From: Gleb Mazovetskiy <glex.spb@gmail.com>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH] relocate-sdk: Allow relocating to any location
Date: Sat, 17 Jul 2021 11:33:02 +0100 [thread overview]
Message-ID: <CADjhFRBgVC8veGTAoZFn1wuxdQNOi8MV1sr7LzRkX=8KOuMDcQ@mail.gmail.com> (raw)
In-Reply-To: <20210717102603.GQ12203@scaer>
> The problem with that, is that users are not free to have the SDK where
they want.
The users are still free to have the SDK wherever they want!
They just need to run relocate-sdk.sh if using the non-default location,
exactly as before.
This approach also does not leak the SDK maintainer's build directory path.
It's basically just a sensible default, instead of no default.
> I don't think this is a good way to use the SDK... I think the relocation
step should really be run by the person extracting and using the SDK, on
their machine.
The vast majority of our users install a single version into the default
location. Many of them are beginners.
> I'm still leaving this patch pending, because I still feel the burn from
last time [...]
I don't know the history behind this but thanks for having a look at these,
and I hope the other SDK maintainers will chip in here.
On Sat, Jul 17, 2021 at 11:26 AM Yann E. MORIN <yann.morin.1998@free.fr>
wrote:
> Gleb, All,
>
> Thanks for the feedback.
>
> On 2021-07-17 11:08 +0100, Gleb Mazovetskiy spake thusly:
> > This fixes a minor frustrating issue. It addresses the following
> scenario:
> > 1. I'm the SDK maintainer. I tell users to unpack to a specific
> directory. E.g. /opt/mysdk.
> > 2. I build the SDK in such a way that the users do not need to run
> relocate-sdk.sh if they unpack to /opt/mysdk.
> > The problem is with step 2: currently it requires the build process to
> move the SDK to /opt/mysdk, then run relocate-sdk.sh, before
> > producing the final tarball.
> > This is often problematic because on the SDK maintainer's machine,
> `/opt/mysdk` usually already exists and contains the latest
> > stable release.
> > With this new argument, the build process can be simplified for the SDK
> maintainer.
>
> The problem with that, is that users are not free to have the SDK where
> they want.
>
> That would prevents them from having multiple versions of the SDK
> installed at once (e.g. when they have to work on a maintenance branch
> to fix bugs in an older releae for example), without removing an SDK
> before they install a new one.
>
> And as a consequence, that also prevents the person responsible for the
> SDK, to work on older SDK to pinpoint issues without removing the
> "latest" SDK first...
>
> I don't think this is a good way to use the SDK... I think the
> relocation step should really be run by the person extracting and using
> the SDK, on their machine.
>
> I'm still leaving this patch pending, because I still feel the burn from
> last time, so if another maintainer is interested in this, they will
> have to apply it (and expand the commit log to carry the explanations
> you provided in this reply, thanks!).
>
> Regards,
> Yann E. MORIN.
>
> > On Sat, Jul 17, 2021 at 10:39 AM Yann E. MORIN < [1]
> yann.morin.1998 at free.fr> wrote:
> >
> > Gleb, All,
> >
> > On 2020-03-28 23:49 +0000, Gleb Mazovetskiy spake thusly:
> > > Previously, the script could only relocate to the current location.
> >
> > I am not sure to understand what you mean...
> >
> > The SDK can already be relocated to any location:
> >
> > $ cd /path/to/anywhere
> > $ tar xf /path/to/sdk.tar
> > $ ./relocate-sdk
> >
> > > This caused issues for SDK maintainers, as the default target
> location
> > > (e.g. /opt/my-sdk) would need to be created and the content copied
> > > to it during the build process, which is error-prone.
> > >
> > > With this change you can now pass a target location argument to
> > > ./relocate-sdk.sh (this does not move the directory).
> >
> > I fail to see the purpose for that... If the SDK is not there, what is
> > the point of relocating to a location where the SDK is not?
> >
> > Eventually, the user will have to either extract the SDK in the proper
> > location, or move it there, so whether they are able to run the script
> > prior to the move does not seem very interesting...
> >
> > So, I am far from convinced by this change... :-(
> >
> > Additionally, last time I had to touch on the SDK topic, I got
> seriously
> > burnt, so I'd like that people that have a better grasp, the history,
> and
> > some beliefs, in the SDK, look at that patch, please...
> >
> > Regards,
> > Yann E. MORIN.
> >
> > > ---
> > > support/misc/relocate-sdk.sh | 19 ++++++++++++-------
> > > 1 file changed, 12 insertions(+), 7 deletions(-)
> > >
> > > diff --git a/support/misc/relocate-sdk.sh
> b/support/misc/relocate-sdk.sh
> > > index caabeaa6f6..981d272425 100755
> > > --- a/support/misc/relocate-sdk.sh
> > > +++ b/support/misc/relocate-sdk.sh
> > > @@ -1,15 +1,20 @@
> > > #!/bin/sh
> > > -#
> > > -if [ "$#" -ne 0 ]; then
> > > - echo "Run this script to relocate the buildroot SDK at that
> location"
> > > +
> > > +if [ "$#" -gt 1 ]; then
> > > + echo "Usage: $0 [path]"
> > > + echo "Run this script to relocate the buildroot SDK to the
> current location"
> > > + echo "If [path] is given, sets the location to [path] (without
> moving it)"
> > > exit 1
> > > fi
> > >
> > > -LOCFILE="share/buildroot/sdk-location"
> > > -FILEPATH="$(readlink -f "$0")"
> > > -NEWPATH="$(dirname "${FILEPATH}")"
> > > +cd "$(dirname "$(readlink -f "$0")")"
> > > +if [ "$#" -eq 1 ]; then
> > > + NEWPATH="$1"
> > > +else
> > > + NEWPATH="${PWD}"
> > > +fi
> > >
> > > -cd "${NEWPATH}"
> > > +LOCFILE="share/buildroot/sdk-location"
> > > if [ ! -r "${LOCFILE}" ]; then
> > > echo "Previous location of the buildroot SDK not found!"
> > > exit 1
> > > --
> > > 2.20.1
> > >
> > > _______________________________________________
> > > buildroot mailing list
> > > [2]buildroot at busybox.net
> > > [3]http://lists.busybox.net/mailman/listinfo/buildroot
> >
> > --
> >
> .-----------------.--------------------.------------------.--------------------.
> > | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics'
> conspiracy: |
> > | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___
>
> > |
> > | +33 561 099 427 `------------.-------: X AGAINST | \e/
> There is no |
> > | [4]http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v
> conspiracy. |
> >
> '------------------------------^-------^------------------^--------------------'
> >
> > Links:
> > 1. mailto:yann.morin.1998 at free.fr
> > 2. mailto:buildroot at busybox.net
> > 3. http://lists.busybox.net/mailman/listinfo/buildroot
> > 4. http://ymorin.is-a-geek.org/
>
> --
>
> .-----------------.--------------------.------------------.--------------------.
> | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics'
> conspiracy: |
> | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___
> |
> | +33 561 099 427 `------------.-------: X AGAINST | \e/ There is
> no |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v
> conspiracy. |
>
> '------------------------------^-------^------------------^--------------------'
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20210717/b9820fcc/attachment-0001.html>
WARNING: multiple messages have this Message-ID (diff)
From: Gleb Mazovetskiy <glex.spb@gmail.com>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH] relocate-sdk: Allow relocating to any location
Date: Sat, 17 Jul 2021 11:33:02 +0100 [thread overview]
Message-ID: <CADjhFRBgVC8veGTAoZFn1wuxdQNOi8MV1sr7LzRkX=8KOuMDcQ@mail.gmail.com> (raw)
Message-ID: <20210717103302.w6kUqju_IvXFAfCYnZKa_5frK3r93zaUTgxvgKtcYQY@z> (raw)
In-Reply-To: <20210717102603.GQ12203@scaer>
> The problem with that, is that users are not free to have the SDK where
they want.
The users are still free to have the SDK wherever they want!
They just need to run relocate-sdk.sh if using the non-default location,
exactly as before.
This approach also does not leak the SDK maintainer's build directory path.
It's basically just a sensible default, instead of no default.
> I don't think this is a good way to use the SDK... I think the relocation
step should really be run by the person extracting and using the SDK, on
their machine.
The vast majority of our users install a single version into the default
location. Many of them are beginners.
> I'm still leaving this patch pending, because I still feel the burn from
last time [...]
I don't know the history behind this but thanks for having a look at these,
and I hope the other SDK maintainers will chip in here.
On Sat, Jul 17, 2021 at 11:26 AM Yann E. MORIN <yann.morin.1998@free.fr>
wrote:
> Gleb, All,
>
> Thanks for the feedback.
>
> On 2021-07-17 11:08 +0100, Gleb Mazovetskiy spake thusly:
> > This fixes a minor frustrating issue. It addresses the following
> scenario:
> > 1. I'm the SDK maintainer. I tell users to unpack to a specific
> directory. E.g. /opt/mysdk.
> > 2. I build the SDK in such a way that the users do not need to run
> relocate-sdk.sh if they unpack to /opt/mysdk.
> > The problem is with step 2: currently it requires the build process to
> move the SDK to /opt/mysdk, then run relocate-sdk.sh, before
> > producing the final tarball.
> > This is often problematic because on the SDK maintainer's machine,
> `/opt/mysdk` usually already exists and contains the latest
> > stable release.
> > With this new argument, the build process can be simplified for the SDK
> maintainer.
>
> The problem with that, is that users are not free to have the SDK where
> they want.
>
> That would prevents them from having multiple versions of the SDK
> installed at once (e.g. when they have to work on a maintenance branch
> to fix bugs in an older releae for example), without removing an SDK
> before they install a new one.
>
> And as a consequence, that also prevents the person responsible for the
> SDK, to work on older SDK to pinpoint issues without removing the
> "latest" SDK first...
>
> I don't think this is a good way to use the SDK... I think the
> relocation step should really be run by the person extracting and using
> the SDK, on their machine.
>
> I'm still leaving this patch pending, because I still feel the burn from
> last time, so if another maintainer is interested in this, they will
> have to apply it (and expand the commit log to carry the explanations
> you provided in this reply, thanks!).
>
> Regards,
> Yann E. MORIN.
>
> > On Sat, Jul 17, 2021 at 10:39 AM Yann E. MORIN < [1]
> yann.morin.1998 at free.fr> wrote:
> >
> > Gleb, All,
> >
> > On 2020-03-28 23:49 +0000, Gleb Mazovetskiy spake thusly:
> > > Previously, the script could only relocate to the current location.
> >
> > I am not sure to understand what you mean...
> >
> > The SDK can already be relocated to any location:
> >
> > $ cd /path/to/anywhere
> > $ tar xf /path/to/sdk.tar
> > $ ./relocate-sdk
> >
> > > This caused issues for SDK maintainers, as the default target
> location
> > > (e.g. /opt/my-sdk) would need to be created and the content copied
> > > to it during the build process, which is error-prone.
> > >
> > > With this change you can now pass a target location argument to
> > > ./relocate-sdk.sh (this does not move the directory).
> >
> > I fail to see the purpose for that... If the SDK is not there, what is
> > the point of relocating to a location where the SDK is not?
> >
> > Eventually, the user will have to either extract the SDK in the proper
> > location, or move it there, so whether they are able to run the script
> > prior to the move does not seem very interesting...
> >
> > So, I am far from convinced by this change... :-(
> >
> > Additionally, last time I had to touch on the SDK topic, I got
> seriously
> > burnt, so I'd like that people that have a better grasp, the history,
> and
> > some beliefs, in the SDK, look at that patch, please...
