT_FUNCTION(ming) /* {{{ */
 {
+	/* EH: Sat Dec  3 09:41:51 EST 2011 */
+	static int init = 0;
+	setlinebuf(stderr);
+	if (init == 0) {
+		init++;
+		fprintf(stderr,"EH: %s -- first init\n", __FUNCTION__);
+	} else {
+		init++;
+		fprintf(stderr,"EH: %s -- init call %d\n", __FUNCTION__, init);
+		return SUCCESS;
+	}
 	/* XXX - this didn't work so well last I tried.. */

 	if (Ming_init() != 0) {

<end patch>

With the static int guarding re-init, and line buffered
messages, I see this in error.log (over several runs):

EH: zm_activate_ming -- first init
EH: zm_activate_ming -- first init
EH: zm_activate_ming -- first init
EH: zm_activate_ming -- first init
EH: zm_activate_ming -- first init
EH: zm_activate_ming -- init call 2
EH: zm_activate_ming -- init call 2
EH: zm_activate_ming -- init call 2
EH: zm_activate_ming -- init call 2
EH: zm_activate_ming -- init call 2
EH: zm_activate_ming -- init call 3
EH: zm_activate_ming -- init call 3
[Sat Dec 03 11:39:46 2011] [error] [client 192.168.1.187] File does not
exist: /var/www/favicon.ico
EH: zm_activate_ming -- init call 3
EH: zm_activate_ming -- init call 4

BTW, those lines begin after a fresh restart of Apache.
Also, note that the warning I asked about does not appear;
returns w/o Ming_init() after the 1st call.

My next question is: can libming handle reuse, possibly
from any number of scripts doing different things, maybe
setting different versions?  Could the simple guard
against reinit work in the real world?

- Ed




> 
> --strk; 
> 
>   ()   Free GIS & Flash consultant/developer
>   /\   http://strk.keybit.net/services.html
> 
> --

Re: [Ming-users] PHP module spams httpd error.log

[Sandro S. <st...@ke...>]

On Fri, Dec 02, 2011 at 09:10:48PM -0500, Ed Hynan wrote:
> 
> Using Ming PHP module with Apache on Ubuntu 10.10
> (package version; Ming 0.4.3).
> 
> This:
> 
> WARNING: changing SWF target version during a run
>          might result in malformed SWF output.
>          You don't have to worry if you're careful about
>          not mixing different version blocks in a movie.
> 
> 
> is written again and again to Apache's error.log.  My
> code is not changing the version anywhere; it's set
> with SWFMovie() and ming_useswfversion(), same in both cases.
> 
> I haven't found a way to stop it. Any ideas?

Idea for your case: don't use both calls, 1 is enough.

Idea for improving Ming: send a patch checking for actual
version and not raising the warning if they match...

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

[Ming-users] PHP module spams httpd error.log

[Ed H. <ev...@ve...>]

Using Ming PHP module with Apache on Ubuntu 10.10
(package version; Ming 0.4.3).

This:

WARNING: changing SWF target version during a run
         might result in malformed SWF output.
         You don't have to worry if you're careful about
         not mixing different version blocks in a movie.


is written again and again to Apache's error.log.  My
code is not changing the version anywhere; it's set
with SWFMovie() and ming_useswfversion(), same in both cases.

I haven't found a way to stop it. Any ideas?

- Ed

Re: [Ming-users] Ming-0.4.4 released

[Sandro S. <st...@ke...>]

On Fri, Oct 28, 2011 at 10:56:12AM +0200, Sandro Santilli wrote:

> The one for ActionScriptTest.c sounds suspicious as
> the vsaprintf.c implementation is already in decompile.c
> and I don't see it used from ActionScriptTest itself.

Pushed a provisional fix for your ActionScriptTest.c
issue with 815f18295602dfabfad53b754fbcaad91e2198bc

> I still have to check the configure.in patch.

This was also committed.

Please test and report back. Thanks !

