Re: serial in branch-1-5

[Ralf Wildenhues]

* Alexandre Duret-Lutz wrote on Tue, Feb 01, 2005 at 11:15:21AM CET:
> >>> "Ralf" == Ralf Wildenhues <address@hidden> writes:
> 
>  Ralf> sorry for self-followup.
>  Ralf> * Ralf Wildenhues wrote on Tue, Feb 01, 2005 at 08:42:44AM CET:
>  >> I just noticed that, while branch-2-0 and HEAD have m4/ltversion.m4,
>  >> branch-1-5 needs manual update of `serial' in libtool.m4 and ltdl.m4.
>  >> Is that correct?  If so, then it needs to be bumped up before the
>  >> release.
>  >> 
>  >> Would it have to be updated for every checkin?
> 
>  Ralf> And what's more: aclocal from CVS HEAD does not like lines like
>  Ralf>  # serial 7 LTDL_INIT
> 
> Wow.  I planned to prepare and send patches tonight, but you
> reacted faster.

Pure luck.  I do read auto{conf,make}-patches, though, and after the
almost-debacle concerning ${datarootdir} (libtoolize from 1.5.10 is
unusable with CVS autoconf; that is one reason we must do another 1.5
release), I figured it'd be good to always check Libtool with the
current tools.

>  Ralf> Automake docs don't specify the format, and I don't know it.
> 
> It's documented, but apparently not in the right place.  Where
> did you look?

Just searched for `serial' in `info Autoconf'.

> Anyway the explanation of the format in the documentation is the
> same that is output when aclocal warns it does not understand a
> serial number.  It looks pretty explicit to me:
> 
> /home/adl/usr/share/aclocal/libtool.m4:30: malformed serial number `49 
> LT_INIT', expecting only digits and dots

Not specific enough for me.  Do you allow multiple dots?  Can the serial
start or end with a dot?  What are the semantics, (in)equality, supposed
way of updating?  (A pointer to the docs would be sufficient.)

There are, by the way, other packages which put certain macro names on
this line.

Regards,
Ralf