kdtree - Project Hosting on Google Code

Last 30 days Dec 20, 2009 issue 7 (free_nodes have a memory leak using node allocator) reported by rixiomorales - What steps will reproduce the problem? 1. Port the library to MSCV environment as static library (MT and MD). 2. Use windows threads to allocate and free the result nodes. 3. Debug displayed memory leaks with no details. What is the expected output? What do you see instead? -Expected: No memory leaks using allocator. -Just one of many (Memory leak of 24bytes at 0x057f8ae0) What version of the product are you using? On what operating system? kd-tree lib 0.55/MS Windows XP Please provide any additional information below. I tracked the leak and found that the free_nodes global static variable is not freed after the use. So I have solve the problem (added a clear free_node when freeing the kdtree if there are free_nodes), I will send the code to the project owner the modifications and the ported code If he/she wants. I've also test the code in MinGW. Further unitary testing is required. Thank you for the lib is very useful. I'm new to this Google code stuff, so If I've made any mistake I beg your pardon. Just needed to contact with the owner and collaborate. Regards. What steps will reproduce the problem? 1. Port the library to MSCV environment as static library (MT and MD). 2. Use windows threads to allocate and free the result nodes. 3. Debug displayed memory leaks with no details. What is the expected output? What do you see instead? -Expected: No memory leaks using allocator. -Just one of many (Memory leak of 24bytes at 0x057f8ae0) What version of the product are you using? On what operating system? kd-tree lib 0.55/MS Windows XP Please provide any additional information below. I tracked the leak and found that the free_nodes global static variable is not freed after the use. So I have solve the problem (added a clear free_node when freeing the kdtree if there are free_nodes), I will send the code to the project owner the modifications and the ported code If he/she wants. I've also test the code in MinGW. Further unitary testing is required. Thank you for the lib is very useful. I'm new to this Google code stuff, so If I've made any mistake I beg your pardon. Just needed to contact with the owner and collaborate. Regards. Earlier this year Nov 16, 2009 issue 3 (kd_insertf causes segfault with less then 16 dimensions) commented on by haomen2 - there is a similar problem for kd_nearest_range() there is a similar problem for kd_nearest_range() Sep 21, 2009 issue 5 (bug in realse mode) commented on by joseluisblancoc - Hi billconan, I've tested this issue and the point is that... "assert" is replaced by nothing in Release, so no points are added to the kd-tree. So I think this bug can be closed. Regards. Hi billconan, I've tested this issue and the point is that... "assert" is replaced by nothing in Release, so no points are added to the kd-tree. So I think this bug can be closed. Regards. Sep 14, 2009 issue 6 (same node problem) reported by haomen2 - What steps will reproduce the problem? 1. insert 10000 same node into kd-tree, d=3, the kd tree can not be built 2. 3. What is the expected output? What do you see instead? should be able to finish building the tree, <= nodes should be on the left side, but is inserting large number of same nodes, the program is not able to finish.(maybe depth problem?) What version of the product are you using? On what operating system? 5.5, under visual studio 2008 on win xp. Please provide any additional information below. What steps will reproduce the problem? 1. insert 10000 same node into kd-tree, d=3, the kd tree can not be built 2. 3. What is the expected output? What do you see instead? should be able to finish building the tree, <= nodes should be on the left side, but is inserting large number of same nodes, the program is not able to finish.(maybe depth problem?) What version of the product are you using? On what operating system? 5.5, under visual studio 2008 on win xp. Please provide any additional information below. Sep 04, 2009 r28 (implementing N nearest neighbors search ) committed by jtsiomb - implementing N nearest neighbors search implementing N nearest neighbors search Aug 27, 2009 kdtree-0.5.5.tar.gz (kdtree 0.5.5) file uploaded by jtsiomb Aug 27, 2009 r27 (tagging 0.5.5 release ) committed by jtsiomb - tagging 0.5.5 release tagging 0.5.5 release Aug 27, 2009 r26 (removing empty kdtree dir ) committed by jtsiomb - removing empty kdtree dir removing empty kdtree dir Aug 27, 2009 r25 (moving arround files ) committed by jtsiomb - moving arround files moving arround files Aug 27, 2009 r24 (moving current code to trunk ) committed by jtsiomb - moving current code to trunk moving current code to trunk Aug 10, 2009 issue 5 (bug in realse mode) reported by billconan - the following test code acts differently in release and debug mode under visual studio 2008. int main(int argc, char *argv) { int i, vcount = 10; kdtree kd; kdres set; unsigned int msec=0, start=0; if(argc > 1 && isdigit(argv[1][0])) { vcount = atoi(argv[1]); } printf("inserting %d random vectors... ", vcount); fflush(stdout); kd = kd_create(3); assert(kd_insert3(kd, 42.1f,44.1f,44.1f, 0)==0); assert(kd_insert3(kd, 43.2f,44.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,43.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,45.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,42.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,47.1f,44.1f, 0)==0); set = kd_nearest_range3(kd, 44.0f, 44.0f, 44.0f, 10.0f); printf("range query returned %d items in %.5f sec\n", kd_res_size(set), (float)msec / 1000.0); kd_res_free(set); kd_free(kd); getchar(); return 0; } in the debug mode, there will be 6 nearby neighbors returned by the query, which is correct. but in the release mode, 0 neighbor will return. the following test code acts differently in release and debug mode under visual studio 2008. int main(int argc, char argv) { int i, vcount = 10; kdtree kd; kdres set; unsigned int msec=0, start=0; if(argc > 1 && isdigit(argv[1][0])) { vcount = atoi(argv[1]); } printf("inserting %d random vectors... ", vcount); fflush(stdout); kd = kd_create(3); assert(kd_insert3(kd, 42.1f,44.1f,44.1f, 0)==0); assert(kd_insert3(kd, 43.2f,44.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,43.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,45.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,42.1f,44.1f, 0)==0); assert(kd_insert3(kd, 44.1f,47.1f,44.1f, 0)==0); set = kd_nearest_range3(kd, 44.0f, 44.0f, 44.0f, 10.0f); printf("range query returned %d items in %.5f sec\n", kd_res_size(set), (float)msec / 1000.0); kd_res_free(set); kd_free(kd); getchar(); return 0; } in the debug mode, there will be 6 nearby neighbors returned by the query, which is correct. but in the release mode, 0 neighbor will return. Jun 18, 2009 Welcome Wiki page edited by jtsiomb Jun 08, 2009 issue 3 (kd_insertf causes segfault with less then 16 dimensions) commented on by Wizarth - That's correct, I don't know where I got 231 from. That's correct, I don't know where I got 231 from. Jun 05, 2009 issue 3 (kd_insertf causes segfault with less then 16 dimensions) commented on by 1ec5.org - This fix worked for me on the same version, but the relevant line was 192 in kdtree.c. This fix worked for me on the same version, but the relevant line was 192 in kdtree.c. May 29, 2009 issue 4 (kd_nearest_rangef causes segfault when dimnsions <= 16) reported by Wizarth - In kd_nearest_rangef, buf is not set to sbuf when dim < 16 Line 301: bptr = buf = sbuf; In kd_nearest_rangef, buf is not set to sbuf when dim < 16 Line 301: bptr = buf = sbuf; May 28, 2009 issue 3 (kd_insertf causes segfault with less then 16 dimensions) reported by Wizarth - What steps will reproduce the problem? 1. Use of kd_insertf with less then 16 dimensions What version of the product are you using? On what operating system? 0.5.4 Please provide any additional information below. buf is not set in the path where dim <= 16 Fix: Line 231: bptr = buf = sbuf; What steps will reproduce the problem? 1. Use of kd_insertf with less then 16 dimensions What version of the product are you using? On what operating system? 0.5.4 Please provide any additional information below. buf is not set in the path where dim <= 16 Fix: Line 231: bptr = buf = sbuf; Mar 16, 2009 r22 (Apparently I had the rlist_insert ordering comparison backwa...) committed by jtsiomb - Apparently I had the rlist_insert ordering comparison backwards... Apparently I had the rlist_insert ordering comparison backwards... Mar 16, 2009 r21 (- applied Tamas Nepusz's patch for single nearest neighbor s...) committed by jtsiomb - - applied Tamas Nepusz's patch for single nearest neighbor search - applied Tamas Nepusz's patch for single nearest neighbor search Jan 25, 2009 kdtree-0.5.4.tar.gz (kdtree 0.5.4) file uploaded by jtsiomb Jan 16, 2009 r20 (fixed a serious bug in kdres_end_something ) committed by jtsiomb - fixed a serious bug in kdres_end_something fixed a serious bug in kdres_end_something Older Oct 26, 2008 r19 (fixed a %s/../.. that apparently went wrong and resulted in ...) committed by jtsiomb - fixed a %s/../.. that apparently went wrong and resulted in the following copyright statement: "This file is part of ``struct kdtree '', a library for working with kd-trees." :) fixed a %s/../.. that apparently went wrong and resulted in the following copyright statement: "This file is part of ``struct kdtree *'', a library for working with kd-trees." :) Aug 27, 2008 r18 (applied Justin Carlson's patch (slightly modified) which cha...) committed by jtsiomb - applied Justin Carlson's patch (slightly modified) which changes all the void pointers into opaque struct pointers, to ease readability. applied Justin Carlson's patch (slightly modified) which changes all the void pointers into opaque struct pointers, to ease readability.

issue 7 (free_nodes have a memory leak using node allocator) reported by rixiomorales - What steps will reproduce the problem? 1. Port the library to MSCV environment as static library (MT and MD). 2. Use windows threads to allocate and free the result nodes. 3. Debug displayed memory leaks with no details. What is the expected output? What do you see instead? -Expected: No memory leaks using allocator. -Just one of many (Memory leak of 24bytes at 0x057f8ae0) What version of the product are you using? On what operating system? kd-tree lib 0.55/MS Windows XP Please provide any additional information below. I tracked the leak and found that the free_nodes global static variable is not freed after the use. So I have solve the problem (added a clear free_node when freeing the kdtree if there are free_nodes), I will send the code to the project owner the modifications and the ported code If he/she wants. I've also test the code in MinGW. Further unitary testing is required. Thank you for the lib is very useful. I'm new to this Google code stuff, so If I've made any mistake I beg your pardon. Just needed to contact with the owner and collaborate. Regards.

What steps will reproduce the problem? 1. Port the library to MSCV environment as static library (MT and MD). 2. Use windows threads to allocate and free the result nodes. 3. Debug displayed memory leaks with no details. What is the expected output? What do you see instead? -Expected: No memory leaks using allocator. -Just one of many (Memory leak of 24bytes at 0x057f8ae0) What version of the product are you using? On what operating system? kd-tree lib 0.55/MS Windows XP Please provide any additional information below. I tracked the leak and found that the free_nodes global static variable is not freed after the use. So I have solve the problem (added a clear free_node when freeing the kdtree if there are free_nodes), I will send the code to the project owner the modifications and the ported code If he/she wants. I've also test the code in MinGW. Further unitary testing is required. Thank you for the lib is very useful. I'm new to this Google code stuff, so If I've made any mistake I beg your pardon. Just needed to contact with the owner and collaborate. Regards.

Last 30 days

Earlier this year

Older