[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: librcsb-core-wrapper read for inspection



* Laszlo Kajan <lkajan@rostlab.org>, 2012-08-16, 21:07:
I am ready with librcsb-core-wrapper for inspection.

As per Python Policy §2.2, the binary package name should be python-corepywrap. (If that sounds awfully generic to you, then it's only because the module name upstream chose is like that...)

I recommend not using ${python:Provides}. See
<http://lists.debian.org/20110324164804.GA5919@jwilk.net> for rationale.

Lintian says:

W: python-librcsb-core-wrapper: hardening-no-fortify-functions usr/lib/python2.6/dist-packages/CorePyWrap.so
W: python-librcsb-core-wrapper: hardening-no-fortify-functions usr/lib/python2.7/dist-packages/CorePyWrap.so

which might be false-positive, but on the other hand blhc seems to confirm that *FLAGS are lost somewhere:

CFLAGS missing (-g -O2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security): libtool --mode=compile gcc -D_FORTIFY_SOURCE=2 -O  -fPIC    -DHAVE_STRCASECMP -DINCL_TEMPLATE_SRC -DHAVE_PLACEMENT_NEW  -I./include -I../include    -DPOSIX_MISTAKE -c src/regcomp.c -o ./obj/regcomp.o
[snip - more complaints about CFLAGS missing]
LDFLAGS missing (-Wl,-z,relro): g++ -D_FORTIFY_SOURCE=2  -w -L/usr/lib obj/xml2mmcif.o ../lib/pdbml-parser.a ../lib/dict-obj-file.a ../lib/cif-file-util.a ../lib/cif-file.a ../lib/cifparse-obj.a ../lib/tables.a ../lib/common.a ../lib/regex.a -lxerces-c  -lm -o ./bin/xml2mmcif

dh_python2 manpage says that python-all-dev build-dependency should be
versioned: (>= 2.6.6-3~). That's only needed to be pedantically correct,
because the needed version is already in squeeze.

It's customary to build extension modules also with python2.X-dbg interpreters, and put them into a separate python-foo-dbg package. If you build-depend on python-all-dbg then dh_auto_* will do most of the work for you. (It's a feature added in debhelper 7.3.5, so you should bump debhelper build-dependency if you decide to use it.)

--
Jakub Wilk


Reply to: