pfopen - failure to open fatal?

[Brian Behlendorf <br...@hyperreal.org>]

PR #817 notes that .htaccess files are ignored if they are unreadable.  I
think
this should be fixed; though I'm unsure if it's something we should keep
specific to .htaccess file parsing, or implement it lower-level, in
pfopen().  Thoughts?

	Brian


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
"Why not?" - TL           brian@organic.com - hyperreal.org - apache.org

Re: pfopen - failure to open fatal?

[Brian Behlendorf <br...@organic.com>]

I'm going for the "members-only" jacket with the embroidered logo I get at
10,000 points.  

At 11:01 PM 7/20/97 -0600, you wrote:
>Oh, while I remember, you are building up lots more Apache Guru Points
>Brian.  They will be dropping from the sky on your head shortly.  Anyone
>else who wants more should deal with PRs too. 

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
"Why not?" - TL           brian@organic.com - hyperreal.org - apache.org

Re: pfopen - failure to open fatal?

[Marc Slemko <ma...@worldgate.com>]

It should not be done in pfopen().  The only thing that could possibly be
done there is some changes to make error codes easier to find; would have
to look, probably not necessary. 

Every call to *open() should be examined to be sure different types of
failure are handled correctly.  The trick with .htaccess is that some
types of failure (eg. file doesn't exist) should be handled one way,
others (eg. permissions deny access) a different way.  

On Sun, 20 Jul 1997, Brian Behlendorf wrote:

> PR #817 notes that .htaccess files are ignored if they are unreadable.  I
> think
> this should be fixed; though I'm unsure if it's something we should keep
> specific to .htaccess file parsing, or implement it lower-level, in
> pfopen().  Thoughts?
> 
> 	Brian
> 
> 
> --=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
> "Why not?" - TL           brian@organic.com - hyperreal.org - apache.org
> 

Re: pfopen - failure to open fatal?

[Marc Slemko <ma...@worldgate.com>]

On Mon, 21 Jul 1997, Rob Hartill wrote:

> Ignoring an unreadible .htaccess file doesn't sound too bad to me. I
> assume you want to log a warning and that's it.

Ignoring an unreadable .htaccess is bad by the idea of least exposure.
Better to be overly restrictive rather than overly permissive.  Say you
accidentally forget to make a htaccess world readable; the server knows it
is there yet can't deal with it, so it should refuse requests.  Say your
disk blows up so trying to read the .htaccess returns a read error.  The
server knows it is there yet can't read it, so it should refuse requests.

Oh yes, does anyone think the server error message should be expanded a
little to say "tell the admin to check the error log for possible
details"?

Re: pfopen - failure to open fatal?

[Rob Hartill <ro...@imdb.com>]

On Sun, 20 Jul 1997, Brian Behlendorf wrote:

> PR #817 notes that .htaccess files are ignored if they are unreadable.  I
> think
> this should be fixed; though I'm unsure if it's something we should keep
> specific to .htaccess file parsing, or implement it lower-level, in
> pfopen().  Thoughts?


What is the "fix" ?, you only seem to be discussing where to put it.

Ignoring an unreadible .htaccess file doesn't sound too bad to me. I
assume you want to log a warning and that's it.

rob

Re: pfopen - failure to open fatal?

[Marc Slemko <ma...@worldgate.com>]

Oh, while I remember, you are building up lots more Apache Guru Points
Brian.  They will be dropping from the sky on your head shortly.  Anyone
else who wants more should deal with PRs too. 

On Sun, 20 Jul 1997, Brian Behlendorf wrote:

> PR #817 notes that .htaccess files are ignored if they are unreadable.  I
> think
> this should be fixed; though I'm unsure if it's something we should keep
> specific to .htaccess file parsing, or implement it lower-level, in
> pfopen().  Thoughts?
> 
> 	Brian
> 
> 
> --=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
> "Why not?" - TL           brian@organic.com - hyperreal.org - apache.org
>