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] Ming-0.4.4 released

[Sandro S. <st...@ke...>]

On Fri, Oct 28, 2011 at 09:38:29AM +0200, Sandro Santilli wrote:
> On Fri, Oct 28, 2011 at 04:08:02PM +0900, Shigeharu TAKENO wrote:

> > After applied mypatch, all test passed in "make check". The 
> > modified patch is the following:
> 
> Ok, sounds better, altought I'd prefer multiple patches
> for multiple issues (and ideally also one ticket for each).
> 
> Tickets go to: http://bugs.libming.org

I've pushed your patches for read.c and parser.c
The one for ActionScriptTest.c sounds suspicious as
the vsaprintf.c implementation is already in decompile.c
and I don't see it used from ActionScriptTest itself.
Maybe the problem is with makeswf.c using it w/out
implementing it, in wich case linking decompile.c in
would fix the linking for ActionScriptTest.c too.

I still have to check the configure.in patch.

--strk;

Re: [Ming-users] Ming-0.4.4 released

[Sandro S. <st...@ke...>]

On Fri, Oct 28, 2011 at 09:38:29AM +0200, Sandro Santilli wrote:

> There's a test/dotests.pl script which is the one invoked as a wrapper
> for most of the testsuite (all but test/actionscript). That script takes
> care of generating the SWF and comparing it with a reference SWF.
> It does it for each enabled language binding.
> The way it generates the SWF is custom for each binding.
> See for example line 157 of the script. There, the commandline to invoke
> a PHP test is generated. You could do something similar for python and 
> for perl and what.
> 
> I gave it a first go, attached is my patch for it.
> See if it works for you.

FYI: I've pushed the change upstream,
togheter with another enhancing testsuite running when you have another
Ming installed.

Patches here:
http://github.com/libming/libming/commit/1ade2aff0865940aaea2aef67c10b5bd7d592394.patch
http://github.com/libming/libming/commit/f55456c1b799bf49547eb10790e9b5f0c78cb5b5.patch

Tests welcome.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] Ming-0.4.4 released

[Sandro S. <st...@ke...>]

On Fri, Oct 28, 2011 at 04:08:02PM +0900, Shigeharu TAKENO wrote:
> shige 10/28 2011
> ----------------
> 
> Thanks for your reply.
> 
> Sandro Santilli <st...@ke...> wrote:
> | > 2) I needed to change paths of the 1st line of test scripts:
> | > 
> | >  #!/usr/bin/python, #!/usr/bin/perl, #!/usr/bin/php
> | > 
> | > to the proper paths
> | 
> | We want to find a way to recuce maintainance cost, maybe
> | 
> | #!/usr/bin/env python
> | 
> | and similar would work for you ?
> 
> This works for me too. It is a nice solution.

I just asked an openbsd user about this and he mentioned there's no
"python" binary there, just "python2.6" or "python2.7"
This means the only way for it to work there is having it set
at ./configure time (--with-python=xxx) and thus either substituted
in each and every test file (too much work, IMHO :) or only substituted
in the dotests.pl file, which is the script that actually invokes
each and every test. See below.

> | Alternatively we should work something out from the dotests.pl
> | runner, so that the path to interpreter is only substituted
> | in that single place.
> | 
> | Could you please file a ticket for this on bugzilla ?
> 
> Sorry, I can't understand your proposal.

There's a test/dotests.pl script which is the one invoked as a wrapper
for most of the testsuite (all but test/actionscript). That script takes
care of generating the SWF and comparing it with a reference SWF.
It does it for each enabled language binding.
The way it generates the SWF is custom for each binding.
See for example line 157 of the script. There, the commandline to invoke
a PHP test is generated. You could do something similar for python and 
for perl and what.

I gave it a first go, attached is my patch for it.
See if it works for you.


> | > 3) I used the patch below at the compilation. Especially, I think
> | > it is serious for util/parser.c.
> | 
> | Is there any test failing w/out your patch and working with ?
> 
> Sorry, I made a mistake. A patch for util/parser.c is not correct.
> I attached correct version of it below.
> 
> I tested again on my Solaris 9. Since the OS does not have 
> vasprintf(), doing "CCLD ActionScriptTest" is faild without 
> my patch for actionscript/ActionScriptTest.c. 
> 
> Without the patch for util/read.c about endian, "make check" says:
> 
> > Compiling `forin.as.pp'... successfully compiled 174 bytes bytecode.
> > ../../util/listswf forin.swf | diff - ./forin.ref
> > Testing legacy_functions with swfversion 6
> > Preprocessing ./legacy_functions.as... 168c168
> > <      [000]   Float: 0.000000
> > ---
> > >      [000]   Float: 1.000000
> > 175c175
> > <      [000]   Float: 0.000000
> > ---
> > >      [000]   Float: 1.000000
> > done.
> > .............
> > FAIL: ActionScriptTest
> > ==================
> > 1 of 1 test failed
> > ==================
> 
> Without the patch for util/parser.c, "make check" says:
> 
> > Executing './test03 .' in /home/users/shige/tmp/work6/ming-0.4.4/test/Font
> > 64c64
> > <    GlyphIndex[0] = 5ee90    GlyphAdvance[0] = 0086
> > ---
> > >    GlyphIndex[0] = 0000    GlyphAdvance[0] = 0086
> > test03 failed (c). Problem comparing against ./test03.ref
> 
> for test03. This may be a problem for initialization of 
> GlyphIndex[0].
> 
> After applied mypatch, all test passed in "make check". The 
> modified patch is the following:

Ok, sounds better, altought I'd prefer multiple patches
for multiple issues (and ideally also one ticket for each).

Tickets go to: http://bugs.libming.org

Thanks again.

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] Ming-0.4.4 released

[Shigeharu T. <sh...@ie...>]

shige 10/28 2011
----------------

Thanks for your reply.

Sandro Santilli <st...@ke...> wrote:
| > 2) I needed to change paths of the 1st line of test scripts:
| > 
| >  #!/usr/bin/python, #!/usr/bin/perl, #!/usr/bin/php
| > 
| > to the proper paths
| 
| We want to find a way to recuce maintainance cost, maybe
| 
| #!/usr/bin/env python
| 
| and similar would work for you ?

This works for me too. It is a nice solution.


| Alternatively we should work something out from the dotests.pl
| runner, so that the path to interpreter is only substituted
| in that single place.
| 
| Could you please file a ticket for this on bugzilla ?

Sorry, I can't understand your proposal.


| > 3) I used the patch below at the compilation. Especially, I think
| > it is serious for util/parser.c.
| 
| Is there any test failing w/out your patch and working with ?

Sorry, I made a mistake. A patch for util/parser.c is not correct.
I attached correct version of it below.

I tested again on my Solaris 9. Since the OS does not have 
vasprintf(), doing "CCLD ActionScriptTest" is faild without 
my patch for actionscript/ActionScriptTest.c. 

Without the patch for util/read.c about endian, "make check" says:

> Compiling `forin.as.pp'... successfully compiled 174 bytes bytecode.
> ../../util/listswf forin.swf | diff - ./forin.ref
> Testing legacy_functions with swfversion 6
> Preprocessing ./legacy_functions.as... 168c168
> <      [000]   Float: 0.000000
> ---
> >      [000]   Float: 1.000000
> 175c175
> <      [000]   Float: 0.000000
> ---
> >      [000]   Float: 1.000000
> done.
> .............
> FAIL: ActionScriptTest
> ==================
> 1 of 1 test failed
> ==================

Without the patch for util/parser.c, "make check" says:

> Executing './test03 .' in /home/users/shige/tmp/work6/ming-0.4.4/test/Font
> 64c64
> <    GlyphIndex[0] = 5ee90    GlyphAdvance[0] = 0086
> ---
> >    GlyphIndex[0] = 0000    GlyphAdvance[0] = 0086
> test03 failed (c). Problem comparing against ./test03.ref

for test03. This may be a problem for initialization of 
GlyphIndex[0].

After applied mypatch, all test passed in "make check". The 
modified patch is the following:

----- From here -----
diff -uN ming-0.4.4/configure.in.ORG ming-0.4.4/configure.in
--- ming-0.4.4/configure.in.ORG	Wed Oct 26 16:29:42 2011
+++ ming-0.4.4/configure.in	Fri Oct 28 08:55:52 2011
@@ -150,6 +150,8 @@
 		AC_MSG_ERROR([Can't build tcl extension, as tcl executable could not be found])
 	fi
 
+tclbindir=`dirname $TCL`
+
 tcllibdirs_default="\
 $prefix/lib/itcl \
 $prefix/lib \
@@ -157,7 +159,8 @@
 $HOME/lib \
 /usr/local/lib \
 /usr/lib64 \
-/usr/lib"
+/usr/lib \
+`dirname $tclbindir`/lib"
 
 	for i in $tcllibdirs_default; do
 	  for suf in 8.3 8.4 ""; do
@@ -193,7 +196,8 @@
 $HOME/include \
 /usr/local/include \
 /usr/include/tcl \
-/usr/include"
+/usr/include \
+`dirname $tclbindir`/include"
 
 	for i in $tclincdirs_default; do
 	   if test -f $i/tcl.h; then
diff -uN ming-0.4.4/test/actionscript/ActionScriptTest.c.ORG ming-0.4.4/test/actionscript/ActionScriptTest.c
--- ming-0.4.4/test/actionscript/ActionScriptTest.c.ORG	Wed Oct 26 15:33:18 2011
+++ ming-0.4.4/test/actionscript/ActionScriptTest.c	Fri Oct 28 12:36:01 2011
@@ -41,6 +41,46 @@
 #include <limits.h>
 #include <makeswf.h>
 
+#ifndef HAVE_VASPRINTF
+/* Workaround for the lack of vasprintf()
+ * As found on: http://unixpapa.com/incnote/stdio.html
+ * Seems to be Public Domain
+ */
+int
+vasprintf(char **ret, const char *format, va_list ap)
+{
+	va_list ap2;
+	int len = 100;        /* First guess at the size */
+
+	if ((*ret = (char *) malloc(len)) == NULL)
+	{
+	    return -1;
+	}
+	while (1)
+	{
+	    int nchar;
+	    va_copy(ap2, ap);
+	    nchar= vsnprintf(*ret, len, format, ap2);
+	    if (nchar > -1 && nchar < len)
+	    {
+		return nchar;
+	    }
+	    if (nchar > len)
+	    {
+		len= nchar+1;
+	    } else
+	    {
+		len*= 2;
+	    }
+	    if ((*ret = (char *) realloc(*ret, len)) == NULL)
+	    {
+		free(*ret);
+		return -1;
+	    }
+	}
+}
+#endif
+
 static SWFMovie
 compile(const char* filename, const char* ppfile, int version)
 {
diff -uN ming-0.4.4/util/parser.c.ORG ming-0.4.4/util/parser.c
--- ming-0.4.4/util/parser.c.ORG	Fri Oct 28 14:22:39 2011
+++ ming-0.4.4/util/parser.c	Fri Oct 28 15:44:33 2011
@@ -242,6 +242,7 @@
   int i;
 
   gerec->GlyphIndex = malloc((glyphbits+31)/32 * sizeof(UI32) );
+  gerec->GlyphIndex[0] = 0; /* for glyphbits == 0 */
   for( i=0; glyphbits; i++ ) {
 	  if( glyphbits > 32 ) {
 	  	gerec->GlyphIndex[i] = readBits(f, 32);
@@ -253,6 +254,7 @@
   }
 
   gerec->GlyphAdvance = malloc((advancebits+31)/32 * sizeof(UI32) );
+  gerec->GlyphAdvance[0] = 0; /* for advancebits == 0 */
   for( i=0; advancebits; i++ ) {
 	  if( advancebits > 32 ) {
 	  	gerec->GlyphAdvance[i] = readBits(f, 32);
diff -uN ming-0.4.4/util/read.c.ORG ming-0.4.4/util/read.c
--- ming-0.4.4/util/read.c.ORG	Wed Oct 26 15:33:18 2011
+++ ming-0.4.4/util/read.c	Fri Oct 28 12:42:22 2011
@@ -122,6 +122,7 @@
 {
   char data[8];
 
+#ifdef SWF_LITTLE_ENDIAN
   data[4] = readUInt8(f);
   data[5] = readUInt8(f);
   data[6] = readUInt8(f);
@@ -130,6 +131,17 @@
   data[1] = readUInt8(f);
   data[2] = readUInt8(f);
   data[3] = readUInt8(f);
+#else
+  data[3] = readUInt8(f);
+  data[2] = readUInt8(f);
+  data[1] = readUInt8(f);
+  data[0] = readUInt8(f);
+  data[7] = readUInt8(f);
+  data[6] = readUInt8(f);
+  data[5] = readUInt8(f);
+  data[4] = readUInt8(f);
+#endif
+
 
   return *((double *)data);
 }
@@ -138,10 +150,17 @@
 {
   char data[4];
 
+#ifdef SWF_LITTLE_ENDIAN
   data[0] = readUInt8(f);
   data[1] = readUInt8(f);
   data[2] = readUInt8(f);
   data[3] = readUInt8(f);
+#else
+  data[3] = readUInt8(f);
+  data[2] = readUInt8(f);
+  data[1] = readUInt8(f);
+  data[0] = readUInt8(f);
+#endif
 
   return *((float *)data);
 }
@@ -374,8 +393,6 @@
 {
 	_dumpBytes(f, length, 1 );
 }
-
-  int j=0, i, k, l=0;
 
 void dumpBuffer(unsigned char *buf, int length)
 {
----- To here -----

+========================================================+
 Shigeharu TAKENO     NIigata Institute of Technology
                       kashiwazaki,Niigata 945-1195 JAPAN
 sh...@ie...   TEL(&FAX): +81-257-22-8161
+========================================================+

Re: [Ming-users] Ming-0.4.4 released

[Sandro S. <st...@ke...>]

On Thu, Oct 27, 2011 at 01:57:34PM +0900, Shigeharu TAKENO wrote:

> 1) When "make check", I saw the following message:

Pushed, thanks (missing font-kerntest.fdb)

> 2) I needed to change paths of the 1st line of test scripts:
> 
>  #!/usr/bin/python, #!/usr/bin/perl, #!/usr/bin/php
> 
> to the proper paths

We want to find a way to recuce maintainance cost, maybe

#!/usr/bin/env python

and similar would work for you ?

Alternatively we should work something out from the dotests.pl
runner, so that the path to interpreter is only substituted
in that single place.

Could you please file a ticket for this on bugzilla ?

> I needed to add the following line:
> 
>   set_time_limit(0);
> 
> to test/Video/test0{1,4,5}.php, since my machine does not work
> so fast. 

Again, best would be putting the limit in a common place.
See if you can do it in dotests.pl (seek for "php -n -d").
And again, a ticket on bugzilla (or github) is appreciated.

> 3) I used the patch below at the compilation. Especially, I think
> it is serious for util/parser.c.

Is there any test failing w/out your patch and working with ?
If not we'll want one to go with the fix.
And we do want a ticket to track this too.

Thanks a lot for your report, I'll be glad to ship 0.4.5 shortly
if you can help fixing these issues.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] Ming-0.4.4 released

[Shigeharu T. <sh...@ie...>]

shige 10/27 2011
----------------

I tried to compile of ming-0.4.4, and I found some problems. I 
did it on Solaris 9 (Sparc).

1) When "make check", I saw the following message:

  Executing './test07 .' in /hoge/hoge/ming-0.4.4/test/Font
  open font file failed
  ./../Media//font-kerntest.fdb: No such file or directory
  Invokation of test builder './test07' returned 

I found the file "font-kerntest.fdb" at

 http://bugs.libming.org/attachment.cgi?id=29

and it should be included in the distribution.

2) I needed to change paths of the 1st line of test scripts:

 #!/usr/bin/python, #!/usr/bin/perl, #!/usr/bin/php

to the proper paths and I needed to add the following line:

  set_time_limit(0);

to test/Video/test0{1,4,5}.php, since my machine does not work
so fast. 

3) I used the patch below at the compilation. Especially, I think
it is serious for util/parser.c.

----- From here -----
diff -uN ming-0.4.4/configure.in.ORG ming-0.4.4/configure.in
--- ming-0.4.4/configure.in.ORG	Wed Oct 26 16:29:42 2011
+++ ming-0.4.4/configure.in	Thu Oct 27 13:02:59 2011
@@ -150,6 +150,8 @@
 		AC_MSG_ERROR([Can't build tcl extension, as tcl executable could not be found])
 	fi
 
+tclbindir=`dirname $TCL`
+
 tcllibdirs_default="\
 $prefix/lib/itcl \
 $prefix/lib \
@@ -157,7 +159,8 @@
 $HOME/lib \
 /usr/local/lib \
 /usr/lib64 \
-/usr/lib"
+/usr/lib \
+`dirname $tclbindir`/lib"
 
 	for i in $tcllibdirs_default; do
 	  for suf in 8.3 8.4 ""; do
@@ -193,7 +196,8 @@
 $HOME/include \
 /usr/local/include \
 /usr/include/tcl \
-/usr/include"
+/usr/include \
+`dirname $tclbindir`/include"
 
 	for i in $tclincdirs_default; do
 	   if test -f $i/tcl.h; then
diff -uN ming-0.4.4/test/actionscript/ActionScriptTest.c.ORG ming-0.4.4/test/actionscript/ActionScriptTest.c
--- ming-0.4.4/test/actionscript/ActionScriptTest.c.ORG	Wed Oct 26 15:33:18 2011
+++ ming-0.4.4/test/actionscript/ActionScriptTest.c	Thu Oct 27 10:04:05 2011
@@ -41,6 +41,46 @@
 #include <limits.h>
 #include <makeswf.h>
 
+#ifndef HAVE_VASPRINTF
+/* Workaround for the lack of vasprintf()
+ * As found on: http://unixpapa.com/incnote/stdio.html
+ * Seems to be Public Domain
+ */
+int
+vasprintf(char **ret, const char *format, va_list ap)
+{
+	va_list ap2;
+	int len = 100;        /* First guess at the size */
+
+	if ((*ret = (char *) malloc(len)) == NULL)
+	{
+	    return -1;
+	}
+	while (1)
+	{
+	    int nchar;
+	    va_copy(ap2, ap);
+	    nchar= vsnprintf(*ret, len, format, ap2);
+	    if (nchar > -1 && nchar < len)
+	    {
+		return nchar;
+	    }
+	    if (nchar > len)
+	    {
+		len= nchar+1;
+	    } else
+	    {
+		len*= 2;
+	    }
+	    if ((*ret = (char *) realloc(*ret, len)) == NULL)
+	    {
+		free(*ret);
+		return -1;
+	    }
+	}
+}
+#endif
+
 static SWFMovie
 compile(const char* filename, const char* ppfile, int version)
 {
diff -uN ming-0.4.4/util/parser.c.ORG ming-0.4.4/util/parser.c
--- ming-0.4.4/util/parser.c.ORG	Wed Oct 26 15:33:18 2011
+++ ming-0.4.4/util/parser.c	Thu Oct 27 10:15:43 2011
@@ -242,7 +242,7 @@
   int i;
 
   gerec->GlyphIndex = malloc((glyphbits+31)/32 * sizeof(UI32) );
-  for( i=0; glyphbits; i++ ) {
+  for( i=0; i <= glyphbits; i++ ) {
 	  if( glyphbits > 32 ) {
 	  	gerec->GlyphIndex[i] = readBits(f, 32);
 	  	glyphbits -= 32;
@@ -253,7 +253,7 @@
   }
 
   gerec->GlyphAdvance = malloc((advancebits+31)/32 * sizeof(UI32) );
-  for( i=0; advancebits; i++ ) {
+  for( i=0; i <= advancebits; i++ ) {
 	  if( advancebits > 32 ) {
 	  	gerec->GlyphAdvance[i] = readBits(f, 32);
 	  	advancebits -= 32;
diff -uN ming-0.4.4/util/read.c.ORG ming-0.4.4/util/read.c
--- ming-0.4.4/util/read.c.ORG	Wed Oct 26 15:33:18 2011
+++ ming-0.4.4/util/read.c	Thu Oct 27 10:17:20 2011
@@ -122,6 +122,7 @@
 {
   char data[8];
 
+#ifdef SWF_LITTLE_ENDIAN
   data[4] = readUInt8(f);
   data[5] = readUInt8(f);
   data[6] = readUInt8(f);
@@ -130,6 +131,17 @@
   data[1] = readUInt8(f);
   data[2] = readUInt8(f);
   data[3] = readUInt8(f);
+#else
+  data[3] = readUInt8(f);
+  data[2] = readUInt8(f);
+  data[1] = readUInt8(f);
+  data[0] = readUInt8(f);
+  data[7] = readUInt8(f);
+  data[6] = readUInt8(f);
+  data[5] = readUInt8(f);
+  data[4] = readUInt8(f);
+#endif
+
 
   return *((double *)data);
 }
@@ -138,10 +150,17 @@
 {
   char data[4];
 
+#ifdef SWF_LITTLE_ENDIAN
   data[0] = readUInt8(f);
   data[1] = readUInt8(f);
   data[2] = readUInt8(f);
   data[3] = readUInt8(f);
+#else
+  data[3] = readUInt8(f);
+  data[2] = readUInt8(f);
+  data[1] = readUInt8(f);
+  data[0] = readUInt8(f);
+#endif
 
   return *((float *)data);
 }
@@ -374,8 +393,6 @@
 {
 	_dumpBytes(f, length, 1 );
 }
-
-  int j=0, i, k, l=0;
 
 void dumpBuffer(unsigned char *buf, int length)
 {
----- To here -----

+=========================================================+
| Shigeharu TAKENO      NIigata Institute of Technology   |
|                        Kashiwazaki,Niigata 945-11 JAPAN |
|(sh...@ie...)  TEL(&FAX): +81-257-22-8161       |
+=========================================================+

[Ming-users] Ming-0.4.4 released

[Sandro S. <st...@ke...>]

Ming-0.4.4 was just released, overdue after almost 2 years since previous.

[ http://strk.keybit.net/blog/2011/10/26/ming-0-4-4-released/ ]

Changes since 0.4.3:

  * Generally improve swftoscript and decompiler
  * Change makefdb to name output files by font ID, to play nicer
    with swftoscript.
  * Add support for 'class A extends B' syntax in actioncompiler
  * Fix bug in 'makeswf' failing to catch some compile errors
    (bugzilla #94) and being too silent in swf embedding errors
  * Fix bug in action compiler dealing with class methods (bugzilla #94)
  * Add support for libpng > 1.4 (bugzilla #96)
  * Add font kernings support (bugzilla #95)
  * Add button characters export capabilities
  * Add support for 'swfAction <code>' syntax in asm blocks

I'll remind that the new Ming source code is on github:

  http://github.com/libming/libming
                                                                                Releases are still accessible from sourceforge:

  http://sourceforge.net/projects/ming/files/Releases/

Updated informations are still on http://www.libming.org


--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

--

what comes below this line is just spam, dont bother scrolling...




















































    still here ?

Re: [Ming-users] Hi, what can i do if my web-space provider doen't install Ming???

[Sandro S. <st...@ke...>]

On Fri, May 27, 2011 at 05:16:09PM +0800, Yihe Chen wrote:
> Hi, all:
>    I want to do a simple  flv  palyer  using  Ming, but the web-space provider
> doesn't install ming extension in their  host  config.  What can i do????
> 
>    If  I  make  the  flv player locally,  namely in my own linux box.
> How  do I use the player in my rented space, I have no  permission to
> install  Ming  in this space.

You just upload the SWF and it'll work.

--strk; 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html

-- 

What comes next is just spam, don't bother scrolling 
































































    ... still here ?


































 well.. you've been warned

[Ming-users] Hi, what can i do if my web-space provider doen't install Ming???

[Yihe C. <yi...@gm...>]

Hi, all:
   I want to do a simple  flv  palyer  using  Ming, but the web-space provider
doesn't install ming extension in their  host  config.  What can i do????

   If  I  make  the  flv player locally,  namely in my own linux box.
How  do I use the player in my rented space, I have no  permission to
install  Ming  in this space.

[Ming-users] (no subject)

[Hujan A. <huj...@gm...>]

http://www.elifsongur.com/wp-content/themes/twentyten/myinfo.html

Re: [Ming-users] [Ming-dev] ming as3 support

[Hujan A. <huj...@gm...>]

some one have tutorial how to instal ming for linux ubuntu ?

2011/3/31 Sandro Santilli <st...@ke...>

> On Wed, Mar 30, 2011 at 05:21:04PM -0400, Michael Roberts wrote:
>
> > What references are out there for what AS3 support would require?
>
> SWFTools include an as3 compiler, check it out.
> It is already flex/bison based, like Ming's, so it'd be
> the natural thing to look at.
>
> --strk;
>
> What comes next is just spam, don't bother scrolling
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> Create and publish websites with WebMatrix
> Use the most popular FREE web apps or write code yourself;
> WebMatrix provides all the features you need to develop and
> publish your website. http://p.sf.net/sfu/ms-webmatrix-sf
> _______________________________________________
> Ming-users mailing list
> Min...@li...
> https://lists.sourceforge.net/lists/listinfo/ming-users
>