I just got a bug report on python-numarray (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=240289) indicating that the size of the installed package has increased from around 6.7MB in FC6 (1.5.2-1.fc6) to around 56MB currently (1.5.2-2.fc7). Now, the only this that was done was to rebuild the package for python 2.5 back in December.
If I rebuild the package now the size is back to "normal". The differences are in the size of the shared libraries:
< 1272 /usr/lib/python2.5/site-packages/numarray/_bytes.so < 1264 /usr/lib/python2.5/site-packages/numarray/_chararray.so < 1308 /usr/lib/python2.5/site-packages/numarray/_conv.so < 1268 /usr/lib/python2.5/site-packages/numarray/_converter.so < 1280 /usr/lib/python2.5/site-packages/numarray/_ndarray.so < 1280 /usr/lib/python2.5/site-packages/numarray/_numarray.so < 1260 /usr/lib/python2.5/site-packages/numarray/_numerictype.so < 1264 /usr/lib/python2.5/site-packages/numarray/_objectarray.so < 1260 /usr/lib/python2.5/site-packages/numarray/_operator.so < 1376 /usr/lib/python2.5/site-packages/numarray/_sort.so < 1292 /usr/lib/python2.5/site-packages/numarray/_ufunc.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncBool.so < 1312 /usr/lib/python2.5/site-packages/numarray/_ufuncComplex32.so < 1312 /usr/lib/python2.5/site-packages/numarray/_ufuncComplex64.so < 1308 /usr/lib/python2.5/site-packages/numarray/_ufuncFloat32.so < 1308 /usr/lib/python2.5/site-packages/numarray/_ufuncFloat64.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncInt16.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncInt32.so < 1324 /usr/lib/python2.5/site-packages/numarray/_ufuncInt64.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncInt8.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt16.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt32.so < 1324 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt64.so < 1320 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt8.so ---
28 /usr/lib/python2.5/site-packages/numarray/_bytes.so 20 /usr/lib/python2.5/site-packages/numarray/_chararray.so 72 /usr/lib/python2.5/site-packages/numarray/_conv.so 20 /usr/lib/python2.5/site-packages/numarray/_converter.so 36 /usr/lib/python2.5/site-packages/numarray/_ndarray.so 36 /usr/lib/python2.5/site-packages/numarray/_numarray.so 12 /usr/lib/python2.5/site-packages/numarray/_numerictype.so 16 /usr/lib/python2.5/site-packages/numarray/_objectarray.so 16 /usr/lib/python2.5/site-packages/numarray/_operator.so 136 /usr/lib/python2.5/site-packages/numarray/_sort.so 48 /usr/lib/python2.5/site-packages/numarray/_ufunc.so 80 /usr/lib/python2.5/site-packages/numarray/_ufuncBool.so 72 /usr/lib/python2.5/site-packages/numarray/_ufuncComplex32.so 72 /usr/lib/python2.5/site-packages/numarray/_ufuncComplex64.so 72 /usr/lib/python2.5/site-packages/numarray/_ufuncFloat32.so 72 /usr/lib/python2.5/site-packages/numarray/_ufuncFloat64.so 76 /usr/lib/python2.5/site-packages/numarray/_ufuncInt16.so 76 /usr/lib/python2.5/site-packages/numarray/_ufuncInt32.so 80 /usr/lib/python2.5/site-packages/numarray/_ufuncInt64.so 76 /usr/lib/python2.5/site-packages/numarray/_ufuncInt8.so 76 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt16.so 76 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt32.so 84 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt64.so 76 /usr/lib/python2.5/site-packages/numarray/_ufuncUInt8.so
and so on.
I did a quick compare of FE 6 to the last FE devel and found that the following packages have increased in size by more than a factor of 2. Some of these may be fine, I haven't looked into it in detail.
What's up?
banshee 2536364 -> 6770157 cppunit-doc 4504311 -> 10063237 duplicity 936214 -> 2255520 fwbackups 347859 -> 894038 graphviz-tcl 298038 -> 2020760 libburn-devel 47334 -> 944521 libcaca-devel 507545 -> 1482955 libhugetlbfs 97382 -> 206633 libhugetlbfs-test 1031501 -> 135632714 libpreludedb-python 91466 -> 1355682 linux-libertine-fonts 2589874 -> 10225720 mercurial 2063651 -> 6487353 nsd 446095 -> 1665847 ogre 6374746 -> 29476134 openbabel-perl 1206495 -> 3376667 openbabel-python 1501764 -> 3509293 ortp-devel 197141 -> 3761868 php-pear-PHPUnit 126756 -> 1106041 plplot-wxGTK 54287 -> 144831 pybluez 341942 -> 1617662 pyflowtools 35170 -> 1302402 pyfribidi 27095 -> 1299303 pygpgme 79943 -> 1349853 python-CDDB 55383 -> 1324169 python-GeoIP 30419 -> 1299395 python-adns 60844 -> 1327860 python-alsaaudio 28238 -> 1297182 python-chm 121632 -> 2670312 python-durus 419912 -> 1718919 python-numarray 6771218 -> 56416870 python-protocols 766774 -> 2075564 python-pycurl 212612 -> 1485863 python-quixote 1853963 -> 7008959 python-ruledispatch 933214 -> 2272446 python-smbpasswd 34293 -> 1302773 python-twisted-runner 74839 -> 1349825 python-zope-interface 1013800 -> 2358938 pyxmms 327987 -> 2871939 rdiff-backup 1286381 -> 3896666 supertux 15057058 -> 49666684 wallpapoz 110963 -> 363890 xpa 945338 -> 1979801 zvbi-fonts 43808 -> 132817
Orion Poplawski (orion@cora.nwra.com) said:
I just got a bug report on python-numarray (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=240289) indicating that the size of the installed package has increased from around 6.7MB in FC6 (1.5.2-1.fc6) to around 56MB currently (1.5.2-2.fc7). Now, the only this that was done was to rebuild the package for python 2.5 back in December.
If I rebuild the package now the size is back to "normal". The differences are in the size of the shared libraries:
Do you turn off debug packages?
Bill
Bill Nottingham wrote:
Orion Poplawski (orion@cora.nwra.com) said:
I just got a bug report on python-numarray (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=240289) indicating that the size of the installed package has increased from around 6.7MB in FC6 (1.5.2-1.fc6) to around 56MB currently (1.5.2-2.fc7). Now, the only this that was done was to rebuild the package for python 2.5 back in December.
If I rebuild the package now the size is back to "normal". The differences are in the size of the shared libraries:
Do you turn off debug packages?
Bill
I do nothing but rebuild. Perhaps debug packages were broken for a bit?
On 5/17/07, Orion Poplawski orion@cora.nwra.com wrote:
I do nothing but rebuild. Perhaps debug packages were broken for a bit?
Is there a way to systematically check for this possibility via postmortem analysis of the current package tree? Like can we do a check for unstripped debug symbols on a rawhide installed system and map back to build timestamps in the associated rpm?
-jef
Orion Poplawski wrote:
I do nothing but rebuild. Perhaps debug packages were broken for a bit?
Although the .so files in -2.fc7 appear to be stripped:
/usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), stripped
# nm /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so nm: /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so: no symbols
# ls -l /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so -rwxr-xr-x 1 root root 1308096 2006-12-12 22:22 /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so
It appeara that the "big" one doesn't have a "NEEDED libpython2.5.so.1.0" entry, perhaps it was getting included directly? This would have to have been an effect of a change in some other package though.
Big:
# objdump -x /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so
/usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so: file format elf32-i386 /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so architecture: i386, flags 0x00000150: HAS_SYMS, DYNAMIC, D_PAGED start address 0x0001c810
Program Header: LOAD off 0x00000000 vaddr 0x00000000 paddr 0x00000000 align 2**12 filesz 0x00119e0c memsz 0x00119e0c flags r-x LOAD off 0x0011a000 vaddr 0x0011a000 paddr 0x0011a000 align 2**12 filesz 0x000250c0 memsz 0x0002b90c flags rw- DYNAMIC off 0x0011a0ec vaddr 0x0011a0ec paddr 0x0011a0ec align 2**2 filesz 0x000000d0 memsz 0x000000d0 flags rw- EH_FRAME off 0x001001c8 vaddr 0x001001c8 paddr 0x001001c8 align 2**2 filesz 0x000052cc memsz 0x000052cc flags r-- STACK off 0x00000000 vaddr 0x00000000 paddr 0x00000000 align 2**2 filesz 0x00000000 memsz 0x00000000 flags rw-
Dynamic Section: NEEDED libm.so.6 NEEDED libpthread.so.0 NEEDED libc.so.6 INIT 0x18d78 FINI 0xdf334 GNU_HASH 0xd4 STRTAB 0x75f8 SYMTAB 0x25a8 STRSZ 0x5203 SYMENT 0x10 PLTGOT 0x11a528 PLTRELSZ 0x1d38 PLTREL 0x11 JMPREL 0x17040 REL 0xd2e8 RELSZ 0x9d58 RELENT 0x8 VERNEED 0xd208 VERNEEDNUM 0x3 VERSYM 0xc7fc RELCOUNT 0x11fb
Version References: required from libpthread.so.0: 0x0d696912 0x00 09 GLIBC_2.2 0x0d696911 0x00 08 GLIBC_2.1 0x0d696910 0x00 04 GLIBC_2.0 required from libm.so.6: 0x0d696910 0x00 03 GLIBC_2.0 required from libc.so.6: 0x09691f73 0x00 12 GLIBC_2.1.3 0x0d696914 0x00 11 GLIBC_2.4 0x0d696913 0x00 10 GLIBC_2.3 0x09691974 0x00 07 GLIBC_2.3.4 0x0d696912 0x00 06 GLIBC_2.2 0x0d696911 0x00 05 GLIBC_2.1 0x0d696910 0x00 02 GLIBC_2.0
Sections: Idx Name Size VMA LMA File off Algn 0 .gnu.hash 000024d4 000000d4 000000d4 000000d4 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 1 .dynsym 00005050 000025a8 000025a8 000025a8 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 2 .dynstr 00005203 000075f8 000075f8 000075f8 2**0 CONTENTS, ALLOC, LOAD, READONLY, DATA 3 .gnu.version 00000a0a 0000c7fc 0000c7fc 0000c7fc 2**1 CONTENTS, ALLOC, LOAD, READONLY, DATA 4 .gnu.version_r 000000e0 0000d208 0000d208 0000d208 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 5 .rel.dyn 00009d58 0000d2e8 0000d2e8 0000d2e8 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 6 .rel.plt 00001d38 00017040 00017040 00017040 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 7 .init 00000017 00018d78 00018d78 00018d78 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 8 .plt 00003a80 00018d90 00018d90 00018d90 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 9 .text 000c2b24 0001c810 0001c810 0001c810 2**4 CONTENTS, ALLOC, LOAD, READONLY, CODE 10 .fini 0000001c 000df334 000df334 000df334 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 11 .rodata 00020e68 000df360 000df360 000df360 2**5 CONTENTS, ALLOC, LOAD, READONLY, DATA 12 .eh_frame_hdr 000052cc 001001c8 001001c8 001001c8 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 13 .eh_frame 00014978 00105494 00105494 00105494 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 14 .ctors 00000008 0011a000 0011a000 0011a000 2**2 CONTENTS, ALLOC, LOAD, DATA 15 .dtors 00000008 0011a008 0011a008 0011a008 2**2 CONTENTS, ALLOC, LOAD, DATA 16 .jcr 00000004 0011a010 0011a010 0011a010 2**2 CONTENTS, ALLOC, LOAD, DATA 17 .data.rel.ro 000000cc 0011a020 0011a020 0011a020 2**5 CONTENTS, ALLOC, LOAD, DATA 18 .dynamic 000000d0 0011a0ec 0011a0ec 0011a0ec 2**2 CONTENTS, ALLOC, LOAD, DATA 19 .got 0000036c 0011a1bc 0011a1bc 0011a1bc 2**2 CONTENTS, ALLOC, LOAD, DATA 20 .got.plt 00000ea8 0011a528 0011a528 0011a528 2**2 CONTENTS, ALLOC, LOAD, DATA 21 .data 00023ce0 0011b3e0 0011b3e0 0011b3e0 2**5 CONTENTS, ALLOC, LOAD, DATA 22 .bss 0000684c 0013f0c0 0013f0c0 0013f0c0 2**5 ALLOC 23 .gnu_debuglink 00000018 00000000 00000000 0013f0c0 2**2 CONTENTS, READONLY SYMBOL TABLE: no symbols
Small: /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so: file format elf32-i386 /usr/lib/python2.5/site-packages/numarray/random_array/ranlib2.so architecture: i386, flags 0x00000150: HAS_SYMS, DYNAMIC, D_PAGED start address 0x00001480
Program Header: LOAD off 0x00000000 vaddr 0x00000000 paddr 0x00000000 align 2**12 filesz 0x0000843c memsz 0x0000843c flags r-x LOAD off 0x00009000 vaddr 0x00009000 paddr 0x00009000 align 2**12 filesz 0x000005f0 memsz 0x00000ea4 flags rw- DYNAMIC off 0x00009018 vaddr 0x00009018 paddr 0x00009018 align 2**2 filesz 0x000000d8 memsz 0x000000d8 flags rw- EH_FRAME off 0x00007bac vaddr 0x00007bac paddr 0x00007bac align 2**2 filesz 0x000001cc memsz 0x000001cc flags r-- STACK off 0x00000000 vaddr 0x00000000 paddr 0x00000000 align 2**2 filesz 0x00000000 memsz 0x00000000 flags rw-
Dynamic Section: NEEDED libpython2.5.so.1.0 NEEDED libm.so.6 NEEDED libpthread.so.0 NEEDED libc.so.6 INIT 0x1118 FINI 0x7104 GNU_HASH 0xd4 STRTAB 0x878 SYMTAB 0x2a8 STRSZ 0x383 SYMENT 0x10 PLTGOT 0x917c PLTRELSZ 0x1a0 PLTREL 0x11 JMPREL 0xf78 REL 0xd18 RELSZ 0x260 RELENT 0x8 VERNEED 0xcb8 VERNEEDNUM 0x2 VERSYM 0xbfc RELCOUNT 0x29
Version References: required from libc.so.6: 0x09691f73 0x00 05 GLIBC_2.1.3 0x0d696910 0x00 04 GLIBC_2.0 0x09691974 0x00 03 GLIBC_2.3.4 required from libm.so.6: 0x0d696910 0x00 02 GLIBC_2.0
Sections: Idx Name Size VMA LMA File off Algn 0 .gnu.hash 000001d4 000000d4 000000d4 000000d4 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 1 .dynsym 000005d0 000002a8 000002a8 000002a8 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 2 .dynstr 00000383 00000878 00000878 00000878 2**0 CONTENTS, ALLOC, LOAD, READONLY, DATA 3 .gnu.version 000000ba 00000bfc 00000bfc 00000bfc 2**1 CONTENTS, ALLOC, LOAD, READONLY, DATA 4 .gnu.version_r 00000060 00000cb8 00000cb8 00000cb8 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 5 .rel.dyn 00000260 00000d18 00000d18 00000d18 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 6 .rel.plt 000001a0 00000f78 00000f78 00000f78 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 7 .init 00000017 00001118 00001118 00001118 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 8 .plt 00000350 00001130 00001130 00001130 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 9 .text 00005c84 00001480 00001480 00001480 2**4 CONTENTS, ALLOC, LOAD, READONLY, CODE 10 .fini 0000001c 00007104 00007104 00007104 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 11 .rodata 00000a8a 00007120 00007120 00007120 2**5 CONTENTS, ALLOC, LOAD, READONLY, DATA 12 .eh_frame_hdr 000001cc 00007bac 00007bac 00007bac 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 13 .eh_frame 000006c4 00007d78 00007d78 00007d78 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 14 .ctors 00000008 00009000 00009000 00009000 2**2 CONTENTS, ALLOC, LOAD, DATA 15 .dtors 00000008 00009008 00009008 00009008 2**2 CONTENTS, ALLOC, LOAD, DATA 16 .jcr 00000004 00009010 00009010 00009010 2**2 CONTENTS, ALLOC, LOAD, DATA 17 .data.rel.ro 00000004 00009014 00009014 00009014 2**2 CONTENTS, ALLOC, LOAD, DATA 18 .dynamic 000000d8 00009018 00009018 00009018 2**2 CONTENTS, ALLOC, LOAD, DATA 19 .got 0000008c 000090f0 000090f0 000090f0 2**2 CONTENTS, ALLOC, LOAD, DATA 20 .got.plt 000000dc 0000917c 0000917c 0000917c 2**2 CONTENTS, ALLOC, LOAD, DATA 21 .data 00000390 00009260 00009260 00009260 2**5 CONTENTS, ALLOC, LOAD, DATA 22 .bss 000008a4 00009600 00009600 000095f0 2**5 ALLOC 23 .gnu_debuglink 00000018 00000000 00000000 000095f0 2**2 CONTENTS, READONLY SYMBOL TABLE: no symbols
Here's a diff of the objdumps. Looks like different glibc versions as well?
< = big, > = small
6c6 < start address 0x0001c810 ---
start address 0x00001480
10,16c10,16 < filesz 0x00119e0c memsz 0x00119e0c flags r-x < LOAD off 0x0011a000 vaddr 0x0011a000 paddr 0x0011a000 align 2**12 < filesz 0x000250c0 memsz 0x0002b90c flags rw- < DYNAMIC off 0x0011a0ec vaddr 0x0011a0ec paddr 0x0011a0ec align 2**2 < filesz 0x000000d0 memsz 0x000000d0 flags rw- < EH_FRAME off 0x001001c8 vaddr 0x001001c8 paddr 0x001001c8 align 2**2 < filesz 0x000052cc memsz 0x000052cc flags r-- ---
filesz 0x0000843c memsz 0x0000843c flags r-x LOAD off 0x00009000 vaddr 0x00009000 paddr 0x00009000 align 2**12 filesz 0x000005f0 memsz 0x00000ea4 flags rw-
DYNAMIC off 0x00009018 vaddr 0x00009018 paddr 0x00009018 align 2**2 filesz 0x000000d8 memsz 0x000000d8 flags rw- EH_FRAME off 0x00007bac vaddr 0x00007bac paddr 0x00007bac align 2**2 filesz 0x000001cc memsz 0x000001cc flags r--
20a21
NEEDED libpython2.5.so.1.0
24,25c25,26 < INIT 0x18d78 < FINI 0xdf334 ---
INIT 0x1118 FINI 0x7104
27,29c28,30 < STRTAB 0x75f8 < SYMTAB 0x25a8 < STRSZ 0x5203 ---
STRTAB 0x878 SYMTAB 0x2a8 STRSZ 0x383
31,32c32,33 < PLTGOT 0x11a528 < PLTRELSZ 0x1d38 ---
PLTGOT 0x917c PLTRELSZ 0x1a0
34,36c35,37 < JMPREL 0x17040 < REL 0xd2e8 < RELSZ 0x9d58 ---
JMPREL 0xf78 REL 0xd18 RELSZ 0x260
38,41c39,42 < VERNEED 0xd208 < VERNEEDNUM 0x3 < VERSYM 0xc7fc < RELCOUNT 0x11fb ---
VERNEED 0xcb8 VERNEEDNUM 0x2 VERSYM 0xbfc RELCOUNT 0x29
44,46c45,46 < required from libpthread.so.0: < 0x0d696912 0x00 09 GLIBC_2.2 < 0x0d696911 0x00 08 GLIBC_2.1 ---
required from libc.so.6: 0x09691f73 0x00 05 GLIBC_2.1.3
47a48
0x09691974 0x00 03 GLIBC_2.3.4
49,56d49 < 0x0d696910 0x00 03 GLIBC_2.0 < required from libc.so.6: < 0x09691f73 0x00 12 GLIBC_2.1.3 < 0x0d696914 0x00 11 GLIBC_2.4 < 0x0d696913 0x00 10 GLIBC_2.3 < 0x09691974 0x00 07 GLIBC_2.3.4 < 0x0d696912 0x00 06 GLIBC_2.2 < 0x0d696911 0x00 05 GLIBC_2.1 61c54 < 0 .gnu.hash 000024d4 000000d4 000000d4 000000d4 2**2 ---
0 .gnu.hash 000001d4 000000d4 000000d4 000000d4 2**2
63c56 < 1 .dynsym 00005050 000025a8 000025a8 000025a8 2**2 ---
1 .dynsym 000005d0 000002a8 000002a8 000002a8 2**2
65c58 < 2 .dynstr 00005203 000075f8 000075f8 000075f8 2**0 ---
2 .dynstr 00000383 00000878 00000878 00000878 2**0
67c60 < 3 .gnu.version 00000a0a 0000c7fc 0000c7fc 0000c7fc 2**1 ---
3 .gnu.version 000000ba 00000bfc 00000bfc 00000bfc 2**1
69c62 < 4 .gnu.version_r 000000e0 0000d208 0000d208 0000d208 2**2 ---
4 .gnu.version_r 00000060 00000cb8 00000cb8 00000cb8 2**2
71c64 < 5 .rel.dyn 00009d58 0000d2e8 0000d2e8 0000d2e8 2**2 ---
5 .rel.dyn 00000260 00000d18 00000d18 00000d18 2**2
73c66 < 6 .rel.plt 00001d38 00017040 00017040 00017040 2**2 ---
6 .rel.plt 000001a0 00000f78 00000f78 00000f78 2**2
75c68 < 7 .init 00000017 00018d78 00018d78 00018d78 2**2 ---
7 .init 00000017 00001118 00001118 00001118 2**2
77c70 < 8 .plt 00003a80 00018d90 00018d90 00018d90 2**2 ---
8 .plt 00000350 00001130 00001130 00001130 2**2
79c72 < 9 .text 000c2b24 0001c810 0001c810 0001c810 2**4 ---
9 .text 00005c84 00001480 00001480 00001480 2**4
81c74 < 10 .fini 0000001c 000df334 000df334 000df334 2**2 ---
10 .fini 0000001c 00007104 00007104 00007104 2**2
83c76 < 11 .rodata 00020e68 000df360 000df360 000df360 2**5 ---
11 .rodata 00000a8a 00007120 00007120 00007120 2**5
85c78 < 12 .eh_frame_hdr 000052cc 001001c8 001001c8 001001c8 2**2 ---
12 .eh_frame_hdr 000001cc 00007bac 00007bac 00007bac 2**2
87c80 < 13 .eh_frame 00014978 00105494 00105494 00105494 2**2 ---
13 .eh_frame 000006c4 00007d78 00007d78 00007d78 2**2
89c82 < 14 .ctors 00000008 0011a000 0011a000 0011a000 2**2 ---
14 .ctors 00000008 00009000 00009000 00009000 2**2
91c84 < 15 .dtors 00000008 0011a008 0011a008 0011a008 2**2 ---
15 .dtors 00000008 00009008 00009008 00009008 2**2
93c86 < 16 .jcr 00000004 0011a010 0011a010 0011a010 2**2 ---
16 .jcr 00000004 00009010 00009010 00009010 2**2
95c88 < 17 .data.rel.ro 000000cc 0011a020 0011a020 0011a020 2**5 ---
17 .data.rel.ro 00000004 00009014 00009014 00009014 2**2
97c90 < 18 .dynamic 000000d0 0011a0ec 0011a0ec 0011a0ec 2**2 ---
18 .dynamic 000000d8 00009018 00009018 00009018 2**2
99c92 < 19 .got 0000036c 0011a1bc 0011a1bc 0011a1bc 2**2 ---
19 .got 0000008c 000090f0 000090f0 000090f0 2**2
101c94 < 20 .got.plt 00000ea8 0011a528 0011a528 0011a528 2**2 ---
20 .got.plt 000000dc 0000917c 0000917c 0000917c 2**2
103c96 < 21 .data 00023ce0 0011b3e0 0011b3e0 0011b3e0 2**5 ---
21 .data 00000390 00009260 00009260 00009260 2**5
105c98 < 22 .bss 0000684c 0013f0c0 0013f0c0 0013f0c0 2**5 ---
22 .bss 000008a4 00009600 00009600 000095f0 2**5
107c100 < 23 .gnu_debuglink 00000018 00000000 00000000 0013f0c0 2**2 ---
23 .gnu_debuglink 00000018 00000000 00000000 000095f0 2**2
Looks like the bad time frame for python packages is about Dec 8 -> Jan 6. python-enchant built on Jan 14 looks okay.
Relevant python changelog:
* Sat Jan 06 2007 Jeremy Katz katzj@redhat.com - 2.5.3-8 - fix extensions to use shared libpython (#219564) - all 64bit platforms need the regex fix (#122304)
Looks like all non-noarch python modules/apps built before 2.5.3-8 should be rebuilt.
Orion Poplawski wrote:
Looks like the bad time frame for python packages is about Dec 8 -> Jan 6. python-enchant built on Jan 14 looks okay.
Relevant python changelog:
- Sat Jan 06 2007 Jeremy Katz katzj@redhat.com - 2.5.3-8
- fix extensions to use shared libpython (#219564)
- all 64bit platforms need the regex fix (#122304)
Looks like all non-noarch python modules/apps built before 2.5.3-8 should be rebuilt.
These are all the ones with "py" in the binary rpm name. Probably should check for BR python-devel or something.
apt-0.5.15lorg3.2-9.fc7.src.rpm hamlib-1.2.5-4.fc7.src.rpm libpreludedb-0.9.11.1-2.fc7.src.rpm MySQL-python-1.2.1_p2-2.src.rpm notify-python-0.1.0-4.fc7.src.rpm OpenIPMI-2.0.6-7.fc7.src.rpm pybluez-0.9.1-3.fc7.src.rpm pyflowtools-0.3-8.fc7.src.rpm pyfribidi-0.6.0-3.fc7.src.rpm pygpgme-0.1-4.fc7.src.rpm pygsl-0.3.2-7.fc7.src.rpm pyOpenSSL-0.6-1.p24.9.src.rpm pypoker-eval-133.0-1.fc7.src.rpm python-adns-1.1.0-5.fc7.src.rpm python-alsaaudio-0.2-2.fc7.src.rpm python-CDDB-1.4-1.fc7.src.rpm python-chm-0.8.4-1.fc7.src.rpm python-durus-3.5-3.fc7.src.rpm python-GeoIP-1.2.1-6.fc7.src.rpm python-krbV-1.0.13-5.fc7.src.rpm python-ldap-2.2.0-3.src.rpm python-numarray-1.5.2-2.fc7.src.rpm python-psycopg-1.1.21-6.fc7.src.rpm python-pycurl-7.16.0-0.1.20061207.fc7.src.rpm python-quixote-2.4-5.fc7.src.rpm python-sexy-0.1.9-3.fc7.src.rpm python-smbpasswd-1.0.1-5.fc7.src.rpm python-twisted-conch-0.7.0-4.fc7.src.rpm python-twisted-core-2.4.0-6.fc7.src.rpm python-twisted-lore-0.2.0-4.fc7.src.rpm python-twisted-mail-0.3.0-4.fc7.src.rpm python-twisted-names-0.3.0-3.fc7.src.rpm python-twisted-runner-0.2.0-4.fc7.src.rpm python-twisted-web-0.6.0-4.fc7.src.rpm python-twisted-words-0.4.0-3.fc7.src.rpm python-zope-interface-3.0.1-7.fc7.src.rpm pyxmms-2.06-4.fc7.src.rpm
Replying to Orion Poplawski:
pyflowtools-0.3-8.fc7.src.rpm
Can I consider this a blocker and rebuild my package in F-7?
Thanks.
On Thu, May 17, 2007 at 04:04:45PM -0600, Orion Poplawski wrote:
Looks like the bad time frame for python packages is about Dec 8 -> Jan 6. python-enchant built on Jan 14 looks okay.
Relevant python changelog:
- Sat Jan 06 2007 Jeremy Katz katzj@redhat.com - 2.5.3-8
- fix extensions to use shared libpython (#219564)
- all 64bit platforms need the regex fix (#122304)
Looks like all non-noarch python modules/apps built before 2.5.3-8 should be rebuilt.
Yet another reason for doing a mass rebuild at the end of the development cycle. Bad time windows of python, gcc and whatnot have always existed and will continue to do so. Sometimes we notice, sometimes not.
That's the price for piecing together a distribution from different states of the environment from over 8 months. Oh and did I mention that changes in glibc (chownat) *run-time* broke the fakeroot packages as well? So much for glibc hasn't changed enough to warrant rebuilds.
Mass rebuilds for F7 are out of the question now, but they should become mandatory for F8 again. Once towards the end of a development cycle before freezing.
I just hope this isn't the tip of the iceberg, otherwise F7 will be re-codenamed Titanic. :(
Orion Poplawski wrote, at 05/18/2007 05:35 AM +9:00:
I did a quick compare of FE 6 to the last FE devel and found that the following packages have increased in size by more than a factor of 2. Some of these may be fine, I haven't looked into it in detail.
wallpapoz 110963 -> 363890
Well, wallpapoz is maintained by me. During FE6->F7 wallpapoz version changed FE-6: 0.3-1.fc6 F-7: 0.4-0.5.rc2.fc7 (Currently I don't have a plan to update FE-6 wallpapoz to 0.4rc2) and actually the contents increased (with new function, translation, etc...). So for this package thereis no problem.
BTW I like wallpapoz very much.
Mamoru
Orion Poplawski wrote:
I did a quick compare of FE 6 to the last FE devel and found that the following packages have increased in size by more than a factor of 2. Some of these may be fine, I haven't looked into it in detail.
Thanks for the effort, however it looks like you've looked at the combined size of all subpackages of a package, for example ogre which is mine:
ogre 6374746 -> 29476134
Has a main package of only 2 Mb, most of the size increase is explained by the adding of a new samples sub package which ways in at 15Mb (previously the samples were unpackaged).
The rest of the size increase seems to be due to errors in the way you've calculated the size, the FE6 ogre, excluding debuginfo ways in at 10 Mb (x86_64), the F7 one at 25 Mb.
Regards,
Hans
Hans de Goede wrote:
Orion Poplawski wrote:
I did a quick compare of FE 6 to the last FE devel and found that the following packages have increased in size by more than a factor of 2. Some of these may be fine, I haven't looked into it in detail.
Thanks for the effort, however it looks like you've looked at the combined size of all subpackages of a package, for example ogre which is mine:
ogre 6374746 -> 29476134
Has a main package of only 2 Mb, most of the size increase is explained by the adding of a new samples sub package which ways in at 15Mb (previously the samples were unpackaged).
The rest of the size increase seems to be due to errors in the way you've calculated the size, the FE6 ogre, excluding debuginfo ways in at 10 Mb (x86_64), the F7 one at 25 Mb.
I'm not looking at the size of the rpms, I'm looking at installed sizes as reported by rpm -q --qf "%{SIZE}\n" -p <rpm>.
# rpm -q --qf '%{SIZE}\n' -p development/i386/ogre-1.2.5-1.fc7.i386.rpm 29476134 # rpm -q --qf '%{SIZE}\n' -p 6/i386/ogre-1.2.2-2.p1.fc6.i386.rpm 6374746
I installed ogre on my devel system however and the total installed size appears to be about 8476 KB. So perhaps a bug in rpm's size calculation?
# rpm -ql ogre | xargs du -k | sort -n | awk '{ tot+=$1 }; END { print tot }' 8476
Orion Poplawski wrote:
Hans de Goede wrote:
Orion Poplawski wrote:
I did a quick compare of FE 6 to the last FE devel and found that the following packages have increased in size by more than a factor of 2. Some of these may be fine, I haven't looked into it in detail.
Thanks for the effort, however it looks like you've looked at the combined size of all subpackages of a package, for example ogre which is mine:
ogre 6374746 -> 29476134
Has a main package of only 2 Mb, most of the size increase is explained by the adding of a new samples sub package which ways in at 15Mb (previously the samples were unpackaged).
The rest of the size increase seems to be due to errors in the way you've calculated the size, the FE6 ogre, excluding debuginfo ways in at 10 Mb (x86_64), the F7 one at 25 Mb.
I'm not looking at the size of the rpms, I'm looking at installed sizes as reported by rpm -q --qf "%{SIZE}\n" -p <rpm>.
# rpm -q --qf '%{SIZE}\n' -p development/i386/ogre-1.2.5-1.fc7.i386.rpm 29476134 # rpm -q --qf '%{SIZE}\n' -p 6/i386/ogre-1.2.2-2.p1.fc6.i386.rpm 6374746
I installed ogre on my devel system however and the total installed size appears to be about 8476 KB. So perhaps a bug in rpm's size calculation?
Perhaps, or maybe rpm is showing the size of all subpackages together?
Regards,
Hans
On Fri, 2007-05-18 at 19:36 +0200, Hans de Goede wrote:
Orion Poplawski wrote:
I'm not looking at the size of the rpms, I'm looking at installed sizes as reported by rpm -q --qf "%{SIZE}\n" -p <rpm>.
# rpm -q --qf '%{SIZE}\n' -p development/i386/ogre-1.2.5-1.fc7.i386.rpm 29476134 # rpm -q --qf '%{SIZE}\n' -p 6/i386/ogre-1.2.2-2.p1.fc6.i386.rpm 6374746
I installed ogre on my devel system however and the total installed size appears to be about 8476 KB. So perhaps a bug in rpm's size calculation?
Perhaps, or maybe rpm is showing the size of all subpackages together?
It's a bug in rpm. Namely, these lines from ogre.spec are throwing things off: %exclude %{_bindir}/Ogre-Samples %exclude %{_libdir}/OGRE/Samples %exclude %{_datadir}/OGRE/Samples
Rpm's size calculation doesn't properly handle %exclude.
-Toshio
On Fri, 2007-05-18 at 11:58 -0700, Toshio Kuratomi wrote:
It's a bug in rpm. Namely, these lines from ogre.spec are throwing things off: %exclude %{_bindir}/Ogre-Samples %exclude %{_libdir}/OGRE/Samples %exclude %{_datadir}/OGRE/Samples
Rpm's size calculation doesn't properly handle %exclude.
Hmm we should be:
Patch13: rpm-4.4.2-excluded-size.patch
--- rpm-4.4.2/build/files.c.excludedsize 2005-11-29 16:21:12.000000000 -0500 +++ rpm-4.4.2/build/files.c 2005-11-29 16:21:21.000000000 -0500 @@ -1678,7 +1678,7 @@ } else i = fl->fileListRecsUsed;
- if (S_ISREG(flp->fl_mode) && i >= fl->fileListRecsUsed) + if (!(flp->flags & RPMFILE_EXCLUDE) && S_ISREG(flp->fl_mode) && i >= fl->fileListRecsUsed) fl->totalFileSize += flp->fl_size; }
Paul
On Mon, 21 May 2007 16:40:16 +0100 Paul Nasrat pnasrat@redhat.com wrote:
On Fri, 2007-05-18 at 11:58 -0700, Toshio Kuratomi wrote:
It's a bug in rpm. Namely, these lines from ogre.spec
are throwing
things off: %exclude %{_bindir}/Ogre-Samples %exclude %{_libdir}/OGRE/Samples %exclude %{_datadir}/OGRE/Samples
Rpm's size calculation doesn't properly handle
%exclude.
Hmm we should be:
Patch13: rpm-4.4.2-excluded-size.patch
--- rpm-4.4.2/build/files.c.excludedsize 2005-11-29 16:21:12.000000000 -0500 +++ rpm-4.4.2/build/files.c 2005-11-29 16:21:21.000000000 -0500 @@ -1678,7 +1678,7 @@ } else i = fl->fileListRecsUsed;
- if (S_ISREG(flp->fl_mode) && i >= fl->fileListRecsUsed)
- if (!(flp->flags & RPMFILE_EXCLUDE) &&
S_ISREG(flp->fl_mode) && i >= fl->fileListRecsUsed) fl->totalFileSize += flp->fl_size; }
Well,
I can confirm that this does NOT work properly on a recent rawhide system.
Maybe the problem is that I exclude dirs, and that the size calculation still dives into the dir, and doesn't see the files in there as excluded?
Regards,
Hans
On Mon, 2007-05-21 at 18:00 +0200, Goede, J.W.R. de wrote:
Well,
I can confirm that this does NOT work properly on a recent rawhide system.
Maybe the problem is that I exclude dirs, and that the size calculation still dives into the dir, and doesn't see the files in there as excluded?
Yeah that seems to be the case, can you come up with a simple test spec reproducer and file in bugzilla please.
Paul
Paul Nasrat wrote:
On Mon, 2007-05-21 at 18:00 +0200, Goede, J.W.R. de wrote:
Well,
I can confirm that this does NOT work properly on a recent rawhide system.
Maybe the problem is that I exclude dirs, and that the size calculation still dives into the dir, and doesn't see the files in there as excluded?
Yeah that seems to be the case, can you come up with a simple test spec reproducer and file in bugzilla please.
Paul
Did anyone else think their spam filter had failed to catch this thread? I came real close to hitting the junk button in Thunderbird when I read the subject line. :-)
Brent
Paul Nasrat wrote:
On Mon, 2007-05-21 at 18:00 +0200, Goede, J.W.R. de wrote:
Well,
I can confirm that this does NOT work properly on a recent rawhide system.
Maybe the problem is that I exclude dirs, and that the size calculation still dives into the dir, and doesn't see the files in there as excluded?
Yeah that seems to be the case, can you come up with a simple test spec reproducer and file in bugzilla please.
Paul
Hi Paul,
Sometime ago we had the above discussion, should I still BZ this, ifso in which BZ, RH or rpm.org?
Regards,
Hans
On Wed, 4 Jul 2007, Hans de Goede wrote:
Paul Nasrat wrote:
On Mon, 2007-05-21 at 18:00 +0200, Goede, J.W.R. de wrote:
Well,
I can confirm that this does NOT work properly on a recent rawhide system.
Maybe the problem is that I exclude dirs, and that the size calculation still dives into the dir, and doesn't see the files in there as excluded?
Yeah that seems to be the case, can you come up with a simple test spec reproducer and file in bugzilla please.
Paul
Hi Paul,
Sometime ago we had the above discussion, should I still BZ this, ifso in which BZ, RH or rpm.org?
Paul's on vacation but yeah, please bugzilla. rpm.org doesn't at the moment have it's own bug tracking so just use RH BZ.
- Panu -