Re: Tracing what is loading

[Harry Putnam]

Kevin Rodgers <kevin.d.rodgers@gmail.com> writes:


[...]

>> All are in /usr/share/emacs/22.3*
>
> I think those were dumped into the emacs executable.
>
>> Then comes a line (106):
>>
>>     /usr/local/GNUS/lisp/gnus-load.el
>>
>> An address that is not a normal path to be searched.
>> How did emacs know to look there?
>>
>> That is the address to gnus but not the one in the emacs
>> distribution.  This is the cvs version.
>
> Hmmm, perhaps that directory was in load-path at some point,
> or the file was loaded using its absolute file name.
>
>> The information that tells emacs where to look for that lisp directory
>> resides in /usr/local/share/emacs/site-lisp/site-start.el
>>
>> Which according to the list has not yet been loaded.
>>
>> That doesn't show up until line 118
>> /usr/local/share/emacs/site-lisp/site-start.el
>>
>> So is this listing just not accurate to that degree or is something else
>> going on?  I ask because I'm trying to discover when certain init
>> files are loaded.
>
> Does /usr/local/GNUS/lisp/gnus-load.el show up in the list when you
> start emacs with -Q?  What about the other files between it and
> /usr/local/share/emacs/site-lisp/site-start.el?

No and site-start.el doesn't either but then they shouldn't with -Q

Even with -Q though I do see one anomaly.  
The whole list (108) lines), with the exception of two lines near the
end are all in /usr/share/emacs/22.3. and you've suggested how that happens.

  /usr/share/emacs/site-lisp/subdirs.el  (loads at 102)

Again that seams normal enough.

And this one is a little puzzling:

  /home/reader/.abbrev_defs (loads at line 106)

But of course ~/.emacs is not loaded with the -Q flag so what
instructs emacs to load ~/.abbrev_defs?  I thought ~/.emacs was doing
that. I do have code there about .abbrev_defs, But as you can see,
even with -Q something has caused emacs to load ~/.abbrev_defs.

I'm beginning to think my initial thought to track things in the
*Messages* buffer is a better place to track init files.

The history-list shows the cart before the horse in several instances
and the list is cluttered with too much stuff that isn't related to
init files.

Here is a similar example.. loading emacs without -Q:

/anex/usr/local/share/emacs/site-lisp/whats_loading.el (line 110)
/usr/local/GNUS/lisp/gnus-load.el (line 111)

Both of those things are loaded from site-start.el (I'm pretty sure
that's the only place.  And the `shadow' list bears that out too.

  /usr/local/share/emacs/site-lisp/site-start.el (line 119)

The history list has site-start.el loading 8,9 lines after 2 items
that are only loaded from site-start.el.

But looking at the *Messages* output.. things appear in the right
order.  However, I don't see the init files gentoo developers have
stuck in /usr/share/emacs/site-lisp.  

I think they are loading somewhere before my site-start.el file but
neither the history-list or *Messages* shows it.

I can probably get to the bottom of it by putting a message in those
files or something... still the whole init process is pretty confusing.