make depend in loader/ on darwin/osx will give invalid gcc option

[sci-fi@…]

I'm splitting this from bug #784 to keep it separate. The logs there will show that we're seeing invalid gcc options -mno-omit-leaf-frame-pointer during make depend when it traverses into the loader/ subdir. Indeed the Makefile for loader/ is not doing cc_check for that gcc option but instead it is hard-coded.

I don't know why loader/ is needing to be traversed for darwin/osx during make depend, I believe it is completely skipped during normal make all on darwin isn't it? So I don't know how to fix this for real, but offer the attached patch named mplayer_loader_makedepend.patch to remove that offending gcc option just so the rest of make depend can continue for darwin. ;)

Again I know this probably is not the right way to fix it, but something needs to be done to help make depend go smooth for darwin.

Thanks.

[sci-fi@…]

gets rid of the offending gcc option for darwin/osx

[diego@…]

--- loader/Makefile_orig 2007-03-18 04:21:47 -0500
+++ loader/Makefile 2007-04-12 04:41:34 -0500
@@ -32,7 +32,7 @@

-CFLAGS+=-fno-omit-frame-pointer -mno-omit-leaf-frame-pointer
+CFLAGS+=-fno-omit-frame-pointer

Hopefully you're not serious about unconditionally removing CFLAGS that happen to fail on one platform..

Rejected with extreme prejudice

[sci-fi@…]

(In reply to comment #2)

(From update of attachment 352 [details])

--- loader/Makefile_orig 2007-03-18 04:21:47 -0500
+++ loader/Makefile 2007-04-12 04:41:34 -0500
@@ -32,7 +32,7 @@

-CFLAGS+=-fno-omit-frame-pointer -mno-omit-leaf-frame-pointer
+CFLAGS+=-fno-omit-frame-pointer

Hopefully you're not serious about unconditionally removing CFLAGS that happen
to fail on one platform..

Rejected with extreme prejudice

Diego,

I'm having extreme problem with your attitude.
Do you understand what I typed earlier?
I said "I know this probably is not the right way".
I KNOW it wasn't!
Good grief man, why do you consistently rub us the wrong way like this?

No one is blaming you or anyone else. I'm trying to help. If I did not care, I would not want to help.

All I can do is report a problem and present a patch to show what fixed it for us. I KNOW it wouldn't be accepted and said so right from the start -- good grief man you really know how to push ppl's buttons here don't you.

Why don't you try having a nicer tone with us? I KNOW it wasn't an acceptable patch and I ALREADY TOLD YOU I KNEW!

You're easily alienating me when you talk that way. You need to treat us like you want to be treated -- "with respect". Some of us don't have thick skins like you. We're going to stop volunteering if we can't have simple plain respect.

Thank you for understanding -- but act as if you do understand anyway, please.

Dead serious about this.

Now, let's find out what the right way will be. I wish I had your brain...

Thank you for taking time to help with this problem.

[reimar]

Note that not using -mno-omit-leaf-frame-pointer will break the loader code, so it IMO should never even be made conditional (though it might not be necessary for the dshow code parts).
Not checking the loader code at all seems like the only option thus.

[diego@…]

(In reply to comment #3)

(In reply to comment #2)

(From update of attachment 352 [details] [details])

--- loader/Makefile_orig 2007-03-18 04:21:47 -0500
+++ loader/Makefile 2007-04-12 04:41:34 -0500
@@ -32,7 +32,7 @@

-CFLAGS+=-fno-omit-frame-pointer -mno-omit-leaf-frame-pointer
+CFLAGS+=-fno-omit-frame-pointer

Hopefully you're not serious about unconditionally removing CFLAGS that
happen to fail on one platform..

Rejected with extreme prejudice

Diego,

I'm having extreme problem with your attitude.
Do you understand what I typed earlier?
I said "I know this probably is not the right way".
I KNOW it wasn't!
Good grief man, why do you consistently rub us the wrong way like this?

I wasn't really serious about this, but from rereading my post I can see that no joke is visible for an outside observer. Please accept my apologies.

Now, let's find out what the right way will be. I wish I had your brain...

Thank you for taking time to help with this problem.

On a slightly more constructive note: Do I understand correctly that this is only a problem during make depend, not during make?

[diego@…]

This is no longer a problem now that we have per-file dependencies and a non-recursive build system.