Re: [Gnash-commit] [SCM] Gnash branch, master, updated. release_0_8_9_start-62-g3240ec3

[strk]

On Mon, Feb 14, 2011 at 04:56:29PM +0000, Rob Savoye wrote:

> commit 3240ec352539a3c119d3c73fbb948f92b058988f
> Author: Rob Savoye <address@hidden>
> Date:   Mon Feb 14 09:55:43 2011 -0700
> 
>     handle multiple messages in a packet.
[...]
> diff --git a/plugin/npapi/plugin.cpp b/plugin/npapi/plugin.cpp
> index 5c4b15d..f615163 100644
> --- a/plugin/npapi/plugin.cpp
> +++ b/plugin/npapi/plugin.cpp
> @@ -715,7 +717,6 @@ nsPluginInstance::handlePlayerRequests(GIOChannel* 
> iochan, GIOCondition cond)
>          // looping here forever, so this is our escape from that loop.
>          if (retries-- <= 0) {
>              gnash::log_error("Too many attempts to read from the player!");
> -            return false;
>          }
>          error = 0;
>          request = 0;

Was the above intentional ?
The comment right above says that the return is required to avoid
entering an infinite loop...

Full code snippet:

        // When in non-blocking mode, we'll get several iterations of this
        // loop while waiting for data. if data never arrives, we'd be stuck
        // looping here forever, so this is our escape from that loop.
        if (retries-- <= 0) {
            gnash::log_error("Too many attempts to read from the player!");
            return false;
        }

The change above drops the return ...


> diff --git a/plugin/npapi/plugin.h b/plugin/npapi/plugin.h
> index 007df68..9801f93 100644
> --- a/plugin/npapi/plugin.h
> +++ b/plugin/npapi/plugin.h
> @@ -144,7 +144,7 @@ private:
>  //  0: no messages at all
>  //  1: fatal errors (errors preventing the plugin from working as it should)
>  //  2: informational messages
> -#define GNASH_PLUGIN_DEBUG 1
> +#define GNASH_PLUGIN_DEBUG 2
>  
>  // This following logging code is copied from libbase/log.h, but
>  // duplicated here because the plugin only needs a more trimmed down

This ? I don't think you want to release with that debug level ?


> commit 0aecfe8e43995f5de3b0d5890b6a52a9895e2d51
> Author: Rob Savoye <address@hidden>
> Date:   Mon Feb 14 09:46:19 2011 -0700
> 
>     properly handle 64 vs 32 bit for the plugin directory
> 
> diff --git a/macros/kde4.m4 b/macros/kde4.m4
> index 40383cc..2825877 100644
> --- a/macros/kde4.m4
> +++ b/macros/kde4.m4
> @@ -192,9 +192,17 @@ AC_DEFUN([GNASH_PATH_KDE4],
>      fi
>  
>      if test x"${with_kde4_plugindir}" != x ; then 
> -      KDE4_PLUGINDIR="${with_kde4_plugindir}/lib/kde4"
> +      if test -d ${with_kde4_plugindir}/lib64; then
> +        KDE4_PLUGINDIR="${with_kde4_plugindir}/lib64/kde4"
> +      else
> +        KDE4_PLUGINDIR="${with_kde4_plugindir}/lib/kde4"
> +      fi
>      else
> -      KDE4_PLUGINDIR="${KDE4_PREFIX}/lib/kde4"
> +      if test -d ${KDE4_PREFIX}/lib64; then
> +        KDE4_PLUGINDIR="${KDE4_PREFIX}/lib64/kde4"
> +      else
> +        KDE4_PLUGINDIR="${KDE4_PREFIX}/lib/kde4"
> +      fi
>      fi
>      if test x"${with_kde4_servicesdir}" != x ; then 
>        KDE4_SERVICESDIR="${with_kde4_servicesdir}/share/kde4/services"

The version based on ${KDE4_PREFIX} seems good to me (as we're guessing)
but the part based on ${with_kde4_plugindir} is dealing with user
expliclty giving --with-kde4-plugindir=<something>, so wouldn't it be
better to just set KDE4_PLUGINDIR="${with_kde4_plugindir}" trusting
the user for what he's doing ?

--strk; 

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