You are viewing a plain text version of this content. The canonical link for it is here.
Posted to apreq-dev@httpd.apache.org by Joe Schaefer <jo...@sunstarsys.com> on 2002/01/23 12:57:42 UTC

1.0 in CVS?

A few days ago I added a patch for 5.005thread support.
Are there any problems with the current CVS that would
prevent us from releasing it?

-- 
Joe Schaefer


Re: 1.0 in CVS?

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Jim Winstead <ji...@apache.org> writes:

> On Wed, Jan 23, 2002 at 07:25:05AM -0500, Joe Schaefer wrote:
> > The most recent patches were doc patches, mainly for Request.pm and 
> > Cookie.pm (and the ToDo). The INSTALL document is probably wrong wrt autoconf 
> > support- we don't currently include a working ./configure script.  I put 
> > some text in README to help C programmers roll their own libapreq.a, but
> > it's hardly an optimal solution.
> 
> running automake/autoconf is part of the release/tarball process. since
> it is a generated file, configure is not checked into cvs.
> 
> 0.33 included what should be a working configure script.

Thanks; we probably need to backout the most recent changes I made to 
MANIFEST then (I commented out a few lines to shut perl up about missing
files).  Also some of the README text I added regarding a C build probably
should be elided as well.


I just tried to build a copy of libapreq-0.33 I downloaded from CPAN,
using

  % ./configure
  ...
  % make
  cd . && aclocal
  cd . && automake --foreign Makefile
  automake: configure.in: required file `./missing' not found
  make: *** [Makefile.in] Error 1
  %

There's a symlink from ./missing to /usr/local/share/automake/missing in 
apreq's main directory, but there is no such file on my system (a RH Linux
derivative).  Relinking it to /usr/share/automake/missing (which is
where RH installed it) seems to work as expected.  Is there an easy fix
for this, or should we document the potential problem instead?

I haven't tested the resulting libraries at all- just the build 
process.

-- 
Joe Schaefer


Re: 1.0 in CVS?

Posted by Jim Winstead <ji...@apache.org>.
On Wed, Jan 23, 2002 at 07:25:05AM -0500, Joe Schaefer wrote:
> The most recent patches were doc patches, mainly for Request.pm and 
> Cookie.pm (and the ToDo). The INSTALL document is probably wrong wrt autoconf 
> support- we don't currently include a working ./configure script.  I put 
> some text in README to help C programmers roll their own libapreq.a, but
> it's hardly an optimal solution.

running automake/autoconf is part of the release/tarball process. since
it is a generated file, configure is not checked into cvs.

0.33 included what should be a working configure script.

jim

Re: 1.0 in CVS?

Posted by Simon Tamás <si...@westel.hu>.
Hello, I don't have access to the CVs repository, but I have a
suggestion

in C language library there's a file called apache_cookie.h
If you try to use this code fram C++ you wil find that the `extern "C"' 
is missing from this header. (It is added correctly in other header
files

here is my - corrected - version of this file

Please add these changes to the release!

Best regards

Tamas Simon

-*-


#ifndef _APACHE_COOKIE_H
#define _APACHE_COOKIE_H

#include "apache_request.h"
typedef array_header ApacheCookieJar;

typedef struct {
    request_rec *r;
    char *name;
    array_header *values;
    char *domain;
    char *expires;
    char *path;
    int secure;
} ApacheCookie;

#ifdef  __cplusplus
 extern "C" {
#endif 


#define ApacheCookieJarItems(arr) arr->nelts

#define ApacheCookieJarFetch(arr,i) \
((ApacheCookie *)(((ApacheCookie **)arr->elts)[i]))

#define ApacheCookieJarAdd(arr,c) \
*(ApacheCookie **)ap_push_array(arr) = c

#define ApacheCookieItems(c) c->values->nelts

#define ApacheCookieFetch(c,i) \
((char *)(((char **)c->values->elts)[i]))

#define ApacheCookieAddn(c,val) \
    if(val) *(char **)ap_push_array(c->values) = (char *)val

#define ApacheCookieAdd(c,val) \
    ApacheCookieAddn(c, ap_pstrdup(c->r->pool, val))

#define ApacheCookieAddLen(c,val,len) \
    ApacheCookieAddn(c, ap_pstrndup(c->r->pool, val, len))

ApacheCookie *ApacheCookie_new(request_rec *r, ...);
ApacheCookieJar *ApacheCookie_parse(request_rec *r, const char *data);
char *ApacheCookie_as_string(ApacheCookie *c);
char *ApacheCookie_attr(ApacheCookie *c, char *key, char *val);
char *ApacheCookie_expires(ApacheCookie *c, char *time_str);
void ApacheCookie_bake(ApacheCookie *c);

#ifdef __cplusplus
 }
#endif

#define APC_ERROR APLOG_MARK, APLOG_NOERRNO|APLOG_ERR, c->r

#endif // _APACHE_COOKIE_H

Re: 1.0 in CVS?

Posted by Joe Schaefer <jo...@sunstarsys.com>.
Issac Goldstand <ma...@beamartyr.net> writes:

> Joe Schaefer wrote:
> 
> >A few days ago I added a patch for 5.005thread support.
> >Are there any problems with the current CVS that would
> >prevent us from releasing it?
> >
> I haven't had any problems (yet), but I haven't aggressively tested 
> everything yet, either...  Is the documentation complete and up-to-date? 
>  I know I've run into that before a lot.
> 
>   Issac

The most recent patches were doc patches, mainly for Request.pm and 
Cookie.pm (and the ToDo). The INSTALL document is probably wrong wrt autoconf 
support- we don't currently include a working ./configure script.  I put 
some text in README to help C programmers roll their own libapreq.a, but
it's hardly an optimal solution.

To facilitate testing, I tarballed the current CVS.  You should be able
to grab it from

  http://www.apache.org/~joes/libapreq-1.0-rc1.tar.gz

Feedback, even if it's positive, is always appreciated :-)

-- 
Joe Schaefer