Bug#625323: collectd: ftbfs with gcc-4.6 -Werror

[Florian Forster]

Hello Peter,

 On Thu, Aug 11, 2011 at 11:52:02PM +0100, peter green wrote:
 As well as the issue mentioned in this bug report I found two other
 issues that were preventing successfull build.

thank you very much for your patch! I've committed the vast majority of it
verbatim as http://octo.cx/61a1fa9.

Here's what I changed:

 -buffer_ptr = fgets (buffer, sizeof (buffer), global_read_fh);
 +fgets (buffer, sizeof (buffer), global_read_fh);

Leave the buffer_ptr in there and use it to write a warning when an
unexpected EOF is read.

Neither the current collectd-4.10 nor the collectd-5.0 branch compile
without typename_len in src/plugin.c. I've dismissed this hunk completely.

Thanks again :)
—octo



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#625323: collectd: ftbfs with gcc-4.6 -Werror

[Colin Watson]

On Thu, Aug 11, 2011 at 11:52:02PM +0100, peter green wrote:
 As well as the issue mentioned in this bug report I found two other
 issues that were preventing successfull build.

Thanks for this!

I'm not the maintainer of this package, but I think the patch attached
to #614439 is a better fix for the 'struct xt_error_target' problem.

 diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' 
 '--exclude=.arch' '--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
 collectd-4.10.1~/src/teamspeak2.c collectd-4.10.1/src/teamspeak2.c
 --- collectd-4.10.1~/src/teamspeak2.c 2011-08-11 22:31:20.0 +
 +++ collectd-4.10.1/src/teamspeak2.c  2011-08-11 22:31:25.0 +
 @@ -296,9 +296,8 @@
  
   { /* Check that the server correctly identifies itself. */
   char buffer[4096];
 - char *buffer_ptr;
  
 - buffer_ptr = fgets (buffer, sizeof (buffer), global_read_fh);
 + fgets (buffer, sizeof (buffer), global_read_fh);
   buffer[sizeof (buffer) - 1] = 0;
  
   if (memcmp ([TS]\r\n, buffer, 6) != 0)

This causes a further build failure on Ubuntu where the toolchain
defaults to -Werror=unused-result.  Here's a patch to your patch to
avoid this problem, and I've attached a complete amended
gcc-4.6-fixes.dpatch.

--- debian/patches/gcc-4.6-fixes.dpatch.orig2011-08-15 12:28:08.964134920 
+0100
+++ debian/patches/gcc-4.6-fixes.dpatch 2011-08-15 11:55:35.0 +0100
@@ -264,14 +264,17 @@
 diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' 
'--exclude=.arch' '--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
collectd-4.10.1~/src/teamspeak2.c collectd-4.10.1/src/teamspeak2.c
 --- collectd-4.10.1~/src/teamspeak2.c  2011-08-11 22:31:20.0 +
 +++ collectd-4.10.1/src/teamspeak2.c   2011-08-11 22:31:25.0 +
-@@ -296,9 +296,8 @@
+@@ -296,9 +296,11 @@
  
{ /* Check that the server correctly identifies itself. */
char buffer[4096];
 -  char *buffer_ptr;
  
 -  buffer_ptr = fgets (buffer, sizeof (buffer), global_read_fh);
-+  fgets (buffer, sizeof (buffer), global_read_fh);
++  if (fgets (buffer, sizeof (buffer), global_read_fh) == NULL)
++  {
++  /* We'll detect failures by inspecting the buffer. */
++  }
buffer[sizeof (buffer) - 1] = 0;
  
if (memcmp ([TS]\r\n, buffer, 6) != 0)

Thanks,

-- 
Colin Watson   [cjwat...@ubuntu.com]
#! /bin/sh /usr/share/dpatch/dpatch-run
## gcc-4.6-fixes.dpatch by  root@localhost
##
## All lines beginning with `## DP:' are a description of the patch.
## DP: fixes for unused but set variable warnings from gcc-4.6

@DPATCH@
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' collectd-4.10.1~/src/disk.c 
collectd-4.10.1/src/disk.c
--- collectd-4.10.1~/src/disk.c 2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/disk.c  2011-08-11 22:31:25.0 +
@@ -426,7 +426,6 @@
int numfields;
int fieldshift = 0;
 
-   int major = 0;
int minor = 0;
 
counter_t read_sectors  = 0;
@@ -464,7 +463,6 @@
if ((numfields != (14 + fieldshift))  (numfields != 7))
continue;
 
-   major = atoll (fields[0]);
minor = atoll (fields[1]);
 
disk_name = fields[2 + fieldshift];
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' collectd-4.10.1~/src/java.c 
collectd-4.10.1/src/java.c
--- collectd-4.10.1~/src/java.c 2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/java.c  2011-08-11 22:31:25.0 +
@@ -2253,7 +2253,6 @@
   cjni_callback_info_t *cbi;
   jobject o_ocitem;
   const char *name;
-  int status;
   size_t i;
 
   jclass class;
@@ -2308,7 +2307,7 @@
   method = (*jvm_env)-GetMethodID (jvm_env, class,
   config, (Lorg/collectd/api/OConfigItem;)I);
 
-  status = (*jvm_env)-CallIntMethod (jvm_env,
+  (*jvm_env)-CallIntMethod (jvm_env,
   cbi-object, method, o_ocitem);
 
   (*jvm_env)-DeleteLocalRef (jvm_env, o_ocitem);
@@ -3065,9 +3064,8 @@
 
   if (config_block != NULL)
   {
-int status;
 
-status = cjni_config_perform (config_block);
+cjni_config_perform (config_block);
 oconfig_free (config_block);
 config_block = NULL;
   }
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
collectd-4.10.1~/src/libvirt.c collectd-4.10.1/src/libvirt.c
--- collectd-4.10.1~/src/libvirt.c  2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/libvirt.c   2011-08-11 22:31:25.0 +
@@ -655,8 +655,6 @@
 int i, n;
 const char *name;
 char uuid[VIR_UUID_STRING_BUFLEN];
-char  

Bug#625323: collectd: ftbfs with gcc-4.6 -Werror

[peter green]

tags 625323 +patch
thanks

As well as the issue mentioned in this bug report I found two other 
issues that were preventing successfull build.


Patches for all three issues and a new 00list file to apply them in the 
right order (the package uses dpatch) are attatched.


Note that since I don't use collectd (i'm just doing flyby rc bug 
squashing) the functionality of the packages built with theese patches 
has not been tested.
#! /bin/sh /usr/share/dpatch/dpatch-run
## gcc-4.6-fixes.dpatch by  root@localhost
##
## All lines beginning with `## DP:' are a description of the patch.
## DP: fixes for unused but set variable warnings from gcc-4.6

@DPATCH@
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' collectd-4.10.1~/src/disk.c 
collectd-4.10.1/src/disk.c
--- collectd-4.10.1~/src/disk.c 2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/disk.c  2011-08-11 22:31:25.0 +
@@ -426,7 +426,6 @@
int numfields;
int fieldshift = 0;
 
-   int major = 0;
int minor = 0;
 
counter_t read_sectors  = 0;
@@ -464,7 +463,6 @@
if ((numfields != (14 + fieldshift))  (numfields != 7))
continue;
 
-   major = atoll (fields[0]);
minor = atoll (fields[1]);
 
disk_name = fields[2 + fieldshift];
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' collectd-4.10.1~/src/java.c 
collectd-4.10.1/src/java.c
--- collectd-4.10.1~/src/java.c 2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/java.c  2011-08-11 22:31:25.0 +
@@ -2253,7 +2253,6 @@
   cjni_callback_info_t *cbi;
   jobject o_ocitem;
   const char *name;
-  int status;
   size_t i;
 
   jclass class;
@@ -2308,7 +2307,7 @@
   method = (*jvm_env)-GetMethodID (jvm_env, class,
   config, (Lorg/collectd/api/OConfigItem;)I);
 
-  status = (*jvm_env)-CallIntMethod (jvm_env,
+  (*jvm_env)-CallIntMethod (jvm_env,
   cbi-object, method, o_ocitem);
 
   (*jvm_env)-DeleteLocalRef (jvm_env, o_ocitem);
@@ -3065,9 +3064,8 @@
 
   if (config_block != NULL)
   {
-int status;
 
-status = cjni_config_perform (config_block);
+cjni_config_perform (config_block);
 oconfig_free (config_block);
 config_block = NULL;
   }
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
collectd-4.10.1~/src/libvirt.c collectd-4.10.1/src/libvirt.c
--- collectd-4.10.1~/src/libvirt.c  2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/libvirt.c   2011-08-11 22:31:25.0 +
@@ -655,8 +655,6 @@
 int i, n;
 const char *name;
 char uuid[VIR_UUID_STRING_BUFLEN];
-char  *host_ptr;
-size_t host_len;
 
 vl-time = t;
 vl-interval = interval_g;
@@ -664,8 +662,6 @@
 sstrncpy (vl-plugin, libvirt, sizeof (vl-plugin));
 
 vl-host[0] = '\0';
-host_ptr = vl-host;
-host_len = sizeof (vl-host);
 
 /* Construct the hostname field according to HostnameFormat. */
 for (i = 0; i  HF_MAX_FIELDS; ++i) {
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
collectd-4.10.1~/src/match_empty_counter.c 
collectd-4.10.1/src/match_empty_counter.c
--- collectd-4.10.1~/src/match_empty_counter.c  2011-08-11 22:31:20.0 
+
+++ collectd-4.10.1/src/match_empty_counter.c   2011-08-11 22:31:25.0 
+
@@ -73,7 +73,6 @@
 const value_list_t *vl,
 notification_meta_t __attribute__((unused)) **meta, void **user_data)
 {
-  mec_match_t *m;
   int num_counters;
   int num_empty;
   int i;
@@ -81,7 +80,6 @@
   if ((user_data == NULL) || (*user_data == NULL))
 return (-1);
 
-  m = *user_data;
 
   num_counters = 0;
   num_empty = 0;
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
collectd-4.10.1~/src/mysql.c collectd-4.10.1/src/mysql.c
--- collectd-4.10.1~/src/mysql.c2011-08-11 22:31:20.0 +
+++ collectd-4.10.1/src/mysql.c 2011-08-11 22:31:25.0 +
@@ -707,7 +707,6 @@
MYSQL_RES *res;
MYSQL_ROW  row;
char  *query;
-   intfield_num;
 
unsigned long long qcache_hits  = 0ULL;
unsigned long long qcache_inserts   = 0ULL;
@@ -743,7 +742,7 @@
if (res == NULL)
return (-1);
 
-   field_num = mysql_num_fields (res);
+   mysql_num_fields (res);
while ((row = mysql_fetch_row (res)))
{
char *key;
diff -urNad '--exclude=CVS' '--exclude=.svn' '--exclude=.git' '--exclude=.arch' 
'--exclude=.hg' '--exclude=_darcs' '--exclude=.bzr' 
collectd-4.10.1~/src/network.c collectd-4.10.1/src/network.c
--- collectd-4.10.1~/src/network.c 

Bug#625323: collectd: ftbfs with gcc-4.6 -Werror

[Luk Claes]

Hi

The problem seems to be that libupsclient (since 2.6.0) does not provide
the upscli_connect symbol anymore according to the config.log.

Not sure if that's true, but if it is, it very probably is a bug in
libupsclient (from source package nut).

Cheers

Luk



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#625323: collectd: ftbfs with gcc-4.6 -Werror

[Florian Forster]

Hi Matthias,

On Tue, May 03, 2011 at 10:32:36AM +, Matthias Klose wrote:
 This package builds with -Werror, and GCC 4.6 triggers new warnings
 which will make the package fail to build.

 The full build log can be found at:
 http://people.debian.org/~doko/tmp/werror/collectd_4.10.1-2.1_lsid64.buildlog

the bulk of those problems reside in code that's generated by libtool
(errors around checking dlopen) or autoconf (errors involving the
ac_aggr variable).

There are a couple or warnings which we can influence ourselves (the
getmnt checking code) and I'll take a look, but I don't see how this
should fail the build. In particular, I did not find any output of make,
only that configure exited with status 1. Unfortunately, in that big
dump of mostly useless data, I wasn't able to detect the output of
configure before exiting with that status.

Regards,
--octo
-- 
Florian octo Forster
Hacker in training
GnuPG: 0x0C705A15
http://octo.it/


signature.asc
Description: Digital signature