Re: emacs-22.2: regression, cannot compile with heimdal but can with mit-krb5

[Ulrich Mueller <ulm@gentoo.org>]

[-- Attachment #1: Type: text/plain, Size: 1260 bytes --]

>>>>> On Tue, 01 Apr 2008, Martin MOKREJŠ wrote:

> I would like to point you to the bug report at
> http://bugs.gentoo.org/show_bug.cgi?id=215558

> Briefly (credits to Michael Hammer (mueli)):

> Here is the definition of the krb5_error struct in mit-krb5 which
> seams compatible with emacs-22.2:
> [...]
>    krb5_data text;                     /* descriptive text */
>    krb5_data e_data;                   /* additional error-describing data */
> [...]

> and here is the heimdal code. You can see that the struct members
> are named differently: 
> [...]
>  heim_general_string *e_text;
>  heim_octet_string *e_data;
> [...]

Hi,

Two alternative ways have been suggested (see Gentoo bug mentioned
above) to make Emacs compile also with Heimdal:

   1. Honza Macháček has pointed out that there exists the
      following patch from T2, which uses "e_data" instead of "text":
      <http://svn.exactcode.de/t2/branches/7.0/package/security/heimdal/emacs-pop.diff>

   2. The patch (by myself) attached to this message, which adds an
      autoconf test, in order to determine if the struct contains a
      "text" or "e_text" member. This compiles and links against
      MIT Kerberos or against Heimdal.

Ulrich

[-- Attachment #2: emacs-22.2-pop-heimdal.patch --]
[-- Type: text/plain, Size: 1906 bytes --]

--- emacs-22.2-orig/configure.in	2008-04-02 18:28:37.000000000 +0200
+++ emacs-22.2/configure.in	2008-04-02 22:18:38.000000000 +0200
@@ -2700,7 +2700,9 @@
   fi
 
   if test "${with_kerberos5+set}" = set; then
-    AC_CHECK_HEADERS(krb5.h)
+    AC_CHECK_HEADERS(krb5.h,
+      AC_CHECK_MEMBERS([krb5_error.text, krb5_error.e_text],,,
+		       [#include <krb5.h>]))
   else
     AC_CHECK_HEADERS(des.h,,
 		     [AC_CHECK_HEADERS(kerberosIV/des.h,,
--- emacs-22.2-orig/lib-src/pop.c	2008-02-23 14:49:00.000000000 +0100
+++ emacs-22.2/lib-src/pop.c	2008-04-02 22:35:55.000000000 +0200
@@ -1200,11 +1200,12 @@
       krb5_free_principal (kcontext, server);
       if (rem)
 	{
+	  strcpy (pop_error, KRB_ERROR);
+	  strncat (pop_error, error_message (rem),
+		   ERROR_MAX - sizeof (KRB_ERROR));
+#if defined HAVE_KRB5_ERROR_TEXT
 	  if (err_ret && err_ret->text.length)
 	    {
-	      strcpy (pop_error, KRB_ERROR);
-	      strncat (pop_error, error_message (rem),
-		       ERROR_MAX - sizeof (KRB_ERROR));
 	      strncat (pop_error, " [server says '",
 		       ERROR_MAX - strlen (pop_error) - 1);
 	      strncat (pop_error, err_ret->text.data,
@@ -1213,12 +1214,17 @@
 	      strncat (pop_error, "']",
 		       ERROR_MAX - strlen (pop_error) - 1);
 	    }
-	  else
+#elif defined HAVE_KRB5_ERROR_E_TEXT
+	  if (err_ret && err_ret->e_text && strlen(*err_ret->e_text))
 	    {
-	      strcpy (pop_error, KRB_ERROR);
-	      strncat (pop_error, error_message (rem),
-		       ERROR_MAX - sizeof (KRB_ERROR));
+	      strncat (pop_error, " [server says '",
+		       ERROR_MAX - strlen (pop_error) - 1);
+	      strncat (pop_error, *err_ret->e_text,
+		       ERROR_MAX - strlen (pop_error) - 1);
+	      strncat (pop_error, "']",
+		       ERROR_MAX - strlen (pop_error) - 1);
 	    }
+#endif
 	  if (err_ret)
 	    krb5_free_error (kcontext, err_ret);
 	  krb5_auth_con_free (kcontext, auth_context);