Re: [edk2] [PATCH] OvmfPkg/build.sh: update gcc detection

[Jordan Justen]

On 2016-07-05 07:37:58, Laszlo Ersek wrote:
> On 06/30/16 12:16, Laszlo Ersek wrote:
> > On 06/30/16 07:00, Olaf Hering wrote:
> >> On Wed, Jun 29, Jordan Justen wrote:
> >>
> >>> Missing Contributed-under. (See OvmfPkg/Contributions.txt)
> >>
> >> Looks like this project tries to avoid simple fixes from third party.
> > 
> > It's not the case. Every project has its contribution rules (I reckon
> > xen-devel is no exception). It takes some time to set everything up in
> > order to contribute in accordance with the rules, but that's a one time
> > cost for every contributor.
> > 
> > If we encouraged drive-by patches without regard to the process, then it
> > would be a constant cost for long term participants / maintainers.
> > 
> > I do agree with you on two points -- I find these unjustified /
> > gratuitous burdens:
> > - the Contributed-under line
> > - having to subscribe to the mailing list in order to post
> > 
> > The first is alas a legal requirement, and I can't do anything about it.
> > Luckily, it can be automated at least. See for example
> > 
> > https://github.com/tianocore/tianocore.github.io/wiki/Laszlo's-unkempt-git-guide-for-edk2-contributors-and-maintainers#contrib-07
> > 
> > The second (the subscription requirement) is terrible. No other open
> > source project I'm aware of follows that requirement. I've raised it
> > several times, but other subscribers on the list disagree with me.
> > 
> >> I'm done with it.
> > 
> > I'm sorry to hear that. I don't really see why you can't post a v2 of
> > your patch, with all the remarks addressed, considering you will likely
> > carry such a patch in your downstream anyway. I encourage you to send a v2.
> 
> I guess you decided not to do it. :(
> 
> Jordan: do you think I can pick up Olaf's patch and resubmit it with
> your remarks addressed? I'm asking because the original patch did not
> have the Contributed-under line, and I wonder if it's okay if I add:
>

If Olaf is not willing or able to add Contributed-under for his patch,
then I think we need to skip it.

We can re-implement it pretty easily after a few weeks. Maybe we can
just address the issue at the same time that we add support for clang
and gcc6, when Steven Shi finishes those.

-Jordan

>   Signed-off-by: Olaf Hering 
>   Fixes: https://github.com/tianocore/edk2/issues/99
>   [ler...@redhat.com: update commit message]
>   Contributed-under: TianoCore Contribution Agreement 1.0
>   Signed-off-by: Laszlo Ersek 
> 
> I think we've seen such examples on the list recently (i.e., one author
> writing the original patch internally at a company, without the
> Contributed-under line, then another contributor picking it up and
> posting it according to Contributions.txt, preserving the original
> authorship too). I think it should work but I figured I'd ask you first.
> 
> Thanks
> Laszlo
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel

Re: [edk2] [PATCH] OvmfPkg/build.sh: update gcc detection

[Laszlo Ersek]

On 06/30/16 12:16, Laszlo Ersek wrote:
> On 06/30/16 07:00, Olaf Hering wrote:
>> On Wed, Jun 29, Jordan Justen wrote:
>>
>>> Missing Contributed-under. (See OvmfPkg/Contributions.txt)
>>
>> Looks like this project tries to avoid simple fixes from third party.
> 
> It's not the case. Every project has its contribution rules (I reckon
> xen-devel is no exception). It takes some time to set everything up in
> order to contribute in accordance with the rules, but that's a one time
> cost for every contributor.
> 
> If we encouraged drive-by patches without regard to the process, then it
> would be a constant cost for long term participants / maintainers.
> 
> I do agree with you on two points -- I find these unjustified /
> gratuitous burdens:
> - the Contributed-under line
> - having to subscribe to the mailing list in order to post
> 
> The first is alas a legal requirement, and I can't do anything about it.
> Luckily, it can be automated at least. See for example
> 
> https://github.com/tianocore/tianocore.github.io/wiki/Laszlo's-unkempt-git-guide-for-edk2-contributors-and-maintainers#contrib-07
> 
> The second (the subscription requirement) is terrible. No other open
> source project I'm aware of follows that requirement. I've raised it
> several times, but other subscribers on the list disagree with me.
> 
>> I'm done with it.
> 
> I'm sorry to hear that. I don't really see why you can't post a v2 of
> your patch, with all the remarks addressed, considering you will likely
> carry such a patch in your downstream anyway. I encourage you to send a v2.

I guess you decided not to do it. :(

Jordan: do you think I can pick up Olaf's patch and resubmit it with
your remarks addressed? I'm asking because the original patch did not
have the Contributed-under line, and I wonder if it's okay if I add:

  Signed-off-by: Olaf Hering 
  Fixes: https://github.com/tianocore/edk2/issues/99
  [ler...@redhat.com: update commit message]
  Contributed-under: TianoCore Contribution Agreement 1.0
  Signed-off-by: Laszlo Ersek 

I think we've seen such examples on the list recently (i.e., one author
writing the original patch internally at a company, without the
Contributed-under line, then another contributor picking it up and
posting it according to Contributions.txt, preserving the original
authorship too). I think it should work but I figured I'd ask you first.

Thanks
Laszlo
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel

Re: [edk2] [PATCH] OvmfPkg/build.sh: update gcc detection

[Laszlo Ersek]

On 06/30/16 07:00, Olaf Hering wrote:
> On Wed, Jun 29, Jordan Justen wrote:
> 
>> Missing Contributed-under. (See OvmfPkg/Contributions.txt)
> 
> Looks like this project tries to avoid simple fixes from third party.

It's not the case. Every project has its contribution rules (I reckon
xen-devel is no exception). It takes some time to set everything up in
order to contribute in accordance with the rules, but that's a one time
cost for every contributor.

If we encouraged drive-by patches without regard to the process, then it
would be a constant cost for long term participants / maintainers.

I do agree with you on two points -- I find these unjustified /
gratuitous burdens:
- the Contributed-under line
- having to subscribe to the mailing list in order to post

The first is alas a legal requirement, and I can't do anything about it.
Luckily, it can be automated at least. See for example

https://github.com/tianocore/tianocore.github.io/wiki/Laszlo's-unkempt-git-guide-for-edk2-contributors-and-maintainers#contrib-07

The second (the subscription requirement) is terrible. No other open
source project I'm aware of follows that requirement. I've raised it
several times, but other subscribers on the list disagree with me.

> I'm done with it.

I'm sorry to hear that. I don't really see why you can't post a v2 of
your patch, with all the remarks addressed, considering you will likely
carry such a patch in your downstream anyway. I encourage you to send a v2.

Thanks
Laszlo
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel

Re: [edk2] [PATCH] OvmfPkg/build.sh: update gcc detection

[Jordan Justen]

On 2016-06-29 12:30:35, Olaf Hering wrote:
> Adjust the code to look forward instead of backwards.
> Existing gcc releases are known to work or break. Upcoming gcc releases
> are assumed to work like the last handled gcc version does.
> Doing it that way will reduce the burden to update the script for each
> upcoming gcc variant.
> 
> This fixes issue #99.

Fixes: https://github.com/tianocore/edk2/issues/99

> Signed-off-by: Olaf Hering 

Missing Contributed-under. (See OvmfPkg/Contributions.txt)

> ---
>  OvmfPkg/build.sh | 10 +-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/OvmfPkg/build.sh b/OvmfPkg/build.sh
> index eb5eb73..b40e561 100755
> --- a/OvmfPkg/build.sh
> +++ b/OvmfPkg/build.sh
> @@ -83,6 +83,13 @@ case `uname` in
>Linux*)
>  gcc_version=$(gcc -v 2>&1 | tail -1 | awk '{print $3}')
>  case $gcc_version in
> +  4.[3210].*|3.*|2.*)
> +echo "gcc version $gcc_version is unsupported"
> +exit 1
> +;;
> +  4.4.*)
> +TARGET_TOOLS=GCC44
> +;;
>4.5.*)
>  TARGET_TOOLS=GCC45
>  ;;
> @@ -99,7 +106,8 @@ case `uname` in
>  TARGET_TOOLS=GCC49
>  ;;
>*)
> -TARGET_TOOLS=GCC44
> +# assume the latest, we cant keep up with gcc releases

I think we should not add this comment.

-Jordan

> +TARGET_TOOLS=GCC49
>  ;;
>  esac
>  esac
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel

Re: [edk2] [PATCH] OvmfPkg/build.sh: update gcc detection

[Laszlo Ersek]

On 06/29/16 21:30, Olaf Hering wrote:
> Adjust the code to look forward instead of backwards.
> Existing gcc releases are known to work or break. Upcoming gcc releases
> are assumed to work like the last handled gcc version does.
> Doing it that way will reduce the burden to update the script for each
> upcoming gcc variant.
> 
> This fixes issue #99.

(I think Jordan prefers tags of the form

Fixes: https://github.com/tianocore/edk2/issues/99

because they auto-close the github issues.

This can be updated at commit time.)

> 
> Signed-off-by: Olaf Hering 
> ---
>  OvmfPkg/build.sh | 10 +-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/OvmfPkg/build.sh b/OvmfPkg/build.sh
> index eb5eb73..b40e561 100755
> --- a/OvmfPkg/build.sh
> +++ b/OvmfPkg/build.sh
> @@ -83,6 +83,13 @@ case `uname` in
>Linux*)
>  gcc_version=$(gcc -v 2>&1 | tail -1 | awk '{print $3}')
>  case $gcc_version in
> +  4.[3210].*|3.*|2.*)
> +echo "gcc version $gcc_version is unsupported"
> +exit 1
> +;;
> +  4.4.*)
> +TARGET_TOOLS=GCC44
> +;;
>4.5.*)
>  TARGET_TOOLS=GCC45
>  ;;
> @@ -99,7 +106,8 @@ case `uname` in
>  TARGET_TOOLS=GCC49
>  ;;
>*)
> -TARGET_TOOLS=GCC44
> +# assume the latest, we cant keep up with gcc releases

"can't" -- can be fixed up at commit time

With those addressed (no need for a v2):

Reviewed-by: Laszlo Ersek 

I'll defer to Jordan on this, like I said in #99.

Thanks!
Laszlo

> +TARGET_TOOLS=GCC49
>  ;;
>  esac
>  esac
> 

___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel