Control: tags -1 + moreinfo

Hi Florian,

On Thu, Nov 13, 2014 at 10:22:20PM +0100, Florian Schlichting wrote:
> in the meantime, another bugfix release, mpd 0.19.3 was released;
> changelog below, debdiff attached. This upload would additionally fix
> #768094 (important: "crashes with std::logic_error on database update")
> and #769436 (normal: "Fails to play some audio streams")

As much as I would like to see the latest and greatest mpd in jessie, I
think that these changes are out of scope for the freeze policy (not
speaking for the release team here, but assuming that they agree).

> diff -Nru mpd-0.19.1/configure.ac mpd-0.19.3/configure.ac
> --- mpd-0.19.1/configure.ac   2014-10-11 20:25:57.000000000 +0200
> +++ mpd-0.19.3/configure.ac   2014-11-05 18:24:15.000000000 +0100
> @@ -293,7 +293,9 @@
>  AC_ARG_ENABLE(libmpdclient,
>       AS_HELP_STRING([--enable-libmpdclient],
>               [enable support for the MPD client]),,
> -     enable_libmpdclient=$database_auto)
> +     enable_libmpdclient=auto)
> +MPD_DEPENDS([enable_libmpdclient], [enable_database],
> +     [Cannot use --enable-libmpdclient with --disable-database])
>  
>  AC_ARG_ENABLE(expat,
>       AS_HELP_STRING([--enable-expat],

All these new MPD_DEPENDS are clearly improvements, but I don't think
this qualifies as important for jessie.

> diff -Nru mpd-0.19.1/doc/doxygen.conf mpd-0.19.3/doc/doxygen.conf
> --- mpd-0.19.1/doc/doxygen.conf       2014-10-11 20:26:19.000000000 +0200
> +++ mpd-0.19.3/doc/doxygen.conf       2014-11-05 18:24:29.000000000 +0100
> @@ -534,7 +534,7 @@
>  # directories like "/usr/src/myproject". Separate the files or directories
>  # with spaces.
>  
> -INPUT = /home/max/git/mpd/src/
> +INPUT = /home/max/git/stable-mpd/src/
>  
>  # This tag can be used to specify the character encoding of the source files
>  # that doxygen parses. Internally doxygen uses the UTF-8 encoding, which is

Even though 0.19.3 claims to be a bug-fix release it contains quite a
few non-functional changes like this one.

Another recurring example of this kind is adding "#ifdef HAVE_GLIB" even
though Debian's mpd always enables glib.

> diff -Nru mpd-0.19.1/NEWS mpd-0.19.3/NEWS
> --- mpd-0.19.1/NEWS   2014-10-19 01:03:36.000000000 +0200
> +++ mpd-0.19.3/NEWS   2014-11-11 11:21:38.000000000 +0100
> @@ -1,3 +1,38 @@
> +ver 0.19.3 (2014/11/11)
> +* protocol
> +  - fix "(null)" result string to "list" when AlbumArtist is disabled
> +* database
> +  - upnp: fix breakage due to malformed URIs
> +* input
> +  - curl: another fix for redirected streams
> +* decoder
> +  - audiofile: fix crash while playing streams
> +  - audiofile: fix bit rate calculation
> +  - ffmpeg: support opus

New feature.

> +  - opus: fix bogus duration on streams
> +  - opus: support chained streams

New feature.

> +  - opus: improved error logging
> +* fix distorted audio with soxr resampler
> +* fix build failure on Mac OS X with non-Apple compilers

Unrelated to Debian.

> diff -Nru mpd-0.19.1/src/output/plugins/RoarOutputPlugin.cxx 
> mpd-0.19.3/src/output/plugins/RoarOutputPlugin.cxx
> --- mpd-0.19.1/src/output/plugins/RoarOutputPlugin.cxx        2014-09-28 
> 13:24:40.000000000 +0200
> +++ mpd-0.19.3/src/output/plugins/RoarOutputPlugin.cxx        2014-10-27 
> 09:26:34.000000000 +0100
> @@ -46,7 +46,7 @@
>       struct roar_connection con;
>       struct roar_audio_info info;
>       mutable Mutex mutex;
> -     volatile bool alive;
> +     bool alive;
>  
>  public:
>       RoarOutput()

This change seems undocumented in changelog and NEWS. It is not clear,
why this should suddenly become safe when it wasn't earlier.

> diff -Nru mpd-0.19.1/src/util/UriUtil.cxx mpd-0.19.3/src/util/UriUtil.cxx
> --- mpd-0.19.1/src/util/UriUtil.cxx   2014-10-10 22:43:40.000000000 +0200
> +++ mpd-0.19.3/src/util/UriUtil.cxx   2014-11-01 12:51:30.000000000 +0100
> @@ -54,6 +54,23 @@
>       return suffix;
>  }
>  
> +const char *
> +uri_get_suffix(const char *uri, UriSuffixBuffer &buffer)
> +{
> +     const char *suffix = uri_get_suffix(uri);
> +     if (suffix == nullptr)
> +             return nullptr;
> +
> +     const char *q = strchr(suffix, '?');
> +     if (q != nullptr && size_t(q - suffix) < sizeof(buffer.data)) {
> +             memcpy(buffer.data, suffix, q - suffix);
> +             buffer.data[q - suffix] = 0;
> +             suffix = buffer.data;
> +     }
> +
> +     return suffix;
> +}
> +
>  static const char *
>  verify_uri_segment(const char *p)
>  {

Refactoring is going on here. Not bad in itself, but maybe inappropriate
for jessie.

On a whole, the size of the diff makes it hard to correlate individual
hunks to chaneglog or NEWs. This ultimately makes the diff harder to
review and goes against the freeze policy.

Violated freeze policy items:
 * Do not package new upstream releases
 * Document all changes verbosely
 * Non-documentation changes of severity < important

Based on the sheer number of violations I am tagging the bug moreinfo
already and ask you to provide less intrusive changes.

Please don't shoot the messenger. I'm trying to avoid releasing without
mpd. (It's already listed for autoremoval.)

Helmut


-- 
To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141119075502.ga10...@alf.mars

Reply via email to