Fix uname usage and add binlogs

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.

[hoyosjs]

According to https://en.wikipedia.org/wiki/Uname#Examples as pointed out by @danmosemsft matches most of the values we have used over time. Thanks @mdh1418 for the fix suggestion to use the -m flag.

@jkoritzinsky it's either this, or just return init-os-and-arch to the older fallback tree. Any reason to extend the tree as you did? Any impact on RIDs or other component builds? Specially for 1386?

Fixes#33164
Fixes#33211
Fixes#33210

[jkoritzinsky]

I extended the tree based on @am11’s feedback. I didn’t have any particular reason to extend it.

[am11]

@jkoritzinsky, this OSX build is before your PR was merged and has the same issue: https://dev.azure.com/dnceng/public/_build/results?buildId=514798&view=logs&j=8a742514-b8bd-5f7f-a706-89d0e8c57cc6&t=63254f6c-9351-50b1-ef65-7fd2e7d08b71&l=21. I think init-os-and-arch changes are irrelevant to this issue. Please CMIIW. :)

[jkoritzinsky]

There's a slight difference from before my PR was merged and after. Before it was merged, i386 showed up as an "Unknown arch" so we would default to building for x64. After it was merged, we recognized i386, so we recognized it as an unsupported arch, but would still try to build for osx-x86 (which is expected, some of our partners build us for Linux x86 so we allowed it even though we don't support it).

[am11]

Thanks for the info @jkoritzinsky! I didn't realized this difference.
I also think -m is the right default. As a reference, please take a look at CMake's platform detection logic (which has quite a large array of supported platforms): https://github.com/Kitware/CMake/blob/e484b42/Modules/CMakeDetermineSystem.cmake#L46.

[hoyosjs]

Failures are #33291 which are pre-existing. This doesn't touch any Windows bits.