summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
...
* LOGCXX-322: TODO addedtschoening2016-03-011-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1733082 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-456: I just made a really stupid mistake and forgot to add ↵tschoening2016-02-1817-89/+37
| | | | | | | | LOG4CXX_STATIC to my own created projects, so the tests were build with "import" in mind, but there was nothing exported by the lib because it got built statically. I'm reverting my changes. https://quality.embarcadero.com/browse/RSP-12598 git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1731100 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-461: OOS already flushes implicitly in reset.tschoening2015-12-221-1/+0
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1721432 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-461: SocketAppender creates memory issue for log4j server.tschoening2015-12-153-12/+25
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1720199 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-461: hopefully improved readabilitytschoening2015-12-154-348/+395
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1720193 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-322: I applied some changes to address the issue that multi threaded ↵tschoening2015-10-132-2/+12
| | | | | | apps using log4cxx may crash on application exit. To reduce the risk of unwanted side effects, they all depend on APR supporting threads, because in such a case it's most likely that the problems occur. Regarding the bug, there's no fast and easy solution, so I'll just leak resources, which shouldn't harm much on app exit. Without that, to many users need to patch the lib on their own... git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1708287 13f79535-47bb-0310-9956-ffa450edef68
* clang warned about logical AND with a constant operator and this line looks ↵tschoening2015-10-111-1/+1
| | | | | | more like it should use bit wise operator, like the lines before. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707953 13f79535-47bb-0310-9956-ffa450edef68
* I always wondered why I get the folder "xyz" in my root and found the reason ↵tschoening2015-10-111-9/+9
| | | | | | now, it's created during errorhandlertest, which uses fallback1.xml. I moved that folder to the output dir, which seems to be the better place and the test still passes. I guess it doesn't care where the folder is, but only relies on :x.log not being created and that's still the case. Anyways, we dont' want to write to arbitrary places in the file system, especially not into non-temp folders and if we write everything else to "output" already. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707951 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Changed test4, 5 and 6 to delete their generic files before ↵tschoening2015-10-091-11/+97
| | | | | | running and changed test7 to execute 4, 5 and 6 and simulate a non empty directory that way. The tests now pass using empty and non empty directory and I hope to not have totally wasted my time... :-( But all other tests worked with and without empty directory, so I hope my changes have some purpose. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707820 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Pushing back the wait factor because it's not as important as I ↵tschoening2015-10-091-13/+13
| | | | | | thought, if needed at all. The tests still pass with an empty directory. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707807 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: test5 failed because it wrongly expected .gz extension.tschoening2015-10-091-1/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707806 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Reverting my behaviour changes for test 6, because I was on the ↵tschoening2015-10-091-4/+4
| | | | | | completely wrong track. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707805 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Improved debugging so the console logger output always starts on ↵tschoening2015-10-091-1/+1
| | | | | | a new line and is better readable. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707781 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Forgot to commit some additional macros last ime.tschoening2015-10-091-27/+29
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707780 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: I think I've found the reason for at least test6 failing: The ↵tschoening2015-10-091-48/+89
| | | | | | | | | | first rollover is done on the second log statement, which is only issued after some sleeping time. The problem is that the first file name to check for is created using "now" and that could have easily past if the second statement is issued. If the first file name is created with that in mind one second in the future, the first log statement can consume the current second of "now" including the sleep and still the first file will be found successfully, because it is expected to be created with the next second. I've added a waitFactor to logMsgAndSleep, because I think at least for test6 we need to make sure that each subsequent file is created in the next second, because that's how we build the file names: One second each. If we sleep in the logging only for half a second, two we may have two log statements within the same second resulting in another missing file. Interestingly, during my tests that didn't seem to make any difference... Additionally I've created test7 repeatly calling test6, because during my tests test6 succeeded even without my changes sometimes, but not always. With the changes it seems to succeed always. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707768 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Reworked the other tests to reduce redundant code. The same ↵tschoening2015-10-091-206/+162
