[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[Thomas Anderson]

[-- Attachment #1: Type: text/plain, Size: 759 bytes --]

On Sun, Oct 26, 2008 at 05:22:26PM +0000, Greg Kroah-Hartman (gregkh) wrote:
> # Copyright 1999-2008 Gentoo Foundation
> # Distributed under the terms of the GNU General Public License v2
> # $Header: /var/cvsroot/gentoo-x86/net-misc/bti/bti-007.ebuild,v 1.1 2008/10/26 17:22:26 gregkh Exp $
> 
> RDEPEND="${DEPEND}"
> 
> src_compile() {
> 	emake || die "emake failed"
> }
This is the default src_compile.
> 
> src_install() {
> 	doman bti.1
> 	dobin bti
> 	dodoc bti.example README RELEASE-NOTES
> }
You really should have some or all of these functions die on failure.
Since that's all the ebuild installs, the package is completely
nonfunctional if bti is not installed(I'd die on all three but you don't
have to die on the last one).

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[Greg KH]

On Sun, Oct 26, 2008 at 01:36:29PM -0400, Thomas Anderson wrote:
> On Sun, Oct 26, 2008 at 05:22:26PM +0000, Greg Kroah-Hartman (gregkh) wrote:
> > # Copyright 1999-2008 Gentoo Foundation
> > # Distributed under the terms of the GNU General Public License v2
> > # $Header: /var/cvsroot/gentoo-x86/net-misc/bti/bti-007.ebuild,v 1.1 2008/10/26 17:22:26 gregkh Exp $
> > 
> > RDEPEND="${DEPEND}"
> > 
> > src_compile() {
> > 	emake || die "emake failed"
> > }
> This is the default src_compile.

Ah, thanks, I should just drop this then.

> > src_install() {
> > 	doman bti.1
> > 	dobin bti
> > 	dodoc bti.example README RELEASE-NOTES
> > }
> You really should have some or all of these functions die on failure.

Why would any of these fail if the src_compile succeeded?

And, for some reason I thought that the default was that if there was an
error in them, they would "die" on their own.  Is that not the case?

> Since that's all the ebuild installs, the package is completely
> nonfunctional if bti is not installed(I'd die on all three but you don't
> have to die on the last one).

Might as well be consistent :)

thanks for the review,

greg k-h

Re: [gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[Thomas Sachau]

[-- Attachment #1: Type: text/plain, Size: 780 bytes --]

Greg KH schrieb:
>>> src_install() {
>>> 	doman bti.1
>>> 	dobin bti
>>> 	dodoc bti.example README RELEASE-NOTES
>>> }
>> You really should have some or all of these functions die on failure.
> 
> Why would any of these fail if the src_compile succeeded?

What happens, if upstream changes the name of the binary on version bump or changes the compile
output location?

> 
> And, for some reason I thought that the default was that if there was an
> error in them, they would "die" on their own.  Is that not the case?

Sadly not. Some functions do die (like epatch or econf), others do not (like emake, dobin,
doman....). If i remember correctly, those that are external functions do die, the others do not.


-- 
Thomas Sachau

Gentoo Linux Developer


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 315 bytes --]

[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[Thomas Anderson]

[-- Attachment #1: Type: text/plain, Size: 1055 bytes --]

On Sun, Oct 26, 2008 at 02:58:02PM -0700, Greg KH wrote:
> On Sun, Oct 26, 2008 at 01:36:29PM -0400, Thomas Anderson wrote:
> > On Sun, Oct 26, 2008 at 05:22:26PM +0000, Greg Kroah-Hartman (gregkh) wrote:
> > > src_install() {
> > > 	doman bti.1
> > > 	dobin bti
> > > 	dodoc bti.example README RELEASE-NOTES
> > > }
> > You really should have some or all of these functions die on failure.
> 
> Why would any of these fail if the src_compile succeeded?
"Succeeded" and "Not error out" are different. Src_compile could not
error out but still not produce the executables/produce an executable
with a different name. Either way you end up with a broken installation
of bti.
> And, for some reason I thought that the default was that if there was an
> error in them, they would "die" on their own.  Is that not the case?
No, they don't die on their own. As far as I know, very few ebuild
functions do. Econf does, and epatch does as well. I'll look into
writing up a list of which functions die and which don't though, for
convenience.

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: [gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[Greg KH]

On Sun, Oct 26, 2008 at 11:17:14PM +0100, Thomas Sachau wrote:
> Greg KH schrieb:
> >>> src_install() {
> >>> 	doman bti.1
> >>> 	dobin bti
> >>> 	dodoc bti.example README RELEASE-NOTES
> >>> }
> >> You really should have some or all of these functions die on failure.
> > 
> > Why would any of these fail if the src_compile succeeded?
> 
> What happens, if upstream changes the name of the binary on version
> bump or changes the compile output location?

Then I will go and swat upstream upside the head for not realizing that
he also needed to fix the ebuild when he bumbed the version :)

(hint, upstream and me are the same...)

thanks,

greg k-h

[gentoo-dev] Re: [gentoo-commits] gentoo-x86 commit in net-misc/bti: bti-007.ebuild ChangeLog metadata.xml Manifest

[Greg KH]

On Sun, Oct 26, 2008 at 06:29:13PM -0400, Thomas Anderson wrote:
> On Sun, Oct 26, 2008 at 02:58:02PM -0700, Greg KH wrote:
> > On Sun, Oct 26, 2008 at 01:36:29PM -0400, Thomas Anderson wrote:
> > > On Sun, Oct 26, 2008 at 05:22:26PM +0000, Greg Kroah-Hartman (gregkh) wrote:
> > > > src_install() {
> > > > 	doman bti.1
> > > > 	dobin bti
> > > > 	dodoc bti.example README RELEASE-NOTES
> > > > }
> > > You really should have some or all of these functions die on failure.
> > 
> > Why would any of these fail if the src_compile succeeded?
> "Succeeded" and "Not error out" are different. Src_compile could not
> error out but still not produce the executables/produce an executable
> with a different name. Either way you end up with a broken installation
> of bti.
> > And, for some reason I thought that the default was that if there was an
> > error in them, they would "die" on their own.  Is that not the case?
> No, they don't die on their own. As far as I know, very few ebuild
> functions do. Econf does, and epatch does as well. I'll look into
> writing up a list of which functions die and which don't though, for
> convenience.

Ok, thanks for the information, I appreciate it.

greg k-h