Richard W.M. Jones wrote:
On Mon, Sep 17, 2012 at 10:47:52AM +0200, Jim Meyering wrote:
trailing_blank TODO:405: - swap devices (both of block device and file) should be wiped. This may Binary file tests/guests/guest-aux/windows-software matches Binary file tests/guests/guest-aux/windows-system matches tools/virt-win-reg:711: maint.mk: found trailing blank(s) make: *** [sc_trailing_blank] Error 1
Binary files should probably not be matched.
Hey! I've just realized that I can easily filter out those lines. Here's yet another patch. We should be able to use a similar change for a few other rules.
diff --git a/top/maint.mk b/top/maint.mk index 4627bc5..ccf09a2 100644 --- a/top/maint.mk +++ b/top/maint.mk @@ -724,6 +724,7 @@ sc_require_test_exit_idiom: sc_trailing_blank: @prohibit='[ ]$$' \ halt='found trailing blank(s)' \ + exclude='^Binary file .* matches$$' \ $(_sc_search_regexp)
# Match lines like the following, but where there is only one space
Yes, this works to exclude the false binary matches.
Thanks for testing those. I've pushed the changes to gnulib: Subject: 0/4: syntax-check: tweak for fewer false positives Newsgroups: gmane.comp.lib.gnulib.bugs Richard W.M. Jones went through the (tedious!) process of reviewing all of libguestfs wrt gnulib's latest syntax-check rules, and reported many false positives. In a few cases, I found ways to avoid entire classes of false positives: 1/4 maint.mk: relax sc_prohibit_strcmp, to avoid a false positive 2/4 maint.mk: teach sc_prohibit_magic_number_exit to accept 77 3/4 maint.mk: sc_prohibit_path_max_allocation: don't FP for UNIX_PATH_MAX 4/4 maint.mk: exempt trailing blanks found in "binary" files