Comment 52 for bug 228988

On 2010-09-16 18:43, papukaija wrote :
> The files (of the atil site) are stored on the server with ISO-8859-1
> encoding (tested with wget) which is not recommended in Linux servers.
That is no reason for FF displaying the default character set
(ISO-8859-1) incorrectly.
FYI, the atlif's server is not a Linux but a Windows server.
> I'm attaching the frontpage to this report. That file is saved with
> UTF-8 encoding and Firefox displays that page correctly with UTF-8
> encoding but incorrectly with ISO-8859-1 encoding.
What you say is fuzzy. AM I correctly guessing what you did?
You took a correct ISO-8859-1 page, transliterated it to UTF-8 and
stored it as an invalid page. Then you displayed that page with FF and
the display was incorrect because UTF-8 was displayed using default
ISO-8859-1. Then you forced FF to display the page with View/Character
encoding UTF-8 and UTF-8 was displayed correctly.

Invalid pages displaying incorrectly and correctly when tweaked is
perfectly expected results but is unrelated with the bug description.
Why do you explain that?
> What I see, is that
> Firefox has troubles showing pages if the encoding set in FF doesn't
> match with the file's encoding, but that, I think, is obvious and not a
> bug.
There is no encoding set in FF that must match with the page's encoding
when it is loaded.
FF learns the encoding of the page through the declarations in it and
shows in View/Character Encoding (which is not a setting) the encoding
that it has used.
One may use View/Character Encoding *after* the page has displayed to
have FF redraw the page in another correct or incorrect character set
but this is totally unrelated with the bug.

The bug is that an ISO8859-1 encoded page that has to be displayed using
ISO8859-1 by default is *sometimes* (randomly) displayed using UTF-8. I
have shown that with much obviousness in the screenshots.