Re: C file recoginzed as image file

[Juanma Barranquero]

On 1/9/07, Stefan Monnier <address@hidden> wrote:

> Content detection sometimes works well, but sometimes not.

Content detection usually works. After all, that's what an image
library does when it opens an image file and checks that the contents
seem OK, or Microsoft Word when it opens a .doc or .rtf file, etc.
Content detection based on regexp matching of a few bytes of the file
header is flaky, though ;)

> OTOH name-based
> detection overall works more reliably because that's what Unix has used for
> many years.

Well, nobody is suggesting throwing off `auto-mode-alist' and putting
`magic-mode-alist' in its place...

> I think that giving precedence to one over the other is an error.

I happen to disagree, but it is largely irrelevant to this discussion.
I prefer to concentrate my efforts in resolving the issue of the image
libraries, with the hope that we release Emacs 22 someday this year (I
almost wrote "month", but I don't want to be overly optimistic).

I stand for what I said earlier: let's disable image auto-detection,
add a note to NEWS (with big warnings about viruses, or whatever
people feels its necessary) detailing how to enable it, and get over
it.

                   /L/e/k/t/u