Re: Regarding loops and memory leak
"Ken A. Redergård" <[email protected]>
| Newsgroups | gmane.comp.mobile.kannel.devel |
|---|---|
| Message-ID | <[email protected]> |
Hi, Thanks both for testing the example :-) Out of old habit I've been running configure like this: ./configure --enable-warnings --with-defaults=debug --with-malloc=checked I've tried without these arguments now, and the leak disappeared in the test_octstr example. But the sqlbox code that runs the same kind of loop still leak memory without the debug params. I guess I have some debugging left :-( -Ken. Benjamin Lee wrote: >For what it's worth, > >I tested (simply just by watching top output) the code below... and there >doesn't appear to be a memory leak. I used the latest CVS code and rebuilt >the lib .a files too. > > >[ben@xaos:~/cvs/remote/gateway]$ gcc -D_REENTRANT=1 -I. -Igw -D_XOPEN_SOURCE=600 -D_BSD_SOURCE -DBROKEN_PTHREADS=1 -I/usr/include/libxml2 -I/usr/include/openssl -Wall -c test_asdfsd.c >test_asdfsd.c: In function `main': >test_asdfsd.c:8: warning: unused variable `i' >[Mon, 13 Sep 2004 14:24:39 +1000] Home Sweet Home >[ben@xaos:~/cvs/remote/gateway]$ gcc -D_REENTRANT=1 -I. -Igw -D_XOPEN_SOURCE=600 -D_BSD_SOURCE -DBROKEN_PTHREADS=1 -I/usr/include/libxml2 -I/usr/include/openssl -Wall -lssl -lpthread -lxml2 -o teststs test_asdfsd.o libgw.a libgwlib.a >[Mon, 13 Sep 2004 14:24:43 +1000] Home Sweet Home >[ben@xaos:~/cvs/remote/gateway]$ >[Mon, 13 Sep 2004 14:24:44 +1000] Home Sweet Home >[ben@xaos:~/cvs/remote/gateway]$ ./teststs > > > > >On Monday, 2004-09-13 at 01:53:32 AM, "Ken A. Redergård" scribbled: > > >>Hi, >> >>Read your posting in the [email protected] list and have a simliar >>problem myself. >> >>I'm also running SuSE 9.1 and experience a memoryleak which I described >>in this posting: >>http://www.kannel.org/mailman/private/devel/2004-August/020299.html >> >>As Stipe Tolj said it does not show up in valgrind, but the leak is >>present and increasing as long as the loop is beeing run. >> >>Could you try to run this piece of code and tell me if you also get a >>memoryleak ? >>It could be that the while(run) semantics in your example are equal to >>the one in the leaking example below. >> >>/* test_octstr.c */ >> >>#include <unistd.h> >>#include "gwlib/gwlib.h" >> >>int main(void) { >> Octstr *os; >> int i; >> >> gwlib_init(); >> >>while(1) { >> os = octstr_format("sadfasdfasdfas"); >> octstr_append(os, octstr_imm("asdfasdfasdfasdf")); >> octstr_destroy(os); >> gwthread_sleep(1); >> } >> >> /* we never get here */ >> return 0; >>} >> >>With regards, >>Ken A. Redergård >> >> >> >> > > >