> >
> > Regards,
> > Yann E. MORIN.
> >
> > > ---
> > > support/misc/relocate-sdk.sh | 19 ++++++++++++-------
> > > 1 file changed, 12 insertions(+), 7 deletions(-)
> > >
> > > diff --git a/support/misc/relocate-sdk.sh
> b/support/misc/relocate-sdk.sh
> > > index caabeaa6f6..981d272425 100755
> > > --- a/support/misc/relocate-sdk.sh
> > > +++ b/support/misc/relocate-sdk.sh
> > > @@ -1,15 +1,20 @@
> > > #!/bin/sh
> > > -#
> > > -if [ "$#" -ne 0 ]; then
> > > - echo "Run this script to relocate the buildroot SDK at that
> location"
> > > +
> > > +if [ "$#" -gt 1 ]; then
> > > + echo "Usage: $0 [path]"
> > > + echo "Run this script to relocate the buildroot SDK to the
> current location"
> > > + echo "If [path] is given, sets the location to [path] (without
> moving it)"
> > > exit 1
> > > fi
> > >
> > > -LOCFILE="share/buildroot/sdk-location"
> > > -FILEPATH="$(readlink -f "$0")"
> > > -NEWPATH="$(dirname "${FILEPATH}")"
> > > +cd "$(dirname "$(readlink -f "$0")")"
> > > +if [ "$#" -eq 1 ]; then
> > > + NEWPATH="$1"
> > > +else
> > > + NEWPATH="${PWD}"
> > > +fi
> > >
> > > -cd "${NEWPATH}"
> > > +LOCFILE="share/buildroot/sdk-location"
> > > if [ ! -r "${LOCFILE}" ]; then
> > > echo "Previous location of the buildroot SDK not found!"
> > > exit 1
> > > --
> > > 2.20.1
> > >
> > > _______________________________________________
> > > buildroot mailing list
> > > [2]buildroot at busybox.net
> > > [3]http://lists.busybox.net/mailman/listinfo/buildroot
> >
> > --
> >
> .-----------------.--------------------.------------------.--------------------.
> > | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics'
> conspiracy: |
> > | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___
>
> > |
> > | +33 561 099 427 `------------.-------: X AGAINST | \e/
> There is no |
> > | [4]http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v
> conspiracy. |
> >
> '------------------------------^-------^------------------^--------------------'
> >
> > Links:
> > 1. mailto:yann.morin.1998 at free.fr
> > 2. mailto:buildroot at busybox.net
> > 3. http://lists.busybox.net/mailman/listinfo/buildroot
> > 4. http://ymorin.is-a-geek.org/
>
> --
>
> .-----------------.--------------------.------------------.--------------------.
> | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics'
> conspiracy: |
> | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___
> |
> | +33 561 099 427 `------------.-------: X AGAINST | \e/ There is
> no |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v
> conspiracy. |
>
> '------------------------------^-------^------------------^--------------------'
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20210717/b9820fcc/attachment-0002.html>
next prev parent reply other threads:[~2021-07-17 10:33 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-28 23:49 [Buildroot] [PATCH] relocate-sdk: Allow relocating to any location Gleb Mazovetskiy
2021-07-17 9:39 ` Yann E. MORIN
2021-07-17 10:08 ` Gleb Mazovetskiy
2021-07-17 10:08 ` Gleb Mazovetskiy
2021-07-17 10:26 ` Yann E. MORIN
2021-07-17 10:33 ` Gleb Mazovetskiy [this message]
2021-07-17 10:33 ` Gleb Mazovetskiy
2022-01-06 21:28 ` Arnout Vandecappelle
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CADjhFRBgVC8veGTAoZFn1wuxdQNOi8MV1sr7LzRkX=8KOuMDcQ@mail.gmail.com' \
--to=glex.spb@gmail.com \
--cc=buildroot@busybox.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.