diff mbox series

[v3] ncurses: Fix install conflict when enable multilib.

Message ID 1736986779-18213-1-git-send-email-wangmy@fujitsu.com
State New
Headers show
Series [v3] ncurses: Fix install conflict when enable multilib. | expand

Commit Message

Mingyu Wang (Fujitsu) Jan. 16, 2025, 12:19 a.m. UTC 
From: Wang Mingyu <wangmy@fujitsu.com>

The setting of want_xterm_kbs is as following:
case $host_os in
(*linux-gnu|*cygwin|*mingw32|*msys)
    want_xterm_kbs=DEL
    ;;
(*)
    want_xterm_kbs=BS
    ;;
esac

The host_os when enable multilib is as folloing:
host_os of aarch64 : linux-gnu
host_os of aarch32 : linux-gnueabi

So in lib64 package, want_xterm_kbs=DEL, and in lib32 package, want_xterm_kbs=BS.
It results the differences in the terminfo files between lib32 and lib64 packages.

Using --with-xterm-kbs=del to unify the packages of lib32 and lib64 into "want_xterm_kbs=DEL".

Signed-off-by: Lei Maohui <leimaohui@fujitsu.com>
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
---
 meta/recipes-core/ncurses/ncurses.inc | 1 +
 1 file changed, 1 insertion(+)

Comments

Ross Burton Jan. 21, 2025, 3:30 p.m. UTC | #1 
On 16 Jan 2025, at 00:19, wangmy via lists.openembedded.org <wangmy=fujitsu.com@lists.openembedded.org> wrote:
> 
> From: Wang Mingyu <wangmy@fujitsu.com>
> 
> The setting of want_xterm_kbs is as following:
> case $host_os in
> (*linux-gnu|*cygwin|*mingw32|*msys)
>    want_xterm_kbs=DEL
>    ;;
> (*)
>    want_xterm_kbs=BS
>    ;;
> esac
> 
> The host_os when enable multilib is as folloing:
> host_os of aarch64 : linux-gnu
> host_os of aarch32 : linux-gnueabi

Seems like the proper fix would be to change the glob to *linux-gnu* (and sent that upstream first, obviously)?

Ross
Alexander Kanavin Jan. 21, 2025, 3:39 p.m. UTC | #2 
On Tue, 21 Jan 2025 at 16:31, Ross Burton via lists.openembedded.org
<ross.burton=arm.com@lists.openembedded.org> wrote:
> > The host_os when enable multilib is as folloing:
> > host_os of aarch64 : linux-gnu
> > host_os of aarch32 : linux-gnueabi
>
> Seems like the proper fix would be to change the glob to *linux-gnu* (and sent that upstream first, obviously)?

This is an outcome of my review; in the end of it I had asked to send
that proper fix upstream, but provide the oe-core fix as a tweak to
configure flags. Basically because we don't know if upstream will take
the proper fix.

What would help is to include the link to that upstream submission
into this commit message.

Alex
Mingyu Wang (Fujitsu) Jan. 22, 2025, 12:16 a.m. UTC | #3 
This issue has been raised to the upstream, who stated that in addition to linux-gnueabi, there are many other types that need to be matched, so they have included the modifications in the to-do list.

  --
Best Regards
---------------------------------------------------
Wang Mingyu
FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
No.6 Wenzhu Road, Nanjing, 210012, China
TEL：+86+25-86630566--8568
COINS: 79988548
FAX: +86+25-83317685
MAIL: wangmy@fujitsu.com

> -----Original Message-----
> From: Alexander Kanavin <alex.kanavin@gmail.com>
> Sent: Tuesday, January 21, 2025 11:40 PM
> To: ross.burton@arm.com
> Cc: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com>;
> openembedded-core@lists.openembedded.org; Lei, Maohui/雷 茂慧
> <leimaohui@fujitsu.com>
> Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict when
> enable multilib.
> 
> On Tue, 21 Jan 2025 at 16:31, Ross Burton via lists.openembedded.org
> <ross.burton=arm.com@lists.openembedded.org> wrote:
> > > The host_os when enable multilib is as folloing:
> > > host_os of aarch64 : linux-gnu
> > > host_os of aarch32 : linux-gnueabi
> >
> > Seems like the proper fix would be to change the glob to *linux-gnu* (and sent
> that upstream first, obviously)?
> 
> This is an outcome of my review; in the end of it I had asked to send that proper
> fix upstream, but provide the oe-core fix as a tweak to configure flags. Basically
> because we don't know if upstream will take the proper fix.
> 
> What would help is to include the link to that upstream submission into this
> commit message.
> 
> Alex
Alexander Kanavin Jan. 22, 2025, 8:28 a.m. UTC | #4 
Please remember to include the links to upstream discussions into commit
messages. This always helps when at some point in the future someone would
need to understand the issue, do additional fixing or other follow up work.

