Mailing List Archive: No PatchID yet, but....

No PatchID yet, but....

brian at organic

Apr 29, 1995, 6:34 PM

Post #1 of 8 (1473 views)

Rob Thau's patch for the problem we were seeing with
"application/x-httpd-cgi" for in-directory .cgi scripts did indeed fix the
problem (it was a pretty basic patch, so that's not surprising). This gets
my +1, it's so small I wonder if it needs a patch ID or just a three "yay"'s
from the crowd here. I would strongly suggest putting it in, as the bug is
ugly on some of our imagemap scripts here.

*** http_mime.c~ Sat Apr 15 14:05:39 1995
--- http_mime.c Wed Apr 26 18:53:14 1995
***************
*** 312,317 ****
+ * the former on neatness grounds, the latter because */
+
+ strcpy (content_type, default_type);
+
while(1) {
/* why did this use getline, when we can use stdio buffered I/O?? */
/* maybe we should put a timeout on this. */

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

Re: No PatchID yet, but.... [ In reply to ]

Apr 30, 1995, 9:29 AM

Post #2 of 8 (1469 views)

That doesn't have a patch ID yet, because I'm still not sure that
plugging in the DefaultType is the right thing --- the guy did say
that lynx botches redirects with *anything* other than "text/html",
in which case we'll have to use that as the default Content-type:
on at least 302 returns in order to guarantee interoperability with
lynx (and, quite possibly, X Mosaic 2.6b, if the NCSA 1.4 hypernews
bug pages are to be believed). I'll be playing with this a bit this
morning... maybe more in an hour or so.

rst

Re: No PatchID yet, but.... [ In reply to ]

Apr 30, 1995, 10:31 AM

Post #3 of 8 (1465 views)

NEVERMIND... I tried my patch out with both lynx and Mosaic 2.6b1 and a
script which set Location: but no Content-type: --- I even installed it
as index.cgi just to be sure (and downloaded and compiled the latest version
of lynx I could find, to be sure of THAT). The patch seems to work, so it's
now on hyperreal (well, actually, the comment in the version there is a
real, complete sentence!).

rst

Sigh --- it would be nice to confirm this with the guy ahead of releasing,
but if it fixes the bug for other people who have observed it, and we can't
duplicate the problem anymore with this fix, well...

Re: No PatchID yet, but.... [ In reply to ]

Apr 30, 1995, 10:45 AM

Post #4 of 8 (1464 views)

Whoops --- forgot the ID. The patch is B84, in the for_Apache_0.6.2
directory, and of course, I give it a +1. Sigh...

rst

Re: No PatchID yet, but.... [ In reply to ]

Apr 30, 1995, 4:11 PM

Post #5 of 8 (1466 views)

>Whoops --- forgot the ID. The patch is B84, in the for_Apache_0.6.2
>directory, and of course, I give it a +1. Sigh...
+1 from me.

David.

Re: No PatchID yet, but.... [ In reply to ]

brian at organic

May 2, 1995, 2:13 PM

Post #6 of 8 (1472 views)

On Sun, 30 Apr 1995, Robert S. Thau wrote:
> That doesn't have a patch ID yet, because I'm still not sure that
> plugging in the DefaultType is the right thing

It does satisfy my definition of "default"...

> the guy did say
> that lynx botches redirects with *anything* other than "text/html",
> in which case we'll have to use that as the default Content-type:
> on at least 302 returns in order to guarantee interoperability with
> lynx (and, quite possibly, X Mosaic 2.6b, if the NCSA 1.4 hypernews
> bug pages are to be believed). I'll be playing with this a bit this
> morning... maybe more in an hour or so.

Yes, it was causing problems on XMosaic and Lynx (the former prompting
you to save the object passed with the 302 to disk("this object has
moved...", yet still opening the connection anyways, the latter flailing
completely). Specs?

Brian

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

Re: No PatchID yet, but.... [ In reply to ]

brian at organic

May 2, 1995, 3:22 PM

Post #7 of 8 (1462 views)

On Tue, 2 May 1995, Robert S. Thau wrote:
> Date: Tue, 2 May 1995 14:13:33 -0700 (PDT)
> From: Brian Behlendorf <brian@organic.com>
>
> Yes, it was causing problems on XMosaic and Lynx (the former prompting
> you to save the object passed with the 302 to disk("this object has
> moved...", yet still opening the connection anyways, the latter flailing
> completely). Specs?
>
> And, just to dot the i's, the patch does cure the problem? (With my
> server, which has the patch, I couldn't duplicate it with either lynx
> or Mosaic, but I would like to verify this with someone who has, in
> fact, seen the problem at least once, just out of sheer paranoia).

Yes.

Brian

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

Re: No PatchID yet, but.... [ In reply to ]

May 2, 1995, 6:14 PM

Post #8 of 8 (1464 views)

Date: Tue, 2 May 1995 14:13:33 -0700 (PDT)
From: Brian Behlendorf <brian@organic.com>

Yes, it was causing problems on XMosaic and Lynx (the former prompting
you to save the object passed with the 302 to disk("this object has
moved...", yet still opening the connection anyways, the latter flailing
completely). Specs?

And, just to dot the i's, the patch does cure the problem? (With my
server, which has the patch, I couldn't duplicate it with either lynx
or Mosaic, but I would like to verify this with someone who has, in
fact, seen the problem at least once, just out of sheer paranoia).

rst