Re: lipo doesn't touch (in the timestamp modification sense) the output file?
Re: lipo doesn't touch (in the timestamp modification sense) the output file?
- Subject: Re: lipo doesn't touch (in the timestamp modification sense) the output file?
- From: Rush Manbert <email@hidden>
- Date: Mon, 27 Mar 2006 17:12:50 -0800
Alexei Kosut wrote:
On Mar 27, 2006, at 2:08 PM, Rush Manbert wrote:
I am using lipo to create fat libraries (ppc & i386). It runs as part of
a script that is doing this for about a hundred libraries and
executables.
For each individual lipo operation, I have two source libraries. One is
for ppc, and was built and checked in on 22 Feb 2006. The other is Intel
and was built 24 March 2006. When I combine them using lipo, the output
file has a 22 February 2006 timestamp. It doesn't seem to matter which
order I specify the two source libraries, the output always has the
earlier timestamp (and not the current timestamp).
Is this what lipo is supposed to do?
Yes. The source code to lipo is available as part of the cctools
package in the Darwin source release of Xcode Tools 2.2, and if you
take a look, you will find the following comment at <http://
www.opensource.apple.com/darwinsource/DevToolsNov2005/cctools-590.18/
misc/lipo.c>:
/*
* Select the eariliest modifiy time so that if the output file
* contains archives with table of contents lipo will not make them
* out of date. This logic however could make an out of date table of
* contents appear up todate if another file is combined with it that
* has a date early enough.
*/
In other words, lipo is trying hard not to invalidate the table of
contents of a static archive, so you do not have to re-run ranlib just
because you've added or removed an architecture.
So your diagnosis is correct, and if this isn't the behavior you want,
you should change the modification date yourself, e.g., using touch.
And as Eric already replied, you should also file a bug with Apple if
you think the tool's behavior should be changed.
Thanks Alexei! That's interesting, and I can understand the motivation.
I think I stand behind the bug report, though. I'd rather it touched the
timestamp (assuming, of course, that ranlib works for fat libraries...)
Best regards,
Rush
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden