On Tuesday 17 March 2009, Stephen J. Turnbull wrote:
> Michael Sperber writes:
>  > "Stephen J. Turnbull" <stephen at xemacs.org> writes:
>  > >  > > Examples of undocumented differences?
>  > >  >
>  > >  > --infodir's default handling;
>  > >
>  > > You'll have to get Mike to discuss that.  He insists that things be
>  > > done as they are, but I've never personally had any reason to care one
>  > > way or the other (no personal config problems, no need for the
>  > > features he added, and I don't recall ever getting user questions
>  > > either).
>  >
>  > That isn't accurate.  --infodir is a holdover from the old days.
>
> Of course the option itself is a holdover.  But my understanding is
> that by "undocumented difference from autoconf" he's referring to
> INFODIR_USER_DEFINED, and the catering to run-in-place and people who
> schlep whole installed trees around and other cases that you won't
> encounter in an RPM install.
>
> Ville?

Infodir handling has currently so many problems that I'm having a hard time 
explaining it, but I'll give it one more try off the cuff.  Just keep the 
combinations of --prefix, --datadir, --datarootdir and --infodir users may 
set in mind in the following and ignore for the moment $foo vs $foo_expanded.  
Some of the following are caused by same underlying issue(s), and may be 
issues with other *dir settings/options similarly.

1) configure --help says infodir's default is $datarootdir/info.  It is not, 
it's $datarootdir/$instvardir/info (kind of, depending on a few other --*dir 
settings).

2) configure is internally inconsistent, infodir defaults to $datarootdir/info 
in the beginning of it but the code from configure.ac uses $prefix/share/info 
to test if it is set to the default.

3) The issue mentioned in Subject of this thread (--infodir given by the user 
is sometimes ignored).

4) Defining --datadir to something that does not match $prefix/share results 
in INFODIR_USER_DEFINED being set.  This is inconsistent with *both* of the 
values in 2) above: configure --help (it says $datarootdir/info, not 
$datadir/info) and the logic when checking whether infodir is set to the 
default (it checks $prefix/share/info, not $datadir/info).

5) The *_USER_DEFINED defines do not actually get set based on whether the 
user defined the corresponding --*dir option.  This was already discussed.

_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta at xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Michael Sperber writes:
 > "Stephen J. Turnbull" <stephen at xemacs.org> writes:

 > >  > > Examples of undocumented differences?
 > >  > 
 > >  > --infodir's default handling;
 > >
 > > You'll have to get Mike to discuss that.  He insists that things be
 > > done as they are, but I've never personally had any reason to care one
 > > way or the other (no personal config problems, no need for the
 > > features he added, and I don't recall ever getting user questions
 > > either).
 > 
 > That isn't accurate.  --infodir is a holdover from the old days.

Of course the option itself is a holdover.  But my understanding is
that by "undocumented difference from autoconf" he's referring to
INFODIR_USER_DEFINED, and the catering to run-in-place and people who
schlep whole installed trees around and other cases that you won't
encounter in an RPM install.

Ville?

_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta at xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Supersedes issue504, issue505, issue506, issue507.  See those
issues for further details.

"Stephen J. Turnbull" <stephen at xemacs.org> writes:

>  > > Examples of undocumented differences?
>  > 
>  > --infodir's default handling;
>
> You'll have to get Mike to discuss that.  He insists that things be
> done as they are, but I've never personally had any reason to care one
> way or the other (no personal config problems, no need for the
> features he added, and I don't recall ever getting user questions
> either).

That isn't accurate.  --infodir is a holdover from the old days.  When I
re-wrote the path-searching stuff, I didn't tackle any interface issues,
even though oddities such as --infodir certainly bothered me, and I'd
change it in a heartbeat if it was so easy.  The path-searching code has
been remarkably resistant to interface changes simply because changing
it would break people's setups.