Mailing List Archive

On Tue, 2 May 1995, Robert S. Thau wrote:
> I'll add a +1 for E88 and E89 (haven't tried the code, and I'll want to
> before we publically release it, but these things are REAL hard to botch).

+1 on both as well, as long as they work. :)

Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com brian@hyperreal.com http://www.[hyperreal,organic].com/

I'll add a +1 for E88 and E89 (haven't tried the code, and I'll want to
before we publically release it, but these things are REAL hard to botch).

rst

the only thing undecided so far is B90 - a fix for the timefmt
server-side include bug that side-stepped an earlier attempted fix.

If/when someone else okays it, I'd like to build 0.6.3 so that we can
plug it in for our own testing.

Oh, one useful fix still waiting to be patched is the "no multi"
message in the error log. rst, you added it, can you fix it ?

Cliff/Randy, there's a bug report for the virtual host stuff
in the apache-bugs mailbox.

robh.
-=-=-=-=

(n.b. some +1s in the table might be missing, but the patch already
has 3 votes so it doesn't matter)

Votes for 0.6.3 bb rst robh drtr rt rm cs

B73 (virtual hosts) +1 0 0 0 +1 +1 IN
B77 (symlink) 0 -1 0 0 OUT
B79 (asis) +1 +1 +1 +1 IN
B81 (AIX) +1 +1 +1 +1 IN
B82 (redirect) +1 +1 +1 +1 IN
B83 (next) +1 +1 +1 IN
B84 (content-type) +1 +1 +1 IN

E70 (handler) -1 -1 -1 OUT
E88 (arena workaround) +1 +1 +1 IN
E89 (netscape " ) +1 +1 +1 IN
B90 (timefmt inc bug) +1 +1 ????

O58 (makefile) -1 -1 -1 OUT

Fork_Free.txt -1 -1 -1 OUT :-)
logresolve.c +1 +1 +1 IN

P67 (group dbm) +1 +1 +1 +1 IN

What do you mean you can't keep up ??? :-)


Votes for 0.6.3 bb rst robh drtr rt rm cs

B73 (virtual hosts) +1 0 0 0 +1 +1 IN
B77 (symlink) 0 -1 0 0 OUT
B79 (asis) +1 +1 +1 +1 IN
B81 (AIX) +1 +1 +1 +1 IN
B82 (redirect) +1 +1 +1 +1 IN
B83 (next) +1 +1 +1 IN
B84 (content-type) +1 +1 +1 IN

E70 (handler) -1 -1 -1 OUT
E88 (arena workaround) +1 +1 +1 IN
E89 (netscape " ) +1 +1 +1 IN
B90 (timefmt inc bug) +1 +1 ????
B92 (SYSV cgi exit) +1 ????

O58 (makefile) -1 -1 -1 OUT

Fork_Free.txt -1 -1 -1 OUT :-)
logresolve.c +1 +1 +1 IN

P67 (group dbm) +1 +1 +1 +1 IN

On Wed, 3 May 1995, Rob Hartill wrote:
> B90 (timefmt inc bug) +1 +1 ????
> B92 (SYSV cgi exit) +1 ????

+1 on both.

Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com brian@hyperreal.com http://www.[hyperreal,organic].com/

YAU - yet another update.

those marked ???? are in need of votes

-=-=


Votes for 0.6.3 bb rst robh drtr rt rm cs

B73 (virtual hosts) +1 +1 +1 IN
B77 (symlink) -1 OUT
B79 (asis) +1 +1 +1 +1 IN
B81 (AIX) +1 +1 +1 +1 IN
B82 (redirect) +1 +1 +1 +1 IN
B83 (next) +1 +1 +1 IN
B84 (content-type) +1 +1 +1 IN

E70 (handler) -1 -1 -1 OUT
E88 (arena workaround) +1 +1 +1 IN
E89 (netscape " ) +1 +1 +1 IN
B90 (timefmt inc bug) +1 +1 +1 IN
B92 (SYSV cgi exit) +1 +1 ????
E93 (AddType spaces) +1 ????

O58 (makefile) -1 -1 -1 OUT
O91 (makefile CFLAG/name) +1 ????

Fork_Free.txt -1 -1 -1 OUT :-)
logresolve.c +1 +1 +1 IN

P67 (group dbm) +1 +1 +1 +1 IN

Votes for 0.6.3 bb rst robh drtr rt rm cs

B73 (virtual hosts) +1 +1 +1 IN
B77 (symlink) -1 OUT
B79 (asis) +1 +1 +1 +1 IN
B81 (AIX) +1 +1 +1 +1 IN
B82 (redirect) +1 +1 +1 +1 IN
B83 (next) +1 +1 +1 IN
B84 (content-type) +1 +1 +1 IN

E70 (handler) -1 -1 -1 OUT
E88 (arena workaround) +1 +1 +1 IN
E89 (netscape " ) +1 +1 +1 IN
B90 (timefmt inc bug) +1 +1 +1 IN
B92 (SYSV cgi exit) +1 +1 ????
E93 (AddType spaces) +1 +1 ????

O58 (makefile) -1 -1 -1 OUT
O91 (makefile CFLAG/name)+1 +1 ????

Fork_Free.txt -1 -1 -1 OUT :-)
logresolve.c +1 +1 +1 IN

P67 (group dbm) +1 +1 +1 +1 IN




--
Rob Hartill
http://nqcd.lanl.gov/~hartill/

+1 on B92. As for O91, I'm not actually sure it's a good idea --- I can
see a reason for being able to #ifdef out the virtual-host stuff (if it has
portability problems), but if it doesn't cause problems for those who don't
use it (and it doesn't seem to), I really think it ought to be in by default.

rst

> > Date: Thu, 4 May 95 12:08 BST
> > From: drtr@ast.cam.ac.uk (David Robinson)
> >
> > O91 (makefile tidy) +1 but could it also remove the warning about Solaris
> > please?
> >
> >Hmmm... it's a bit late, but I'd like to elevate my reservations about
> >this to a formal -0.5 --- the virtual host stuff seems to be one of
> >our big selling points, and I'm not sure it's wise to have it off by
> >default. If everyone else votes +1, on the other hand, I can live
> >with it...

I also would prefer to see VirtualHost _on_ by default. It seems
to be a fairly inocuous piece of code and would prefer not to
have to have that variable to address when hit with "why doesn't
VirtualHost work?" questions.

I would vote +1 on O91 if we could remove the #ifdef VIRTUALHOST from
the code.

On a related note, when do we plan to wrap this release? If it
will be a couple of days, I could finalize the other Makefile
approach and wrap these changes.

Comments?

>
> > >Hmmm... it's a bit late, but I'd like to elevate my reservations about
> > >this to a formal -0.5 --- the virtual host stuff seems to be one of
> > >our big selling points, and I'm not sure it's wise to have it off by
> > >default. If everyone else votes +1, on the other hand, I can live
> > >with it...
> >
> > In which case, why is it a compile-time option? Why not remove the
> > #if's from the code, or alternatively, #define the constants in httpd.h.
>
> If the #ifdefs get removed, it might be a good idea to clearly mark
> the multi-host code as such, so that when HTTP/1.1 finally arrives,
> it'll be easy to yank the code.
>
> robh

Just a comment...

With the lag in browsers adopting the newer standards, can we really
expect to take advantage of these features before a majority of the
browsers have begun using them? I still have not found MultiViews
as useful as it could be without the browsers doing the right thing.

Am I missing something?

BTW - I don't disagree with Rob's suggestion to mark this code.

> >Hmmm... it's a bit late, but I'd like to elevate my reservations about
> >this to a formal -0.5 --- the virtual host stuff seems to be one of
> >our big selling points, and I'm not sure it's wise to have it off by
> >default. If everyone else votes +1, on the other hand, I can live
> >with it...
>
> In which case, why is it a compile-time option? Why not remove the
> #if's from the code, or alternatively, #define the constants in httpd.h.

If the #ifdefs get removed, it might be a good idea to clearly mark
the multi-host code as such, so that when HTTP/1.1 finally arrives,
it'll be easy to yank the code.

robh

Date: Thu, 4 May 95 12:08 BST
From: drtr@ast.cam.ac.uk (David Robinson)

O91 (makefile tidy) +1 but could it also remove the warning about Solaris
please?

Hmmm... it's a bit late, but I'd like to elevate my reservations about
this to a formal -0.5 --- the virtual host stuff seems to be one of
our big selling points, and I'm not sure it's wise to have it off by
default. If everyone else votes +1, on the other hand, I can live
with it...

rst

In which case, why is it a compile-time option? Why not remove the
#if's from the code, or alternatively, #define the constants in httpd.h.

Agreed completely... (though I'd prefer to #define the things in httpd.h
for 0.6.3 just to minimize the probability of typos, and then delete the
#ifs in 0.7).

rst

O91 (makefile tidy) +1 but could it also remove the warning about Solaris
please?
B92 (SYSV) +1
B94 (multi message) +1
E93 (content-parms) 0

Re E93, which allows spaces in the AddType command.

1. You aren't allowed spaces in the mime.types file; i.e.

text/html; version=2.0 html
doesn't work, whereas
text/html;version=2.0 html
does. I'm not quite sure of the syntax of mime.types, as I couldn't
find any documentation of it anywhere.

2. Parameters on magic types.

httpd doesn't allow parameters on the magic mime types. For example
AddType text/x-server-parsed-html;level=3 .shtml
isn't recognised as a server-parsed included file.

I notice that there is a new magic type of
text/x-server-parsed-html3
but the client still only gets a content-type of text/html when this is used.

David.

> Date: Thu, 4 May 95 12:08 BST
> From: drtr@ast.cam.ac.uk (David Robinson)
>
> O91 (makefile tidy) +1 but could it also remove the warning about Solaris
> please?
>
>Hmmm... it's a bit late, but I'd like to elevate my reservations about
>this to a formal -0.5 --- the virtual host stuff seems to be one of
>our big selling points, and I'm not sure it's wise to have it off by
>default. If everyone else votes +1, on the other hand, I can live
>with it...

In which case, why is it a compile-time option? Why not remove the
#if's from the code, or alternatively, #define the constants in httpd.h.

For testing purposes, I can see that a compile-time option is useful.
For code with portability problems, then you have to switch it at compile time.
(e.g. the dbm stuff).

I'm not sure that the multi-home stuff falls into either of these catagories
any more.

David.