Mailing List Archive: Votes for 0.8.12...

Votes for 0.8.12...

Aug 29, 1995, 10:02 AM

Post #1 of 33 (1862 views)

Add whatever comments you like, but please keep these in the same order...

+1 patch11.access_cleanup --- Eliminate compiler warning from mod_access.c.
[Robert Thau]
+1 patch11.bad-namesrvr --- Behave more sanely when the name server loses
its mind [Sean Welch]
+1 patch11.bsdi-rfc931 --- rfc931 code compiles cleanly on old BSDI releases
[Randy Terbush]
-1 patch11.bsdi-longsz --- handle fsize include directive on old BSDI releases
(does not compile correctly, does not work under other systems).
[Bob Thrush]
+1 patch11.cgi-cleanup --- Delete inaccurate comment [Robert Thau]
+1 patch11.config-script --- Make Configure script accept Module lines which
end in trailing whitespace. [Robert Thau]
+1 patch11.dyn-load --- Properly configure dynamically loaded modules
[R. W. Shufflebotham]
+1 patch11.htaccess-first --- Lets .htaccess override <Directory> in
access control, etc. [David Robinson]
+1 patch11.imap_point --- add "point" directive to imagemap module [James
Cloos]
+1 patch11.inc-cgi-pause --- Don't pause three seconds after SSI includes
output of a CGI script or command which is too slow to die off.
[Robert Thau]
+1 patch11.include-opt --- properly honor Options Includes (or more
accurately, properly honor turning it off). [David Robinson]
+1 patch11.include-text --- Make IncludesNoExec work properly, by being
willing to include at least text/*. [Roy Fielding]
+1 patch11.lookup_uri --- Dike out speed hack in sub_req_lookup_uri which
was keeping some marginal cases (e.g., ScriptAlias of a *particular*
index.html file) from working properly [David Robinson]
+1 patch11.no-auth-no-error --- Don't log an error when the requested URI
requires auth, but the client sent no auth line; this is not an error
(the client just didn't know auth was needed here), and shouldn't
be logged as such. [Robert Thau]
+1 patch11.no-bsd-conf --- Delete BSD #define from conf.h everyplace; also
cures the bsdi-longsz problem above. [ Randy Terbush ]
+1 patch11.rfc931 --- After a request which had ident data is processed,
don't give the same name on future requests, which do not. [David
Robinson]
+1 patch11.virt-srvr-name --- Default ServerName for virtual hosts if
not specified. [Robert Thau]

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 7:16 AM

Post #2 of 33 (1858 views)

At 10:02 AM 8/29/95 EDT, you wrote:
>Add whatever comments you like, but please keep these in the same order...
>
>-1 patch11.access_cleanup --- Eliminate compiler warning from mod_access.c.
> [Robert Thau]
This didn't work for me.... I'm still having a problem with it.
>+1 patch11.bad-namesrvr --- Behave more sanely when the name server loses
> its mind [Sean Welch]
>+1 patch11.bsdi-rfc931 --- rfc931 code compiles cleanly on old BSDI releases
> [Randy Terbush]
>-1 patch11.bsdi-longsz --- handle fsize include directive on old BSDI releases
> (does not compile correctly, does not work under other systems).
> [Bob Thrush]
Didn't compile for me either, didn't include it.
>+1 patch11.cgi-cleanup --- Delete inaccurate comment [Robert Thau]
>+1 patch11.config-script --- Make Configure script accept Module lines which
> end in trailing whitespace. [Robert Thau]
>+1 patch11.dyn-load --- Properly configure dynamically loaded modules
> [R. W. Shufflebotham]
>+1 patch11.htaccess-first --- Lets .htaccess override <Directory> in
> access control, etc. [David Robinson]
>+1 patch11.imap_point --- add "point" directive to imagemap module [James
> Cloos]
>+1 patch11.inc-cgi-pause --- Don't pause three seconds after SSI includes
> output of a CGI script or command which is too slow to die off.
> [Robert Thau]
>+1 patch11.include-opt --- properly honor Options Includes (or more
> accurately, properly honor turning it off). [David Robinson]
>+1 patch11.include-text --- Make IncludesNoExec work properly, by being
> willing to include at least text/*. [Roy Fielding]
>+1 patch11.lookup_uri --- Dike out speed hack in sub_req_lookup_uri which
> was keeping some marginal cases (e.g., ScriptAlias of a *particular*
> index.html file) from working properly [David Robinson]
>+1 patch11.no-auth-no-error --- Don't log an error when the requested URI
> requires auth, but the client sent no auth line; this is not an error
> (the client just didn't know auth was needed here), and shouldn't
> be logged as such. [Robert Thau]
>0 patch11.no-bsd-conf --- Delete BSD #define from conf.h everyplace; also
> cures the bsdi-longsz problem above. [ Randy Terbush ]
Didn't compile in. Having trouble with all BSD patches.
>+1 patch11.rfc931 --- After a request which had ident data is processed,
> don't give the same name on future requests, which do not. [David
> Robinson]
>+1 patch11.virt-srvr-name --- Default ServerName for virtual hosts if
> not specified. [Robert Thau]
--
Aram W. Mirzadeh, MIS Manager, Qosina Corporation
http://www.qosina.com/~awm/, awm@qosina.com

