Re: [Nautilus-list] Nautilus Performance Analysis

[Darin Adler <darin bentspoon com>]

On Tuesday, August 14, 2001, at 11:07  AM, Alex Larsson wrote:

> The problem is in eel_scalable_font_new ():
>
> EelScalableFont *
> eel_scalable_font_new (const char *file_name)
> {
> 	EelScalableFont *font;
> 	gpointer font_handle_as_pointer;
> 	RsvgFTFontHandle font_handle = -1;
>
> 	g_return_val_if_fail (eel_strlen (file_name) > 0, NULL);
> 	g_return_val_if_fail (eel_font_manager_file_is_scalable_font 
> (file_name), NULL);
>
> 	initialize_global_stuff_if_needed ();
>
> The eel_font_manager_file_is_scalable_font() call results in a call to
> gnome-vfs mimetype magic sniffing that opens the file and reads the
> beginning of it.

Such a test is completely inappropriate for g_return_val_if_fail, if it 
results in I/O. We ought to fix this egregious mistake even if we also 
solve the real problem with your patch. The best fix is to simply remove 
this line.

> Here is my patch for this:

The patch looks great. Please commit it (and the removed line) on both 
branches if you would.

Your use of (!x) goes against the Nautilus coding style (used for Eel) 
which requires (x == NULL), but perhaps we can stop trying to be 
consistent about that. I don't feel strongly about it.

    -- Darin