If it’s on the mailing list, those typically have web based archives with
stable links to individual messages.

Alex

On Wed 22. Jan 2025 at 1.17, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com>
wrote:

> This issue has been raised to the upstream, who stated that in addition to
> linux-gnueabi, there are many other types that need to be matched, so they
> have included the modifications in the to-do list.
>
>   --
> Best Regards
> ---------------------------------------------------
> Wang Mingyu
> FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
> No.6 Wenzhu Road, Nanjing, 210012, China
> TEL：+86+25-86630566--8568
> COINS: 79988548
> FAX: +86+25-83317685
> MAIL: wangmy@fujitsu.com
>
> > -----Original Message-----
> > From: Alexander Kanavin <alex.kanavin@gmail.com>
> > Sent: Tuesday, January 21, 2025 11:40 PM
> > To: ross.burton@arm.com
> > Cc: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com>;
> > openembedded-core@lists.openembedded.org; Lei, Maohui/雷 茂慧
> > <leimaohui@fujitsu.com>
> > Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict
> when
> > enable multilib.
> >
> > On Tue, 21 Jan 2025 at 16:31, Ross Burton via lists.openembedded.org
> > <ross.burton=arm.com@lists.openembedded.org> wrote:
> > > > The host_os when enable multilib is as folloing:
> > > > host_os of aarch64 : linux-gnu
> > > > host_os of aarch32 : linux-gnueabi
> > >
> > > Seems like the proper fix would be to change the glob to *linux-gnu*
> (and sent
> > that upstream first, obviously)?
> >
> > This is an outcome of my review; in the end of it I had asked to send
> that proper
> > fix upstream, but provide the oe-core fix as a tweak to configure flags.
> Basically
> > because we don't know if upstream will take the proper fix.
> >
> > What would help is to include the link to that upstream submission into
> this
> > commit message.
> >
> > Alex
>
Mingyu Wang (Fujitsu) Jan. 24, 2025, 2:10 a.m. UTC | #5 
The upstream of ncurses has fixed this issue.
But I'm not sure if it has been completely modified.
I can submit an upgrade patch, do I still need to specify --with-xterm-kbs=del in this patch？

--
Best Regards
---------------------------------------------------
Wang Mingyu
FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
No.6 Wenzhu Road, Nanjing, 210012, China
TEL：+86+25-86630566--8568
COINS: 79988548
FAX: +86+25-83317685
MAIL: wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>

From: Alexander Kanavin <alex.kanavin@gmail.com>
Sent: Wednesday, January 22, 2025 4:29 PM
To: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com>
Cc: ross.burton@arm.com; openembedded-core@lists.openembedded.org; Lei, Maohui/雷 茂慧 <leimaohui@fujitsu.com>
Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict when enable multilib.

Please remember to include the links to upstream discussions into commit messages. This always helps when at some point in the future someone would need to understand the issue, do additional fixing or other follow up work.

If it’s on the mailing list, those typically have web based archives with stable links to individual messages.

Alex

On Wed 22. Jan 2025 at 1.17, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>> wrote:
This issue has been raised to the upstream, who stated that in addition to linux-gnueabi, there are many other types that need to be matched, so they have included the modifications in the to-do list.

  --
Best Regards
---------------------------------------------------
Wang Mingyu
FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
No.6 Wenzhu Road, Nanjing, 210012, China
TEL：+86+25-86630566--8568
COINS: 79988548
FAX: +86+25-83317685
MAIL: wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>

> -----Original Message-----
> From: Alexander Kanavin <alex.kanavin@gmail.com<mailto:alex.kanavin@gmail.com>>
> Sent: Tuesday, January 21, 2025 11:40 PM
> To: ross.burton@arm.com<mailto:ross.burton@arm.com>
> Cc: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>>;
> openembedded-core@lists.openembedded.org<mailto:openembedded-core@lists.openembedded.org>; Lei, Maohui/雷 茂慧
> <leimaohui@fujitsu.com<mailto:leimaohui@fujitsu.com>>
> Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict when
> enable multilib.
>
> On Tue, 21 Jan 2025 at 16:31, Ross Burton via lists.openembedded.org<http://lists.openembedded.org>
> <ross.burton=arm.com@lists.openembedded.org<mailto:arm.com@lists.openembedded.org>> wrote:
> > > The host_os when enable multilib is as folloing:
> > > host_os of aarch64 : linux-gnu
> > > host_os of aarch32 : linux-gnueabi
> >
> > Seems like the proper fix would be to change the glob to *linux-gnu* (and sent
> that upstream first, obviously)?
>
> This is an outcome of my review; in the end of it I had asked to send that proper
> fix upstream, but provide the oe-core fix as a tweak to configure flags. Basically
> because we don't know if upstream will take the proper fix.
>
> What would help is to include the link to that upstream submission into this
> commit message.
>
> Alex
Alexander Kanavin Jan. 24, 2025, 10:45 a.m. UTC | #6 
On Fri, 24 Jan 2025 at 03:11, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com> wrote:
>
> The upstream of ncurses has fixed this issue.
>
> But I'm not sure if it has been completely modified.
>
> I can submit an upgrade patch, do I still need to specify --with-xterm-kbs=del in this patch？

Unfortunately the upgrade patch is not acceptable (I wrote a separate
response why). Can you point to the upstream fix that they added to
their repo?

Alex
Mingyu Wang (Fujitsu) Jan. 25, 2025, 11:57 p.m. UTC | #7 
> why). Can you point to the upstream fix that they added to their repo?

The upstream modify the file configure to fix this bug:

@@ -659,7 +659,7 @@ WHICH_XTERM=$with_xterm_new
AC_SUBST(WHICH_XTERM)

 case $host_os in
- (*linux-gnu|*cygwin|*mingw32|*msys)
+(linux*gnu|linux*gnuabi64|linux*gnuabin32|linux*gnueabi|linux*gnueabihf|linux*gnux32|uclinux*|gnu*|mint*|k*bsd*-gnu|cygwin|msys|mingw*|linux*uclibc|linux*musl)
 	 want_xterm_kbs=DEL
 	 ;;
 (*)
‎

  --
Best Regards
---------------------------------------------------
Wang Mingyu
FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
No.6 Wenzhu Road, Nanjing, 210012, China
TEL：+86+25-86630566--8568
COINS: 79988548
FAX: +86+25-83317685
MAIL: wangmy@fujitsu.com

> -----Original Message-----
> From: Alexander Kanavin <alex.kanavin@gmail.com>
> Sent: Friday, January 24, 2025 6:46 PM
> To: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com>
> Cc: ross.burton@arm.com; openembedded-core@lists.openembedded.org; Lei,
> Maohui/雷 茂慧 <leimaohui@fujitsu.com>
> Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict when
> enable multilib.
> 
> On Fri, 24 Jan 2025 at 03:11, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com>
> wrote:
> >
> > The upstream of ncurses has fixed this issue.
> >
> > But I'm not sure if it has been completely modified.
> >
> > I can submit an upgrade patch, do I still need to specify
> > --with-xterm-kbs=del in this patch？
> 
> Unfortunately the upgrade patch is not acceptable (I wrote a separate response
> why). Can you point to the upstream fix that they added to their repo?
> 
> Alex
Alexander Kanavin Jan. 27, 2025, 8:16 a.m. UTC | #8 
On Sun, 26 Jan 2025 at 00:57, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com> wrote:
>
> > why). Can you point to the upstream fix that they added to their repo?
>
> The upstream modify the file configure to fix this bug:
>
> @@ -659,7 +659,7 @@ WHICH_XTERM=$with_xterm_new
> AC_SUBST(WHICH_XTERM)
>
>  case $host_os in
> - (*linux-gnu|*cygwin|*mingw32|*msys)
> +(linux*gnu|linux*gnuabi64|linux*gnuabin32|linux*gnueabi|linux*gnueabihf|linux*gnux32|uclinux*|gnu*|mint*|k*bsd*-gnu|cygwin|msys|mingw*|linux*uclibc|linux*musl)
>          want_xterm_kbs=DEL
>          ;;
>  (*)

The problem is that upstream doesn't publish individual commits, they
just roll everything into 'snapshots', so you need to extract that one
change, and make a commit with it. I'd say let's leave the
--with-xterm-kbs=del, and wait for the next major release, 6.6 or so
when we can drop the setting perhaps.

Alex
Mingyu Wang (Fujitsu) Feb. 7, 2025, 2:03 a.m. UTC | #9 
I discussed this issue directly with the maintainer through email, so he only mentioned it in the changelog and did not leave a record of the discussion on the web.

If necessary, I can forward this discussion to mail list to record the detail.


--
Best Regards
---------------------------------------------------
Wang Mingyu
FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
No.6 Wenzhu Road, Nanjing, 210012, China
TEL：+86+25-86630566--8568
COINS: 79988548
FAX: +86+25-83317685
MAIL: wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>

From: Alexander Kanavin <alex.kanavin@gmail.com>
Sent: Wednesday, January 22, 2025 4:29 PM
To: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com>
Cc: ross.burton@arm.com; openembedded-core@lists.openembedded.org; Lei, Maohui/雷 茂慧 <leimaohui@fujitsu.com>
Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict when enable multilib.

Please remember to include the links to upstream discussions into commit messages. This always helps when at some point in the future someone would need to understand the issue, do additional fixing or other follow up work.

If it’s on the mailing list, those typically have web based archives with stable links to individual messages.

Alex

On Wed 22. Jan 2025 at 1.17, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>> wrote:
This issue has been raised to the upstream, who stated that in addition to linux-gnueabi, there are many other types that need to be matched, so they have included the modifications in the to-do list.

  --
Best Regards
---------------------------------------------------
Wang Mingyu
FUJITSU NANJING SOFTWARE TECHNOLOGY CO., LTD. (FNST)
No.6 Wenzhu Road, Nanjing, 210012, China
TEL：+86+25-86630566--8568
COINS: 79988548
FAX: +86+25-83317685
MAIL: wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>

> -----Original Message-----
> From: Alexander Kanavin <alex.kanavin@gmail.com<mailto:alex.kanavin@gmail.com>>
> Sent: Tuesday, January 21, 2025 11:40 PM
> To: ross.burton@arm.com<mailto:ross.burton@arm.com>
> Cc: Wang, Mingyu/王 鸣瑜 <wangmy@fujitsu.com<mailto:wangmy@fujitsu.com>>;
> openembedded-core@lists.openembedded.org<mailto:openembedded-core@lists.openembedded.org>; Lei, Maohui/雷 茂慧
> <leimaohui@fujitsu.com<mailto:leimaohui@fujitsu.com>>
> Subject: Re: [PATCH] [OE-core] [PATCH v3] ncurses: Fix install conflict when
> enable multilib.
>
> On Tue, 21 Jan 2025 at 16:31, Ross Burton via lists.openembedded.org<http://lists.openembedded.org>
> <ross.burton=arm.com@lists.openembedded.org<mailto:arm.com@lists.openembedded.org>> wrote:
> > > The host_os when enable multilib is as folloing:
> > > host_os of aarch64 : linux-gnu
> > > host_os of aarch32 : linux-gnueabi
> >
> > Seems like the proper fix would be to change the glob to *linux-gnu* (and sent
> that upstream first, obviously)?
>
> This is an outcome of my review; in the end of it I had asked to send that proper
> fix upstream, but provide the oe-core fix as a tweak to configure flags. Basically
> because we don't know if upstream will take the proper fix.
>
> What would help is to include the link to that upstream submission into this
> commit message.
>
> Alex
Alexander Kanavin Feb. 7, 2025, 8:42 a.m. UTC | #10 
On Fri, 7 Feb 2025 at 03:03, Mingyu Wang (Fujitsu) <wangmy@fujitsu.com> wrote:
> I discussed this issue directly with the maintainer through email, so he only mentioned it in the changelog and did not leave a record of the discussion on the web.

That's ok; generally it's preferred to hold such discussions through a
public channel. If it was via private email, please mention that
explicitly in the future.

Alex
diff mbox series

Patch

diff --git a/meta/recipes-core/ncurses/ncurses.inc b/meta/recipes-core/ncurses/ncurses.inc
index e96ee075bd..3f93550cf9 100644
--- a/meta/recipes-core/ncurses/ncurses.inc
+++ b/meta/recipes-core/ncurses/ncurses.inc
@@ -70,6 +70,7 @@  ncurses_configure() {
 	        --without-debug \
 	        --without-ada \
 	        --without-gpm \
+		--with-xterm-kbs=del \
 	        --enable-hard-tabs \
 	        --enable-xmc-glitch \
 	        --enable-colorfgbg \