Re: ld: table of contents for archive out of date?
Re: ld: table of contents for archive out of date?
- Subject: Re: ld: table of contents for archive out of date?
- From: Dair Grant <email@hidden>
- Date: Wed, 28 Sep 2005 22:51:53 +0100
Andrew Gallatin wrote:
>>Is this use case still valid? Is this warning useful at all? I
>>prefer the linker to fail if it cannot find symbols to link to, not
>>if it doesn't like the timestamp of the library I am linking to.
>
>Seconded, whole heartedly. This is one of my biggest pet peeves
>about the darwin toolchain.
We've also run into this issue, and found it a little bizarre that
simply copying a .a file around can "break" it by virtue of bumping its
mod date...
Our evil workaround, which really requires cvs, is to run ranlib prior
to check in and then to hack the ,v file by hand to modify the check in
date for that revision. I.e., run ranlib to insert the current time
(22:43) into the .a file, check it in, and edit the ,v file to rewind
the check in time to 22:30.
Provided your cvs client sets the local mod time to match the check in
time when checking out, after check out the mod time will be earlier
than the "ranlib time" and so the library remains linkable.
It is gruesome, but means we can keep 3rd party Mach-O libs under
revision control the same way we keep CW/VC++ libs.
-dair
___________________________________________________
mailto:dair+refnum.com http://www.refnum.com/
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden