| Message ID | 20240824151511.3178807-1-dario.binacchi@amarulasolutions.com |
|---|---|
| State | New |
| Headers | show |
| Series |
|
| Related | show |
Hello Dario, On Sat, 24 Aug 2024 17:15:11 +0200 Dario Binacchi <dario.binacchi@amarulasolutions.com> wrote: > - C++11 is now the minimum required C++ standard > - faster handling of compound expressions by trimatu() and trimatl() > - faster sparse matrix addition, subtraction and element-wise multiplication > - expanded sparse submatrix views to handle the non-contiguous form of X.cols(vector_of_column_indices) > - expanded eigs_sym() and eigs_gen() with optional fine-grained parameters > > Fixes: > - http://autobuild.buildroot.org/results/d4e9e7b453960a4f62a199344b30b729a4f235bc > > Signed-off-by: Dario Binacchi <dario.binacchi@amarulasolutions.com> > --- > package/armadillo/armadillo.hash | 2 +- > package/armadillo/armadillo.mk | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) Same comment as for babeld: bumping to solve a download issue is not a good solution. Indeed, if we have to bump because upstream dropped the tarball, what tells us that they won't drop the new tarball again in the future? This can only be resolved either by finding an alternate upstream location or by talking with upstream so that they restore the previous tarball and understand the issues caused by moving/removing previous tarballs. Also, if they made C++11 required, then I'm pretty sure we will need a BR2_TOOLCHAIN_GCC_AT_LEAST_xyz dependency added, no? Thanks! Thomas
Hello Thomas, On Mon, Aug 26, 2024 at 6:57 PM Thomas Petazzoni <thomas.petazzoni@bootlin.com> wrote: > > Hello Dario, > > On Sat, 24 Aug 2024 17:15:11 +0200 > Dario Binacchi <dario.binacchi@amarulasolutions.com> wrote: > > > - C++11 is now the minimum required C++ standard > > - faster handling of compound expressions by trimatu() and trimatl() > > - faster sparse matrix addition, subtraction and element-wise multiplication > > - expanded sparse submatrix views to handle the non-contiguous form of X.cols(vector_of_column_indices) > > - expanded eigs_sym() and eigs_gen() with optional fine-grained parameters > > > > Fixes: > > - http://autobuild.buildroot.org/results/d4e9e7b453960a4f62a199344b30b729a4f235bc > > > > Signed-off-by: Dario Binacchi <dario.binacchi@amarulasolutions.com> > > --- > > package/armadillo/armadillo.hash | 2 +- > > package/armadillo/armadillo.mk | 2 +- > > 2 files changed, 2 insertions(+), 2 deletions(-) > > Same comment as for babeld: bumping to solve a download issue is not a > good solution. Indeed, if we have to bump because upstream dropped the > tarball, what tells us that they won't drop the new tarball again in > the future? This can only be resolved either by finding an alternate > upstream location or by talking with upstream so that they restore the > previous tarball and understand the issues caused by moving/removing > previous tarballs. This is the response from Conrad Sand, one of the two maintainers of the project: Hi Dario, Armadillo 9.900 is very old, horribly outdated and no longer supported. I recommend upgrading to the latest version, which is currently 14.0.2. https://sourceforge.net/projects/arma/files/armadillo-14.0.2.tar.xz If you really want ancient versions of Armadillo, they can be temporarily found here: https://sourceforge.net/projects/arma/files/retired/ > > Also, if they made C++11 required, then I'm pretty sure we will need a > BR2_TOOLCHAIN_GCC_AT_LEAST_xyz dependency added, no? > > Thanks! > > Thomas > -- > Thomas Petazzoni, CTO, Bootlin > Embedded Linux and Kernel engineering > https://bootlin.com
Hello Thomas, Sorry for the previous email, I sent it prematurely before completing it. On Mon, Aug 26, 2024 at 6:57 PM Thomas Petazzoni <thomas.petazzoni@bootlin.com> wrote: > > Hello Dario, > > On Sat, 24 Aug 2024 17:15:11 +0200 > Dario Binacchi <dario.binacchi@amarulasolutions.com> wrote: > > > - C++11 is now the minimum required C++ standard > > - faster handling of compound expressions by trimatu() and trimatl() > > - faster sparse matrix addition, subtraction and element-wise multiplication > > - expanded sparse submatrix views to handle the non-contiguous form of X.cols(vector_of_column_indices) > > - expanded eigs_sym() and eigs_gen() with optional fine-grained parameters > > > > Fixes: > > - http://autobuild.buildroot.org/results/d4e9e7b453960a4f62a199344b30b729a4f235bc > > > > Signed-off-by: Dario Binacchi <dario.binacchi@amarulasolutions.com> > > --- > > package/armadillo/armadillo.hash | 2 +- > > package/armadillo/armadillo.mk | 2 +- > > 2 files changed, 2 insertions(+), 2 deletions(-) > > Same comment as for babeld: bumping to solve a download issue is not a > good solution. Indeed, if we have to bump because upstream dropped the > tarball, what tells us that they won't drop the new tarball