How to show *Warnings* buffer at startup time? [Re: Pretest next week]

How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Kenichi Handa]

In article <8763g39jwd.fsf@cyd.mit.edu>, Chong Yidong <cyd@stupidchicken.com> writes:

> Hi all,
> I'm planning to make the 23.0.94 pretest next Wednesday, the 20th.

> The release seems to be coming along nicely; the font code seems to have
> converged at last (with the exception of #2667), and most of the
> documentation work is now done.  I'm thinking of cutting the branch
> after the next pretest, so that some of the pent-up patches can start
> being applied.  I have not, however, discussed this in detail with
> Stefan yet.  If you like, please feel free to weigh in with opinions
> and/or concerns about the release process.

I've just noticed that the subject issure is not yet
settled.  I think this is not a serious one but the attached
patch to startup.el is simple and it helps people identify a
fontset problem earlier.

---
Kenichi Handa
handa@m17n.org


In article <E1Ljviv-0002mh-Jc@etlken>, Kenichi Handa <handa@m17n.org> writes:

> I'm going to change create-fontset-from-x-resource
> (fontset.el) as this:

> @@ -1049,8 +1058,11 @@
>      (while (setq fontset-spec (x-get-resource (format "fontset-%d" idx)
>  					      (format "Fontset-%d" idx)))
>        (condition-case nil
> -	  (create-fontset-from-fontset-spec fontset-spec t 'noerror)
> -	(error (message "Fontset-%d: invalid specification in X resource" idx)))
> +	  (create-fontset-from-fontset-spec fontset-spec t)
> +	(error (display-warning
> +		'initialization
> +		(format "Fontset-%d: invalid specification in X resource" idx)
> +		:warning)))
>        (setq idx (1+ idx)))))
 
>  ;;

> But the *Warnings* buffer doesn't show up at startup time
> even if create-fontset-from-fontset-spec causes an error.
> Could someone tell me what is wrong with the above change?

> If it is because the X frame is not yet created at that
> time, don't we need something like the attached patch?

> ---
> Kenichi Handa
> handa@m17n.org

> --- startup.el.~1.527.~	2009-02-23 09:51:02.000000000 +0900
> +++ startup.el	2009-03-18 22:16:35.000000000 +0900
> @@ -1235,7 +1235,13 @@
>    (if (and (boundp 'x-session-previous-id)
>             (stringp x-session-previous-id))
>        (with-no-warnings
> -	(emacs-session-restore x-session-previous-id))))
> +	(emacs-session-restore x-session-previous-id)))
> +
> +  (let ((warning-buf (get-buffer "*Warnings*")))
> +    (if (and warning-buf
> +	     (not (get-buffer-window warning-buf))
> +	     (> (buffer-size warning-buf) 0))
> +	(display-buffer warning-buf))))
 
>  (defcustom initial-scratch-message (purecopy "\
>  ;; This buffer is for notes you don't want to save, and for Lisp evaluation.

Re: How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Chong Yidong]

> I've just noticed that the subject issure is not yet settled.  I think
> this is not a serious one but the attached patch to startup.el is
> simple and it helps people identify a fontset problem earlier.

Sorry, I forgot about this issue when you first raised it.  But I don't
think it's serious enough to carve out an exception to the
regressions-only rule.

Re: How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Kenichi Handa]

In article <87ab5ev286.fsf@cyd.mit.edu>, Chong Yidong <cyd@stupidchicken.com> writes:

> > I've just noticed that the subject issure is not yet settled.  I think
> > this is not a serious one but the attached patch to startup.el is
> > simple and it helps people identify a fontset problem earlier.

> Sorry, I forgot about this issue when you first raised it.  But I don't
> think it's serious enough to carve out an exception to the
> regressions-only rule.

Ok, I see.

---
Kenichi Handa
handa@m17n.org

Re: How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Richard M Stallman]

    Sorry, I forgot about this issue when you first raised it.  But I don't
    think it's serious enough to carve out an exception to the
    regressions-only rule.

If there is a bug in a change since Emacs 23, we should certainly fix it.

Re: How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Chong Yidong]

Richard M Stallman <rms@gnu.org> writes:

>     Sorry, I forgot about this issue when you first raised it.  But I don't
>     think it's serious enough to carve out an exception to the
>     regressions-only rule.
>
> If there is a bug in a change since Emacs 23, we should certainly fix
> it.

This is not a regression, I think.

Re: How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Richard M Stallman]

    > If there is a bug in a change since Emacs 23, we should certainly fix
    > it.

    This is not a regression, I think.

A bad wrinkle in new feature is never a regression,
but now (before it is released) is the best time to fix it.

Re: How to show *Warnings* buffer at startup time? [Re: Pretest next week]

[Chong Yidong]

Richard M Stallman <rms@gnu.org> writes:

>     > If there is a bug in a change since Emacs 23, we should certainly fix
>     > it.
>
>     This is not a regression, I think.
>
> A bad wrinkle in new feature is never a regression,

It's not a new feature either.