Re: libmng data-push mechanism(s)

At 09:38 PM 4/14/2004 +0200, Alexander Opitz wrote:
>Hi Glenn,
>
>> No, the query is in case the user has installed an old libmng instead
>> of the supplied libmng-1.0.8.  You query for the new (1.0.8) functions
>> that might not be present.  Or, you query for the version number and
>> take appropriate action if the version number is too old.  Or, you
>> can just go ahead and use the new functions and let the loader bail
>> out if it cannot find them in the libmng version being loaded.  I'd
>> probably take that third approach and not bother with queries.
>
>the thing in question was, that you tried my old patch if it uses the new 
>functions ... and this can't never work, course the old patch doesn't know of 
>the new existing functions.

No, I tried the old patch with whatever old functions it used.  I made
no new changes to the glue (changing the glue is your task).  I only
verified that if the only change is replacing libmng with yesterday's
CVS it still works.  It is really no surprise that it does work the same
as previously.

>I can try to make the new patch working how you mentioned, but I think this 
>makes the source bigger.

Yes, and the *.so is also about 2k bigger because it contains the new
(but unused) functions.

>And, mngzilla distribute mng inside them self, so 
>the glue code uses the same libmng code for which it is written. You can also 
>softlink to an installed libmng ... but then the glucode should mention it 
>need's >=1.0.8 (cause of the crashers in the lower versions).

Yes, all of that is correct.

Glenn



--
Send the message body "help" to [email protected]