again in > the future? This can only be resolved either by finding an alternate > upstream location or by talking with upstream so that they restore the > previous tarball and understand the issues caused by moving/removing > previous tarballs. This is the response from Conrad Sand, one of the two maintainers of the project: --------------------------------------------------------------------------------------------------- Hi Dario, Armadillo 9.900 is very old, horribly outdated and no longer supported. I recommend upgrading to the latest version, which is currently 14.0.2. https://sourceforge.net/projects/arma/files/armadillo-14.0.2.tar.xz If you really want ancient versions of Armadillo, they can be temporarily found here: https://sourceforge.net/projects/arma/files/retired/ --------------------------------------------------------------------------------------------------- So, should I restore the 9.900 version? Or at this point, would it be acceptable to bump to 10.10? Or should I restore 9.900 and then bump to 10.10? Or should I restore 9.900 and then bump to 14.2? If bumping to 10.10 is acceptable right away, I can then take my time to bump to 14.2 later, but in the meantime, the bug gets fixed. Thanks and regards, Dario > > Also, if they made C++11 required, then I'm pretty sure we will need a > BR2_TOOLCHAIN_GCC_AT_LEAST_xyz dependency added, no? > > Thanks! > > Thomas > -- > Thomas Petazzoni, CTO, Bootlin > Embedded Linux and Kernel engineering > https://bootlin.com
Hello Dario, On Tue, 27 Aug 2024 08:20:30 +0200 Dario Binacchi <dario.binacchi@amarulasolutions.com> wrote: > This is the response from Conrad Sand, one of the two maintainers of > the project: > > --------------------------------------------------------------------------------------------------- > Hi Dario, > > Armadillo 9.900 is very old, horribly outdated and no longer supported. Right, but this is not a reason for the tarball to no longer be available, as it breaks build systems horribly. > I recommend upgrading to the latest version, which is currently 14.0.2. > https://sourceforge.net/projects/arma/files/armadillo-14.0.2.tar.xz > > If you really want ancient versions of Armadillo, they can be > temporarily found here: > > https://sourceforge.net/projects/arma/files/retired/ Did you try using those? > --------------------------------------------------------------------------------------------------- > > So, should I restore the 9.900 version? Yes, for master, I would prefer to keep the 9.900 version. > Or at this point, would it be acceptable to bump to 10.10? > Or should I restore 9.900 and then bump to 10.10? > Or should I restore 9.900 and then bump to 14.2? This last choice is the right one: restore 9.900 in master, and bump to 14.0.2 in next, of course if you're interested by the package. Otherwise, we could also decide to drop the package entirely. Samuel Martin is associated to its in the DEVELOPERS file, but Samuel has stopped contributing to Buildroot years ago, and nobody else got interested by that package. Therefore dropping it would be a totally acceptable alternative. Best regards, Thomas
diff --git a/package/armadillo/armadillo.hash b/package/armadillo/armadillo.hash index cc91fb456aa0..d7a17f226f4e 100644 --- a/package/armadillo/armadillo.hash +++ b/package/armadillo/armadillo.hash @@ -1,4 +1,4 @@ # Locally computed: -sha256 d78658c9442addf7f718eb05881150ee3ec25604d06dd3af4942422b3ce26d05 armadillo-9.900.2.tar.xz +sha256 72e3f1b4b4d1b4df70d9cb1e321a254ea04d7843f68a6c34d82691997d558395 armadillo-10.1.0.tar.xz # License files, locally calculated sha256 3ddf9be5c28fe27dad143a5dc76eea25222ad1dd68934a047064e56ed2fa40c5 LICENSE.txt diff --git a/package/armadillo/armadillo.mk b/package/armadillo/armadillo.mk index 8661bfd9da87..b3ae9ef68fa9 100644 --- a/package/armadillo/armadillo.mk +++ b/package/armadillo/armadillo.mk @@ -4,7 +4,7 @@ # ################################################################################ -ARMADILLO_VERSION = 9.900.2 +ARMADILLO_VERSION = 10.1.0 ARMADILLO_SOURCE = armadillo-$(ARMADILLO_VERSION).tar.xz ARMADILLO_SITE = https://downloads.sourceforge.net/project/arma ARMADILLO_INSTALL_STAGING = YES
- C++11 is now the minimum required C++ standard - faster handling of compound expressions by trimatu() and trimatl() - faster sparse matrix addition, subtraction and element-wise multiplication - expanded sparse submatrix views to handle the non-contiguous form of X.cols(vector_of_column_indices) - expanded eigs_sym() and eigs_gen() with optional fine-grained parameters Fixes: - http://autobuild.buildroot.org/results/d4e9e7b453960a4f62a199344b30b729a4f235bc Signed-off-by: Dario Binacchi <dario.binacchi@amarulasolutions.com> --- package/armadillo/armadillo.hash | 2 +- package/armadillo/armadillo.mk | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)