[PATCH] make dir/file.o is still broken :(

Bernhard Fischer rep.nop at aon.at
Tue Mar 14 18:55:00 UTC 2006


On Tue, Mar 14, 2006 at 01:03:15PM -0500, Rob Landley wrote:
>On Tuesday 14 March 2006 1:11 am, Denis Vlasenko wrote:
>> On Monday 13 March 2006 17:28, Rob Landley wrote:
>> > > make: *** [applets/applets.o] Error 1
>> > > # make applets/applets.s
>> > > make: *** No rule to make target `applets/applets.s'.  Stop.
>> >
>> > Why on earth are you trying to make an assembly file?  We don't do that
>> > anywhere in the tree, to we?  It's not portable...
>>
>> I use it to see what actual code is generated. Helps to understand
>> where exactly it grows too much. Remember those "shrink unlzma by 50%"
>> patches?
>
>Ah.  I generally just fiddle around with objdump -D if I get curious...
>
>> > > It also adds a "make dir/file.i" target for generating preprocessed C
>> > > file.
>> >
>> > Why would we want to do this?
>>
>> For debugging reasons.
>
>Ok.
>
>Bernhard reverted the make thing so it's moot at the moment...

I agree with vda that this is a reasonably handy thing to have. Do we
agree to agree to revisit this later? :)

PS: Rob, i constanly get
libbb/copy_file.c:124: warning: 'dst_fd' may be used uninitialized in
this function

which is really annoying. As i've had my share of borking stuff ATM, can
you please checkin some variant of

Index: libbb/copy_file.c
===================================================================
--- libbb/copy_file.c	(revision 14537)
+++ libbb/copy_file.c	(working copy)
@@ -121,7 +121,7 @@ int copy_file(const char *source, const 
 	} else if (S_ISREG(source_stat.st_mode) || (flags &
FILEUTILS_DEREFERENCE))
 	{
 		int src_fd;
-		int dst_fd;
+		int dst_fd = 0 or -1
 #ifdef CONFIG_FEATURE_PRESERVE_HARDLINKS
 		char *link_name;
 

TIA,



More information about the busybox mailing list