vgough / rlog Goto Github PK

Hello Valient!

Please tell me can I still using this library in production system?!
Are you preparing a new release or enhance a library?!

Regards,
Karol Szkudlarek

What steps will reproduce the problem?
1. compile the attached file (requires rlog & boost)
2. run it with ./a.out -c debug/

What is the expected output? What do you see instead?

expected: 
... [debug/foo/bar/AA] ... A(a)::execute(1)
... [debug/foo/bar/BB/b1] ... B(b1)::execute(10)
... [debug/foo/bar/BB/b2] ... B(b2)::execute(100)

actual (the last channel is b1 instead of b2):
... [debug/foo/bar/AA] ... A(a)::execute(1)
... [debug/foo/bar/BB/b1] ... B(b1)::execute(10)
... [debug/foo/bar/BB/b1] ... B(b2)::execute(100)

What version of the product are you using? On what operating system?
1.4

See:

https://bugzilla.novell.com/show_bug.cgi?id=784948
http://lists.gnu.org/archive/html/bug-rcs/2011-11/msg00000.html

What steps will reproduce the problem?
1. Open project
2. Build

What is the expected output? What do you see instead?
I expect it to compile. Can not find Rlog.lib. It generates Rlog.dll but
nothing else.

What version of the product are you using? On what operating system?
1.4 on Windows XP.

Current support is for VC7.1.

Patch from David Wolfe.

Details pasted to http://pastebin.ca/1346940

Please, is this issue a duplicate of issue 7?

I attached a patch that resolves that problem

Hello,

We have found seriouos problem with logging in several threads and we look
at the sources and find that problem exists in rMessageCall:

# define _rMessageCall(ID, CHANNEL, ...) \
  if(unlikely(ID.publish!=0))  \                       // 1
  { \
    (*ID.publish)( &ID, CHANNEL, ##__VA_ARGS__ ); \    // 2
    rlog::__checkArgs( 0, ##__VA_ARGS__ ); \
  }


two threads can enter in //1 but it can happen that one thread could zero
ID.publish (in //2) and second will crash in //2.

what can you suggest? Maybe some patch?

Regards,
Karol Szkudlarek

What steps will reproduce the problem?
1. define a new channel made up of single letters: 
   chn = DEF_CHANNEL("debug/A/B/CC", rlog::Log_Debug);
2. use the channel for logging:
   rLog(chn, "foo");

What is the expected output? What do you see instead?
expect: channel name as [debug/A/B/CC]
actual: channel name is shown as [debug/CC]

What version of the product are you using? On what operating system?
1.4

Please provide any additional information below.

What steps will reproduce the problem?
1. compile the attatched files

What is the expected output? What do you see instead?
make rlogtest
g++ rlogtest.cc -o rlogtest -I ~/extlib/rlog-1.4/include/
-DRLOG_COMPONENT=”rlog” -L ~/extlib/rlog-1.4/lib/ -lrlog
rlogtest.cc:17: error: _rL_17 causes a section type conflict
make: *** [rlogtest] Error 1

If I place a rlog statement in a class member I can not compile the
programm any more!

What version of the product are you using? On what operating system?

RLog Version 1.4 - June 14, 2008 release on Ubuntu 8.10 with g++4.2 on a
64bit machine.

Please provide any additional information below.

I have no problem to compile the same code with rlog 1.3.7
What's my mistake?

What steps will reproduce the problem?
1. Doxygen is installed, but not latex or pdflatex
2. configure, make fails with missing target for latex/refman.pdf 
3. This also occurs if you pass --disable-docs to configure. 

Please provide any additional information below.

Testing head version from SVN, on OSX 10.5, as of 14 Apr 2009

Hello!

I'm stuck when I'm  trying to write simple function dotest with 
argument name channel and try to log through rLog function to
different channels. RLog library can log only to first specified channel.
The problem arises when you subscribe to a channel in object constructor,
to log each instance of an object to different channel.

Please look at attached test case file (testlib.cpp)

This issue is of course independent of operating system.

Regards,

This is not a problem but a feature request.

What is the expected output? What do you see instead?
I am using rlog in a production project. The executable will be released to
the clients, but not the source code. When logging in INFO, I see the name
of my cpp file and the line number. This is fine in DEBUG, but when I use
INFO I want to hide the name of file/line number. Is there a way to do it
in rlog?

What version of the product are you using? On what operating system?
rlog on Windows XP.

builtin_expect wasn't getting used when building on GCC4, so the compiler
hints weren't being presented.  Fixed in development tree.

Once fixed, dormant log messages are down to as little as 2 clock cycles on
my Pentium-M.

What steps will reproduce the problem?

1. svn checkout http://rlog.googlecode.com/svn/trunk/ rlog
2. cd rlog
3. autoreconf
4. ./configure
5. make

What is the expected output? What do you see instead?

/bin/sh ../libtool --tag=CXX   --mode=compile g++ -DHAVE_CONFIG_H -I. -I..   
-I..   -DRLOG_COMPONENT="rlog" -DUSE_VALGRIND=0 -g -O2 -D_THREAD_SAFE   -MT 
rlog.lo -MD -MP -MF .deps/rlog.Tpo -c -o rlog.lo rlog.cpp
../libtool: line 771: X--tag=CXX: command not found
../libtool: line 804: libtool: ignoring unknown tag : command not found

What version of the product are you using? On what operating system?

svn head

Mac OS X 10.6.5 on x86 (Darwin 10.5.0 Darwin Kernel Version 10.5.0: Fri Nov  5 
23:20:39 PDT 2010; root:xnu-1504.9.17~1/RELEASE_I386 i386)

Please provide any additional information below.

Running "cp /usr/bin/glibtool libtool" after configure seems to fix the 
problem. I guess an incompatible libtool somewhat is configured.

What steps will reproduce the problem?
1. Download rlog.
2. ./configure
3. make

What is the expected output? What do you see instead?

g++ -DHAVE_CONFIG_H -I. -I..   -I..   -DRLOG_COMPONENT="rlog"
-DUSE_VALGRIND=0 -g -O2 -D_THREAD_SAFE   -MT test.o -MD -MP -MF
.deps/test.Tpo -c -o test.o test.cpp
/var/tmp//cctyxeyY.s:335:Expected comma after segment-name
/var/tmp//cctyxeyY.s:335:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:348:Expected comma after segment-name
/var/tmp//cctyxeyY.s:348:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:361:Expected comma after segment-name
/var/tmp//cctyxeyY.s:361:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:374:Expected comma after segment-name
/var/tmp//cctyxeyY.s:374:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:387:Expected comma after segment-name
/var/tmp//cctyxeyY.s:387:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:400:Expected comma after segment-name
/var/tmp//cctyxeyY.s:400:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:413:Expected comma after segment-name
/var/tmp//cctyxeyY.s:413:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:426:Expected comma after segment-name
/var/tmp//cctyxeyY.s:426:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:439:Expected comma after segment-name
/var/tmp//cctyxeyY.s:439:Rest of line ignored. 1st junk character valued 32
( ).
/var/tmp//cctyxeyY.s:452:Expected comma after segment-name
/var/tmp//cctyxeyY.s:452:Rest of line ignored. 1st junk character valued 32
( ).
make[2]: *** [test.o] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

What version of the product are you using? On what operating system?

OS X 10.5.3

Please provide any additional information below.

What steps will reproduce the problem?
1. Try to build on OS X
2. Build fails

What is the expected output? What do you see instead?
Build error:
error: argument to 'section' attribute is not valid for
      this target: mach-o section specifier requires a segment and section separated by a comma

which I trace back to
  # define RLOG_SECTION __attribute__ (( section("RLOG_DATA") ))

What version of the product are you using? On what operating system?
OS X 10.7, clang 3.0. GCC 4.7 also gave a build error, but in the assembler 
phase.


Please provide any additional information below.
Changing this to:
  # define RLOG_SECTION __attribute__ (( section(".rlogseg,RLOG_DATA") ))

Gets it to build

What version of the product are you using? On what operating system?
1.4

Please provide any additional information below.

What steps will reproduce the problem?
1. Try a compile using VS2005 where the project path includes spaces in it.
ie: "c:\Users\c\Documents\Visual Studio 2005\Projects\rlog-1.3.7"

What is the expected output? What do you see instead?
Visual studio will abort the compile because the copy on the custom build
step for common.h.msvc71 will fail.



The fix is to put double quotes around the path in the build step:
was:
copy "$InputPath" ..\rlog\common.h

should be:
copy "$(InputPath)" ..\rlog\common.h

Just wanted to let the maintainers know.

One of features introduced in GCC 4.3 was that it doesn't include by
default many header files it included in previous versions.

This issue may be easily fixed - just add necessary includes (it doesn't
break compilation with other GCC versions).

Patch against ver. 1.3 included (very, very easiest patch, actually :))

We typically use daily log files one per process. This
proves to be quite convenient for tracking down problems.
The log4cxx library has a daily rolling file appender that
supports this. 
http://logging.apache.org/log4cxx/apidocs/classlog4cxx_1_1_daily_rolling_file_ap
pender.html

What would be the best way to add this capability to rlog?

Thanks,

Gerry Wiener

What steps will reproduce the problem?

1. patch -p1 < rlog-section-attibutes.patch
2. autoreconf
3. ./configure
4. make
5. make install

What is the expected output? What do you see instead?

make
cd . && /bin/sh /usr/src/rlog-1.4/config/missing --run aclocal-1.6 -I 
config
cd . && \
  /bin/sh /usr/src/rlog-1.4/config/missing --run automake-1.6 --foreign  
Makefile
configure.ac:45: `automake requires `AM_CONFIG_HEADER', not 
`AC_CONFIG_HEADER'
configure.ac: required file `config/mkinstalldirs' not found

What version of the product are you using? On what operating system?

rlog 1.4
OSX 10.4.8

Please provide any additional information below.

After applying the rlog section attibutes patch I ran autoreconf and that 
gave me the following output:

configure.ac:45: `automake requires `AM_CONFIG_HEADER', not 
`AC_CONFIG_HEADER'
configure.ac: required file `config/mkinstalldirs' not found
rlog/Makefile.am:21: variable `USER_LDFLAGS' not defined

After moving the mkdinstalldirs file to the config/ folder I can actually 
run ./configure, make and make install but when using rlog (in for example 
encfs) it could not find librlog:

checking for RLOG... configure: WARNING: Checking for librlog the hard way
checking for RLogVersion in -lrlog... no
configure: error: EncFS depends on librlog, by the same author.

ls of /usr/local/lib

librlog -> librlog.5
librlog.5
librlog.5.0.0 -> librlog.5
librlog.la

Looks like it is missing the dylib file; after creating a link 
(librlog.dylib -> librlog.5) I could use rlog.