The LD_PRELOAD workaround is an effective workaround for now, so this
isn't actually causing any serious pain at the moment.

It is trivial enough to add the dependency.  I'm just not sure if
that's the correct thing to do.  If you are *positive* I'll believe
you.  I will note that the seamonkey browser is not currently in
Debian, and that this issue does not, to my knowledge, manifest using
any actual Debian browser.  Moreover, since the plugin protocol
sort-of requires libXt, maybe browsers that don't use libXt themselves
should yank in the library anyway in order to service plugins, kind of
like an internal mimic of the effect of an appropriate LD_PRELOAD?  In
fact, maybe that's something they're supposed to be doing already,
making this a seamonkey bug?  I'd really like to get the opinion of
someone actually familiar with this stuff, instead of just guessing.

>> What if the browser itself already uses libXt, and links to a
>> different version of the library than the plugin?

> Then you're screwed whether or not nsdejavu.so properly declares its own
> dependencies, because the browser's version of libXt is not guaranteed to
> provide the ABI that the djvu plugin is expecting.

You know, these are just tiny little plugins!  It seems quite likely
that ABI changes that require an so-lib name rev would be in portions
of the ABI that the plugin does not exercise.  That is a good example
of a situation where an explicit dependency in the plugin itself might
cause trouble in the future.

Upstream has apparently had some actual problematic library conflicts
caused by including an explicit dependency.  Here is what they said
when I asked "why no explicit dependency in build scripts?"

  > The reason was that some browsers were linked with custom or
  > static versions of libXt, and we wanted to use whatever Xt they
  > were using internally.  Note that the plugin protocol sort-of
  > require libXt to capture the events.  Which browser is causing
  > problems?

So, who's a big web plugin expert?
--
Barak A. Pearlmutter
 Hamilton Institute & Dept Comp Sci, NUI Maynooth, Co. Kildare, Ireland
 http://www.bcl.hamilton.ie/~barak/


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to