| | | | | | tests like before fail now, so I shouldn't have broken too much. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707759 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Chaning indentation, because it wasn't really uniform before, ↵tschoening2015-10-091-289/+293
| | | | | | isn't compared to other source files and my IDE is configured to use tabs and I brought them in in the last commit already. But that shouldn't be a problem because it's easier to change form tabs to spaces again than the other way around, if some parts use indentation 8, other 2 and most of the other files something around 4 and 8. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707718 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: I've reworked test1 to reduce the amount of duplicated code: All ↵tschoening2015-10-091-74/+191
| | | | | | tests seem to produce file names in the same way, log some messages and most of them check results using file content comparison. The test still succeeds and I formerly got an access violation in fprintf if not all tests succeeded, which is gone now as well. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707714 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Changed delayUntilNextSecond to delayUntilNextSecondWithMsg, ↵tschoening2015-10-091-3/+3
| | | | | | because I don't see why only some should print the msg, while others don't. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707679 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: Removed hard coded 100 milliseconds to get a more uniform behaviour.tschoening2015-10-091-12/+12
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707678 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-457: delayUntilNextMinute is not used (anymore) and even if it will ↵tschoening2015-10-091-9/+0
| | | | | | be needed in the future it shoudl be implemented calling delayUnitNextSecond. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707676 13f79535-47bb-0310-9956-ffa450edef68
* Added delayUntilNextSecondWithMsg because the same msg was hard coded 3 times.tschoening2015-10-081-9/+10
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707615 13f79535-47bb-0310-9956-ffa450edef68
* Applied the same change like for date.cpp about apr.h defining INT64_C, only ↵tschoening2015-10-081-6/+2
| | | | | | that in this file everythign should have been OK before already because apr_strings.h was included before INT64C was checked. But now the code should look more clearly. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707614 13f79535-47bb-0310-9956-ffa450edef68
* APR may include stdint.h already, which itself may provide INT64_C. So I ↵tschoening2015-10-081-4/+1
| | | | | | swapped both lines to remove a warnign about a wrong redefinition of INT64_C in that case. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707594 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-456: My C++-Builder 10 Seattle seems to wrongly remove getters and ↵tschoening2015-10-0817-37/+89
| | | | | | setters which are not used in the static lib itself, but e.g. only in tests, if they are defined in the header. I'm not sure how to handle this properly, but moving the definition to the CPP implementation file of a class seems to resolve this issue, so I'm doing that for now, because it shouldn't harm anyone. If it does, we can easily revert and document the problem further in the ticket. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707585 13f79535-47bb-0310-9956-ffa450edef68
* Foxed some lines with broken code style, comments where not properly ↵tschoening2015-10-081-9/+8
| | | | | | associated to their methods and such. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707579 13f79535-47bb-0310-9956-ffa450edef68
* We rely on apr anyways, so include it earlier, because APR will then ↵tschoening2015-10-081-2/+3
| | | | | | detect/define WIN32 on it's own an WIn32 users don'T need to manually anymore. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707552 13f79535-47bb-0310-9956-ffa450edef68
* Removed warning about unreachable code by making it reachable. Normally I do ↵tschoening2015-10-081-3/+5
| | | | | | prefer early returns as wlel, but I guess the return NULL in this case has been left by purpose because some compilers7IDEs warn about the method should return some value else. Ignoring the warnign with #pragma seemed to much overhead, because that would need to be compiler specific. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707550 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-426: Let APR deal with windows.h and WIN32_LEAN_AND_MEAN and that ↵tschoening2015-10-081-8/+6
| | | | | | stuff. If we define WINSOCK here, aAPR will itself include windows.h anyways, but without winsock, which results in socklen_t being not available, but used in apr.h. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1707547 13f79535-47bb-0310-9956-ffa450edef68
* Corrected indentationtschoening2015-05-171-37/+58
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1679869 13f79535-47bb-0310-9956-ffa450edef68
* LOG4CXX_CONFIGURATION has been mentioned two times which resulted in bad ↵tschoening2015-02-191-5/+4
| | | | | | expression in my mind. I thought that maybe some other env var was meant, but code only uses those two. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1660906 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-442:Using non standard ports with syslog.Credits to 'Ulrik'.chand2015-01-294-8/+22
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1655676 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-434 : The attached patch fixes some cppcheck warnings. Credits to ↵chand2015-01-075-7/+6