Re: Votes for 0.8.12... [ In reply to ]

randy at zyzzyva

Aug 29, 1995, 8:09 AM

Post #3 of 33 (1855 views)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 8:34 AM

Post #4 of 33 (1856 views)

At 10:09 AM 8/29/95 -0500, you wrote:
>> >0 patch11.no-bsd-conf --- Delete BSD #define from conf.h everyplace; also
>> > cures the bsdi-longsz problem above. [ Randy Terbush ]
>> Didn't compile in. Having trouble with all BSD patches.
>
>Did the patch fail, or did it fail to compile, and if the later,
>what OS?

I had to apply it. Well, change the patch file so it would patch to the
right file.
And then it didn't compile.

Linux 1.2

<Aram>
--
Aram W. Mirzadeh, MIS Manager, Qosina Corporation
http://www.qosina.com/~awm/, awm@qosina.com

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 10:09 AM

Post #5 of 33 (1855 views)

In reply to Robert S. Thau who said
>
> -1 patch11.bsdi-longsz --- handle fsize include directive on old BSDI releases
> (does not compile correctly, does not work under other systems).
> [Bob Thrush]

I think this patch will work though it's a bit cludgy

> +1 patch11.no-bsd-conf --- Delete BSD #define from conf.h everyplace; also
> cures the bsdi-longsz problem above. [ Randy Terbush ]

This won't work because BSD4_4 is a botched define in that it was
also defined for BSD systems between 4.3 and 4.4, i.e. Net/2 and
possible Reno releases so you can't rely on all 4.4 features being
present even if it's defined. It was defined to be 0.5 in pre-4.4
releases and 1 in the real 4.4 release. CSRG don't screw up often
but they certainly did this time.

I prefer the approach taken in this patch though, it just doesn't quite
work.

I'm looking at all this now, hopefully I'll have something sane worked
out by the end of the night but it'll probably touch too much for
the .12 release.
--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 10:09 AM

Post #6 of 33 (1855 views)

In reply to Robert S. Thau who said
>
> -1 patch11.bsdi-longsz --- handle fsize include directive on old BSDI releases
> (does not compile correctly, does not work under other systems).
> [Bob Thrush]

I think this patch will work though it's a bit cludgy

> +1 patch11.no-bsd-conf --- Delete BSD #define from conf.h everyplace; also
> cures the bsdi-longsz problem above. [ Randy Terbush ]

This won't work because BSD4_4 is a botched define in that it was
also defined for BSD systems between 4.3 and 4.4, i.e. Net/2 and
possible Reno releases so you can't rely on all 4.4 features being
present even if it's defined. It was defined to be 0.5 in pre-4.4
releases and 1 in the real 4.4 release. CSRG don't screw up often
but they certainly did this time.

I prefer the approach taken in this patch though, it just doesn't quite
work.

I'm looking at all this now, hopefully I'll have something sane worked
out by the end of the night but it'll probably touch too much for
the .12 release.
--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

randy at zyzzyva

Aug 29, 1995, 10:29 AM

Post #7 of 33 (1855 views)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 10:45 AM

Post #8 of 33 (1853 views)

At 11:57 AM 8/29/95 EDT, you wrote:
>>-1 patch11.access_cleanup --- Eliminate compiler warning from mod_access.c.
>> [Robert Thau]
> This didn't work for me.... I'm still having a problem with it.
>
>What sort of problem? (This patch adds one #include to mod_access.c;
>its sole purpose is to eliminate a compiler warning, and it's been included
>with a bunch of the block patches we've got from other people. It surely
>has nothing to do with whatever problems you're having with includes...).
>

Sorry, wrong patch, change that to a +1. Here is what I need for the linux
system
to work properly. I know it's stupid, but it seems to be what's causing the
problem.
I'm not sure if I did it right or not... could someone please check it
before putting it
up as a patch though. ( It does work, I'm not sure of the impact on other
systems).

patch1:

*** util.c Mon Aug 28 15:47:29 1995
--- util.c.orig Mon Aug 28 15:47:08 1995
***************
*** 765,771 ****
#ifndef MINIMAL_DNS
hptr = gethostbyaddr((char *)iaddr, sizeof(struct in_addr), AF_INET);
if(hptr) {
! conn->remote_host = pstrdup(conn->pool, (void *)hptr->h_name);
conn->remote_name = conn->remote_host;
str_tolower (conn->remote_host);
}
--- 765,771 ----
#ifndef MINIMAL_DNS
hptr = gethostbyaddr((char *)iaddr, sizeof(struct in_addr), AF_INET);
if(hptr) {
! conn->remote_host = pstrdup(conn->pool, hptr->h_name);
conn->remote_name = conn->remote_host;
str_tolower (conn->remote_host);
}
***************
*** 835,841 ****
return pstrdup(a, p->h_aliases[x]);
}
return NULL;
! } else return pstrdup(a, (void *)p->h_name);
}

char *get_local_host(pool *a)
--- 835,841 ----
return pstrdup(a, p->h_aliases[x]);
}
return NULL;
! } else return pstrdup(a, p->h_name);
}

char *get_local_host(pool *a)

patch2:

*** http_main.c Mon Aug 28 15:46:58 1995
--- http_main.c.orig Mon Aug 28 15:46:45 1995
***************
*** 662,668 ****
struct hostent *h = gethostbyaddr ((char *)&(s->host_addr),
sizeof (struct in_addr),
AF_INET);
! s->server_hostname = pstrdup (pconf, (void *)h->h_name);
}
}

--- 662,668 ----
struct hostent *h = gethostbyaddr ((char *)&(s->host_addr),
sizeof (struct in_addr),
AF_INET);
! s->server_hostname = pstrdup (pconf, h->h_name);
}
}

--
Aram W. Mirzadeh, MIS Manager, Qosina Corporation
http://www.qosina.com/~awm/, awm@qosina.com

Re: Votes for 0.8.12... [ In reply to ]

randy at zyzzyva

Aug 29, 1995, 10:46 AM

Post #9 of 33 (1857 views)

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 10:58 AM

Post #10 of 33 (1858 views)

In reply to Randy Terbush who said
>
> Your correct, this is bogus on BSDI 1.1. *sigh*
>
> The problem I was running into was that Sun would not Grok the
> comparison of a value for BSD ( ie BSD > 199305).
>
> The only other solution I can come up with would be to add
> #define HAS_QUAD_T or somesuch.

Having though about this for a while, I think the following
proposal is the way to go but it's not going to be ready for the
next release.

Each non-portable problem has a flag, i.e. HAS_QUAD_T, etc. This
would allow the actual code to be written without worrying about what
OS's it runs on.

Each OS has a config file, e.g.

config-freebsd.h:

#include <sys/param.h>
#define HAS_QUAD_T
#define HAS_FOO

The configure script runs uname to get an id string for the system and
then copies the appropriate config- file to os.h

Note that os.h would be OS specific values only, there should be a
config.h that allows the admin to decide what modules to build, what the
default paths are etc. Some of this should be moved from httpd.h.

This keeps all OS specific flags in one place that the person for that OS
can find and maintain easily. People writing "core" code don't have to add
flags all over the place for each OS, if they run into portability
problems they just add a single global flag and each of the OS porters
can check whether they need to add it to the OS specific config file.
There should be a documentation file that all the flags are listed in
explaining what they do.

I'll implement this as soon as I can if people are interested. It's
basically what other packages have done.

--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 11:17 AM

Post #11 of 33 (1855 views)

+1 patch11.access_cleanup
+1 patch11.bad-namesrvr
+1 patch11.bsdi-rfc931
+1 patch11.cgi-cleanup
+1 patch11.config-script
+1 patch11.htaccess-first
+1 patch11.imap_point
+1 patch11.inc-cgi-pause
+1 patch11.include-opt
+1 patch11.include-text
+1 patch11.no-auth-no-error
+1 patch11.virt-srvr-name

Missing patches - I either didn't try them or understand them.

rob

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 11:55 AM

Post #12 of 33 (1855 views)

In reply to Aram W. Mirzadeh who said
>
> Sorry, wrong patch, change that to a +1. Here is what I need for the linux
> system
> to work properly. I know it's stupid, but it seems to be what's causing the
> problem.

What exactly doesn't work. If that patch is necessary to make things
work on Linux I'd suspect a Linux compiler bug that casting
to void * is masking.

By the way, your patch was reversed.

--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 11:57 AM

Post #13 of 33 (1855 views)

>-1 patch11.access_cleanup --- Eliminate compiler warning from mod_access.c.
> [Robert Thau]
This didn't work for me.... I'm still having a problem with it.

What sort of problem? (This patch adds one #include to mod_access.c;
its sole purpose is to eliminate a compiler warning, and it's been included
with a bunch of the block patches we've got from other people. It surely
has nothing to do with whatever problems you're having with includes...).

rst

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 12:30 PM

Post #14 of 33 (1856 views)

If I accessed a page without asking for it... ie (http://www.qosina.com:8080)
it would ask me if I wanted to download it....

I had x-parse-html turned on to .html files. I had to put a .htaccess file
in it, to
get it to open the file currectly.

Anyway, the previous patches also fixed two warnings that the compiler gave
when compiling. I looked there first, and added the (void *) to the lines, and
then tried it without the .htaccess file and it worked. Some how it's
overriding
the memory location of the parser??! No idea why it's happening, or how
it got fixed. I'm in no way a programmer to be able to tell you why. :(

<Aram>

At 07:55 PM 8/29/95 +0100, you wrote:
>In reply to Aram W. Mirzadeh who said
>>
>> Sorry, wrong patch, change that to a +1. Here is what I need for the linux
>> system
>> to work properly. I know it's stupid, but it seems to be what's causing the
>> problem.
>
>What exactly doesn't work. If that patch is necessary to make things
>work on Linux I'd suspect a Linux compiler bug that casting
>to void * is masking.
>
>By the way, your patch was reversed.
>
>--
> Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
> Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
> Phone: 0370 462071 (Mobile), +44 1222 457651 (home)
>
>
--
Aram W. Mirzadeh, MIS Manager, Qosina Corporation
http://www.qosina.com/~awm/, awm@qosina.com

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 12:47 PM

Post #15 of 33 (1857 views)

In reply to Rob Hartill who said
>
>
> > Note that os.h would be OS specific values only, there should be a
> > config.h that allows the admin to decide what modules to build, what the
> > default paths are etc. Some of this should be moved from httpd.h.
>
> > I'll implement this as soon as I can if people are interested. It's
> > basically what other packages have done.
>
> Sounds reasonable. I also like your idea on function definitions,

I'll start working on this.

>
> int
> foo(...)
>
> there's nothing worse than grep'ing for a function definition and
> getting back a face full of matches in response.

I'll do this in my own code but touching other people's code tends
to get you into trouble when dealing with coding style, I'd really
like to see this principle picked by the other contributers though,
it does make life much easier. If Rob is amenable then I'll also
convert any code I touch as I go through it. Actually, if the general
concensus is that this should be done then Rob'll just have to accept
it :-)

--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 12:50 PM

Post #16 of 33 (1857 views)

Paul Richards wrote..

> Each OS has a config file, e.g.
>
> config-freebsd.h:
>
> #include <sys/param.h>
> #define HAS_QUAD_T
> #define HAS_FOO
>
> The configure script runs uname to get an id string for the system and
> then copies the appropriate config- file to os.h
>
> Note that os.h would be OS specific values only, there should be a
> config.h that allows the admin to decide what modules to build, what the
> default paths are etc. Some of this should be moved from httpd.h.

> I'll implement this as soon as I can if people are interested. It's
> basically what other packages have done.

Sounds reasonable. I also like your idea on function definitions,

int
foo(...)

there's nothing worse than grep'ing for a function definition and
getting back a face full of matches in response.

rob

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 12:57 PM

Post #17 of 33 (1856 views)

In reply to Aram W. Mirzadeh who said
>
>
> If I accessed a page without asking for it... ie (http://www.qosina.com:8080)
> it would ask me if I wanted to download it....
>
> I had x-parse-html turned on to .html files. I had to put a .htaccess file
> in it, to
> get it to open the file currectly.
>
> Anyway, the previous patches also fixed two warnings that the compiler gave
> when compiling. I looked there first, and added the (void *) to the lines, and
> then tried it without the .htaccess file and it worked. Some how it's
> overriding
> the memory location of the parser??! No idea why it's happening, or how
> it got fixed. I'm in no way a programmer to be able to tell you why. :(

Can you send me the definition of struct hostent ? It should be
in /usr/include/netdb.h I don't know about the other problems but the
compiler warning is easy to check.

--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 1:17 PM

Post #18 of 33 (1855 views)

Paul,

Here is the struct you asked for:

/usr/include/netdb.h
[stuff deleted]
/*
* Structures returned by network data base library. All addresses are
* supplied in host order, and returned in network order (suitable for
* use in system calls).
*/
struct hostent {
__const
char *h_name; /* official name of host */
char **h_aliases; /* alias list */
int h_addrtype; /* host address type */
int h_length; /* length of address */
char **h_addr_list; /* list of addresses from name server */
#define h_addr h_addr_list[0] /* address, for backward compatiblity */
};

[stuff deleted]

At 08:57 PM 8/29/95 +0100, you wrote:
>In reply to Aram W. Mirzadeh who said
>>
>>
>> If I accessed a page without asking for it... ie (http://www.qosina.com:8080)
>> it would ask me if I wanted to download it....
>>
>> I had x-parse-html turned on to .html files. I had to put a .htaccess file
>> in it, to
>> get it to open the file currectly.
>>
>> Anyway, the previous patches also fixed two warnings that the compiler gave
>> when compiling. I looked there first, and added the (void *) to the
lines, and
>> then tried it without the .htaccess file and it worked. Some how it's
>> overriding
>> the memory location of the parser??! No idea why it's happening, or how
>> it got fixed. I'm in no way a programmer to be able to tell you why. :(
>
>Can you send me the definition of struct hostent ? It should be
>in /usr/include/netdb.h I don't know about the other problems but the
>compiler warning is easy to check.
>
>--
> Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
> Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
> Phone: 0370 462071 (Mobile), +44 1222 457651 (home)
>
>
--
Aram W. Mirzadeh, MIS Manager, Qosina Corporation
http://www.qosina.com/~awm/, awm@qosina.com

Re: Votes for 0.8.12... [ In reply to ]

paul at netcraft

Aug 29, 1995, 1:27 PM

Post #19 of 33 (1854 views)

In reply to Robert S. Thau who said
>
> Paul, I'm not responding to email right now because I'm trying to get some
> Real Work(tm) done, but I personally do *not* think that it's the right
> thing to go ahead with a major rework of the configuration stuff now. I
> think people here are losing sight of the reason why we have a feature freeze
> on now --- it is to avoid making changes to the code which cause new problems
> while we are trying to chase out the old ones. So, for instance, imap-points
> doesn't bug me so much --- it's simple, self-contained, and has minimal
> impact on what's already there.
>
> However, a massive rework of the way the conf.h stuff works is *certain* to
> cause such problems, and, for that reason, while I do understand the desire
> for it, I do *not* think it would be a good idea to proceed with this sort
> of thing until after 1.0. (If you want to get it ready in the meantime, I'd
> appreciate it, but for 1.0, I feel strongly that the right thing is to stick
> with incremental adaptations of what we have).

Hmm, I certainly had no intention of doing this for an immediate release,
but I think this is enough of a mess that it should be fixed before a
"significant" release such as 1.0, it comes back to the discussion of
releases I guess. If 1.0 is pretty close then I'll just wait until after
that to present my changes.

Anyway, this would be something I go away and work on behind the
scenes and I'll present it when it's rock solid on at least the platforms
I have access to.

--
Paul Richards, Bluebird Computer Systems. FreeBSD core team member.
Internet: paul@FreeBSD.org, http://www.freebsd.org/~paul
Phone: 0370 462071 (Mobile), +44 1222 457651 (home)

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 3:55 PM

Post #20 of 33 (1858 views)

Paul, I'm not responding to email right now because I'm trying to get some
Real Work(tm) done, but I personally do *not* think that it's the right
thing to go ahead with a major rework of the configuration stuff now. I
think people here are losing sight of the reason why we have a feature freeze
on now --- it is to avoid making changes to the code which cause new problems
while we are trying to chase out the old ones. So, for instance, imap-points
doesn't bug me so much --- it's simple, self-contained, and has minimal
impact on what's already there.

However, a massive rework of the way the conf.h stuff works is *certain* to
cause such problems, and, for that reason, while I do understand the desire
for it, I do *not* think it would be a good idea to proceed with this sort
of thing until after 1.0. (If you want to get it ready in the meantime, I'd
appreciate it, but for 1.0, I feel strongly that the right thing is to stick
with incremental adaptations of what we have).

rst

Re: Votes for 0.8.12... [ In reply to ]

brian at organic

Aug 29, 1995, 4:57 PM

Post #21 of 33 (1858 views)

In the interested of validating my +1 votes, I only gave +1's to patches
I actually tested, or find mostly cosmetic.

On Tue, 29 Aug 1995, Robert S. Thau wrote:
> Add whatever comments you like, but please keep these in the same order...
>
> +1 patch11.access_cleanup --- Eliminate compiler warning from mod_access.c.
> [Robert Thau]
cosmetic.

> +1 patch11.bad-namesrvr --- Behave more sanely when the name server loses
> its mind [Sean Welch]
tested - I killed my nameserver and then completely restarted httpd, and
it returned "Cannot resolve host name taz.apache.org --- exiting!". Much
preferable to a core dump.

> +1 patch11.bsdi-rfc931 --- rfc931 code compiles cleanly on old BSDI releases
> [Randy Terbush]
compiles fine on BSDI 2.0

> 0 patch11.bsdi-longsz --- handle fsize include directive on old BSDI releases
> (does not compile correctly, does not work under other systems).
> [Bob Thrush]
not tested, obsoleted apparently by other patches

> +1 patch11.cgi-cleanup --- Delete inaccurate comment [Robert Thau]
cosmetic

> +1 patch11.config-script --- Make Configure script accept Module lines which
> end in trailing whitespace. [Robert Thau]
tested

> 0 patch11.dyn-load --- Properly configure dynamically loaded modules
> [R. W. Shufflebotham]
> 0 patch11.htaccess-first --- Lets .htaccess override <Directory> in
> access control, etc. [David Robinson]
not tested

> +1 patch11.imap_point --- add "point" directive to imagemap module [James
> Cloos]
> +1 patch11.inc-cgi-pause --- Don't pause three seconds after SSI includes
> output of a CGI script or command which is too slow to die off.
> [Robert Thau]
tested

> 0 patch11.include-opt --- properly honor Options Includes (or more
> accurately, properly honor turning it off). [David Robinson]
> 0 patch11.include-text --- Make IncludesNoExec work properly, by being
> willing to include at least text/*. [Roy Fielding]
> 0 patch11.lookup_uri --- Dike out speed hack in sub_req_lookup_uri which
> was keeping some marginal cases (e.g., ScriptAlias of a *particular*
> index.html file) from working properly [David Robinson]
not tested

> +1 patch11.no-auth-no-error --- Don't log an error when the requested URI
> requires auth, but the client sent no auth line; this is not an error
> (the client just didn't know auth was needed here), and shouldn't
> be logged as such. [Robert Thau]
tested

> +1 patch11.no-bsd-conf --- Delete BSD #define from conf.h everyplace; also
> cures the bsdi-longsz problem above. [ Randy Terbush ]
with the caveats as posted in my previous message - HAS_GMTOFF has to be
defined in the __bsdi__ switch in conf.h.

> 0 patch11.rfc931 --- After a request which had ident data is processed,
> don't give the same name on future requests, which do not. [David
> Robinson]
> 0 patch11.virt-srvr-name --- Default ServerName for virtual hosts if
> not specified. [Robert Thau]
not tested.

Brian

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

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 5:14 PM

Post #22 of 33 (1859 views)

Votes (void if patch file changed):

350 Aug 27 15:16 patch11.access_cleanup +1
658 Aug 28 14:33 patch11.bad-namesrvr +1
3209 Aug 25 02:19 patch11.bsdi-rfc931 +1
1433 Aug 28 14:33 patch11.bsdi-longsz 0
502 Aug 27 15:16 patch11.cgi-cleanup +1
553 Aug 27 15:17 patch11.config-script +1
2956 Aug 25 02:19 patch11.dyn-load 0 very ugly fix
2873 Aug 25 02:19 patch11.htaccess-first +1
1347 Aug 25 02:22 patch11.imap_point -1 feature enhancement (+)
2831 Aug 27 15:16 patch11.inc-cgi-pause +1
810 Aug 25 02:19 patch11.include-opt +1
1981 Aug 25 02:19 patch11.include-text +1
4339 Aug 25 23:23 patch11.lookup_uri +1
587 Aug 27 17:14 patch11.no-auth-no-error +1
8831 Aug 29 14:20 patch11.no-bsd-conf +1 *
2470 Aug 25 02:19 patch11.rfc931 +1
1895 Aug 27 17:14 patch11.virt-srvr-name -1 ** bug

* Ok, but remove the #include <sys/files.h> from the #if defined(OSF1)
section. (I put it in to get round a BSD problem, which no longer exists.)

** Nearly there, but it needs to check the return value from gethostbyaddr(),
otherwise Apache could core dump. (As in the bug fixed by bad-namesrvr.)

+ We should be concentrating on bug fixes for now.

David.

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 7:16 PM

Post #23 of 33 (1857 views)

From: "Aram W. Mirzadeh" <awm@qosina.com>
Reply-To: new-httpd@hyperreal.com

If I accessed a page without asking for it... ie (http://www.qosina.com:8080)
it would ask me if I wanted to download it....

I had x-parse-html turned on to .html files. I had to put a .htaccess file
in it, to
get it to open the file currectly.

Anyway, the previous patches also fixed two warnings that the compiler gave
when compiling. I looked there first, and added the (void *) to the lines, and
then tried it without the .htaccess file and it worked. Some how it's
overriding
the memory location of the parser??! No idea why it's happening, or how
it got fixed. I'm in no way a programmer to be able to tell you why. :(

<Aram>

Hmmm... on Linux, at least (or most machines), the casts you added
should not have had *any* effect on the generated code, though they
might have eliminated a warning or two. (I'm not sure they should
have done even that --- char* and void* are supposed to interchange
without warning in ANSI C).

What is more likely to have happened is that there was some sort of
version skew between versions of the object files which were lying
around, and doing an extra 'make' cleared them up...

rst

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 7:30 PM

Post #24 of 33 (1859 views)

From: Paul Richards <paul@netcraft.co.uk>
Date: Tue, 29 Aug 1995 18:09:45 +0100 (BST)
In reply to Robert S. Thau who said

This won't work because BSD4_4 is a botched define in that it was
also defined for BSD systems between 4.3 and 4.4, i.e. Net/2 and
possible Reno releases so you can't rely on all 4.4 features being
present even if it's defined. It was defined to be 0.5 in pre-4.4
releases and 1 in the real 4.4 release. CSRG don't screw up often
but they certainly did this time.

Is this the only thing wrong with Randy's patch (aside from not
setting HAS_GMTOFF in some of the BSD cases (!), as Brian has recently
discovered)?

I may well be missing something here, but if my reading of this
paragraph is correct, Randy's approach can be made to work as follows:

#ifdef BSD_4_4
#if BSD_4_4 > 0.5
sprintf(tag,"%qd",finfo.st_size);
#else
sprintf(tag,"%ld",finfo.st_size);
#endif /* *real* BSD_4_4 */
#else
sprintf(tag,"%ld",finfo.st_size);
#endif

This will, of course, win no prizes for elegance or concision;
however, if it will in fact work on all BSD-oid systems, I'd much
prefer it. The advantage is that it doesn't involve messing around
much with the rest of the code, including the config defines for
systems which we don't necessarily have (and which I *really* don't
want to mess with until post-1.0, at which point conf.h clearly does
deserve serious rework).

rst

Re: Votes for 0.8.12... [ In reply to ]

Aug 29, 1995, 7:30 PM

Post #25 of 33 (1854 views)

From: Paul Richards <paul@netcraft.co.uk>
Date: Tue, 29 Aug 1995 18:09:45 +0100 (BST)
In reply to Robert S. Thau who said

This won't work because BSD4_4 is a botched define in that it was
also defined for BSD systems between 4.3 and 4.4, i.e. Net/2 and
possible Reno releases so you can't rely on all 4.4 features being
present even if it's defined. It was defined to be 0.5 in pre-4.4
releases and 1 in the real 4.4 release. CSRG don't screw up often
but they certainly did this time.

Is this the only thing wrong with Randy's patch (aside from not
setting HAS_GMTOFF in some of the BSD cases (!), as Brian has recently
discovered)?

I may well be missing something here, but if my reading of this
paragraph is correct, Randy's approach can be made to work as follows:

#ifdef BSD_4_4
#if BSD_4_4 > 0.5
sprintf(tag,"%qd",finfo.st_size);
#else
sprintf(tag,"%ld",finfo.st_size);
#endif /* *real* BSD_4_4 */
#else
sprintf(tag,"%ld",finfo.st_size);
#endif

This will, of course, win no prizes for elegance or concision;
however, if it will in fact work on all BSD-oid systems, I'd much
prefer it. The advantage is that it doesn't involve messing around
much with the rest of the code, including the config defines for
systems which we don't necessarily have (and which I *really* don't
want to mess with until post-1.0, at which point conf.h clearly does
deserve serious rework).

rst