| | | | | | 'Christian Ehrlicher git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1649999 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-428: Wildcards in Makefile.am break either VPATH or non-VPATH installsrhys2014-12-1228-221/+534
| | | | | | | | | | | | | | | | | Also, LOGCXX-446: make install fails, trying to overwrite header files Many Makefile.am files spread throughout the source tree used wildcards. This is invalid, as discussed at https://www.gnu.org/software/automake/manual/html_node/Wildcards.html. The use of wildcards was problematic for VPATH builds, in particular there was one Makefile.am which included srcdir and builddir headers twice. Correcting it to install only the builddir log4cxx.h and the remaining headers from srcdir/*.h addresses LOGCXX-446. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1644820 13f79535-47bb-0310-9956-ffa450edef68
* LOG4CXX-443: Return by const reference in Logger::getName()rhys2014-10-212-1/+2
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1633356 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-421: We have a user who wants to build log4cxx under Windows XP, but ↵tschoening2014-09-073-101/+1
| | | | | | the current work to get C++ Builder XE 4 supported introduced a dependency on where.exe, which is not available under Windows XP out of the box. So I had two options: Either remove the added logic or make the use of where.exe optional. I decided the first, because the already implemented logic already seems hard to maintain and simply doesn't work, it looks like some changes to cpptasks is necessary anyways. If that's the case, I could fix all problems directly in cpptasks, which surely is easier than using ANT. And as always, no work done is lost, we can always revert... git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1623040 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-430: My patch didn't work, against my expactations I seem to have ↵tschoening2014-09-073-13/+14
| | | | | | introduced "static initialization order fiasco", which the former implementation didn't suffer from. So I changed the problem of not beeing thread safe with another one and I was surely only lucky in my former runs of the testsuite to succeed. Currently they don't, so I'll revert my changes and reopen the issue again. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1623014 13f79535-47bb-0310-9956-ffa450edef68
* Corrected some docs issues: "optimization" seems to be "optimize" and ↵tschoening2014-07-021-1/+1
| | | | | | enable-shared|static not used at all. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1607481 13f79535-47bb-0310-9956-ffa450edef68
* I changed the description of when to create bugs and when to use the users ↵tschoening2014-07-022-1/+49
| | | | | | mailing list instead first, because no one objected on the dev mailing list. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1607423 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-430: LogManager::getRootLogger is not thread-safetschoening2014-03-143-14/+13
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1577488 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-411: Fixed this in LOGCXX-394.tschoening2014-03-101-0/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1575851 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-416joseph2014-03-031-0/+1
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1573645 13f79535-47bb-0310-9956-ffa450edef68
* LOG4CXX-428: install log4cxx.h in VPATHrhys2014-03-011-1/+2
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1573202 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-394: Doxygen: warning: no uniquely matching class member found fortschoening2014-02-211-3/+4
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1570494 13f79535-47bb-0310-9956-ffa450edef68
* fixed indentationtschoening2014-02-211-7/+7
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1570493 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-425: exceptions in CachedDateFormatTestCase after LOGCXX-420tschoening2014-02-193-29/+12
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1569635 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-425: Changed the calculation of millis in a formatted string again ↵tschoening2014-02-182-54/+78
| | | | | | and added another tests for LOGCXX-420. There's some debug output in there which need to get removed later, but for now I need some help in defining the correct behavior and want others to be able to test themselves. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1569535 13f79535-47bb-0310-9956-ffa450edef68
* Place LevelPtr::operator== into namespace helperrhys2014-02-181-1/+6
| | | | | | | | | | | Fixes the following compilation error on GCC 4.8.1: src/main/include/log4cxx/level.h:279:67: error: specialization of ‘template<class T> bool log4cxx::helpers::ObjectPtrT<T>::operator==(const log4cxx::helpers::ObjectPtrT<T>&) const’ in different namespace [-fpermissive] template<> inline bool LevelPtr::operator==(const LevelPtr& rhs) const git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1569339 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-425: I added a test for LOGCXX-420 which failed in the former ↵tschoening2014-02-182-20/+46
| | | | | | | | codebase, additionally to the problem with the exceptions. The reason for the exceptions were negativ indices, the calculations for offsets in the magicString introduced in LOGCXX-420 were wrong because the magicString is always 3 chars only. This and the failing added test for LOGCXX-420 has been fixed, but test 17 of cacheddateformattestcase.cpp is failing now. I will have a look into this afterwards. git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1569322 13f79535-47bb-0310-9956-ffa450edef68
* LOGCXX-367: Use C++ headers instead of C-ones.tschoening2014-02-182-2/+4
| | | | git-svn-id: http://svn.apache.org/repos/asf/incubator/log4cxx/trunk@1569234 13f79535-47bb-0310-9956-ffa450edef68
View Lorry Controller Status