Bump F# to 4.5 #10919

15:49:58 mkdir: @/external/bockbuild/scratch/fsharp.install@/external/bockbuild/stage/lib/mono/fsharp: Permission denied
15:49:58 
make[1]: *** [install-sdk-lib] Error 1
15:49:58 make: *** [install] Error 2
15:49:58 
15:49:58 fsharp->build_sh:

As far as I can tell, the permission error has been caused by an inadvertent sudo make install from a previous build. I think the bots need manual clean up.

akoeplinger · 2018-10-18T18:29:16Z

Ok, I cleaned the bots, let's try another build.

akoeplinger · 2018-10-18T18:30:05Z

@monojenkins build pkg

nosami · 2018-10-19T06:38:04Z

Same error after clean up. I have no idea where the 'Permission denied' is coming from.

nosami · 2018-10-19T06:42:14Z

I am assuming that there is some issue with the Jenkins bots somewhere. The same script builds on Travis.

I don't see sudo being used anywhere apart from sudo make install (which I patched out - It appears that the bots run make install to a custom folder). It's also used for restoring packages when the script is ran on linux.

dsyme · 2018-10-19T09:40:38Z

Who can add me to those orgs? I should be in xamarin I thought

Can we get direct access to one of these build machines?

akoeplinger · 2018-10-19T11:55:49Z

@dsyme yeah you are in the org but your membership is set to private (so the bot can't see it). Go to https://github.com/orgs/xamarin/people?utf8=%E2%9C%93&query=dsyme and flip it to "public"

akoeplinger · 2018-10-19T11:55:58Z

@nosami yes, let's talk on Slack

alexischr · 2018-10-19T17:46:35Z

It seemed like more uncleanable files from sudo make install

alexischr · 2018-10-19T17:46:43Z

@monojenkins build pkg

alexischr · 2019-04-10T14:25:52Z

@monojenkins build pkg

alexischr · 2019-04-10T15:37:56Z

@directhex given our understanding of the problem now, do you think it's still necessary that both all and install are run on the build step? I may fix that up in a subsequent PR.

directhex · 2019-04-10T15:58:41Z

@alexischr my strong preference is that they NOT be run on the same build step, tbh

alexischr · 2019-04-10T16:33:19Z

It built!! @directhex sounds good, I am not even going to try here though. Merging!

dsyme · 2019-04-10T19:00:35Z

🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉 🎉

🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾 🍾

dsyme · 2019-04-10T19:01:13Z

Thank you so, so much to everyone who helped with this little piece of nightmare.

auduchinok · 2019-04-10T19:03:12Z

BTW, should we talk about bumping it to 4.6 already? :)

dsyme · 2019-04-10T19:03:51Z

@auduchinok First we drink all that champagne I just posted

alexischr · 2019-04-10T19:06:37Z

@monojenkins backport to 2019-02

alexischr · 2019-04-10T19:07:04Z

@monojenkins backport to 2018-10

monojenkins · 2019-04-10T19:07:36Z

@alexischr backporting to 2019-02 failed, the patch results in conflicts:

Applying: attempt F# update
Applying: Revert "attempt F# update"
Applying: revert and integrate
Using index info to reconstruct a base tree...
M	packaging/MacSDK/fsharp.py
.git/rebase-apply/patch:96: space before tab in indent.
 	@echo "Installing $(ASSEMBLY)"
.git/rebase-apply/patch:102: space before tab in indent.
 	@mkdir -p $(DESTDIR)$(monodir)/fsharp
.git/rebase-apply/patch:104: space before tab in indent.
 	@if test "x$(DELAY_SIGN)" = "x1"; then \
.git/rebase-apply/patch:105: space before tab in indent.
 	    echo "Signing $(outdir)$(ASSEMBLY) with Mono key"; \
.git/rebase-apply/patch:124: trailing whitespace.
 
warning: squelched 1 whitespace error
warning: 6 lines add whitespace errors.
Falling back to patching base and 3-way merge...
Removing packaging/MacSDK/patches/fsharp-string-switchName.patch
Removing packaging/MacSDK/patches/fsharp-path-overloads.patch
Removing packaging/MacSDK/patches/fsharp-debug-pinvoke-fix.patch
Auto-merging packaging/MacSDK/fsharp.py
CONFLICT (content): Merge conflict in packaging/MacSDK/fsharp.py
error: Failed to merge in the changes.
Patch failed at 0003 revert and integrate

Please backport manually!

monojenkins · 2019-04-10T19:08:05Z

@alexischr backporting to 2018-10 failed, the patch results in conflicts:

Applying: attempt F# update
Using index info to reconstruct a base tree...
M	packaging/MacSDK/fsharp.py
.git/rebase-apply/patch:96: space before tab in indent.
 	@echo "Installing $(ASSEMBLY)"
.git/rebase-apply/patch:102: space before tab in indent.
 	@mkdir -p $(DESTDIR)$(monodir)/fsharp
.git/rebase-apply/patch:104: space before tab in indent.
 	@if test "x$(DELAY_SIGN)" = "x1"; then \
.git/rebase-apply/patch:105: space before tab in indent.
 	    echo "Signing $(outdir)$(ASSEMBLY) with Mono key"; \
.git/rebase-apply/patch:124: trailing whitespace.
 
warning: squelched 1 whitespace error
warning: 6 lines add whitespace errors.
Falling back to patching base and 3-way merge...
Removing packaging/MacSDK/patches/fsharp-string-switchName.patch
Removing packaging/MacSDK/patches/fsharp-path-overloads.patch
Removing packaging/MacSDK/patches/fsharp-debug-pinvoke-fix.patch
Auto-merging packaging/MacSDK/fsharp.py
CONFLICT (content): Merge conflict in packaging/MacSDK/fsharp.py
error: Failed to merge in the changes.
Patch failed at 0001 attempt F# update

Please backport manually!

alexischr · 2019-04-10T19:08:24Z

I'll do the backporting manually. And yeah we should get started on 4.6 right away :)

auduchinok · 2019-04-14T19:26:34Z

@alexischr Thank you again!
And just to get some insight on Mono release process: what Mono versions this change should be expected to get to and when to expect that to happen?

* Use a custom prefix instead of trying to infer it, which doesn't work under bockbuild. The following line is problematic in the scope of building the Mac SDK package (using bockbuild): https://github.com/fsharp/fsharp/blob/ccfccbacfdb16b2ab143cf83f5ece20f738fd864/mono/config.make#L5 It's problematic because where `mono` lives in this case would be something like `/macsdk-staging-dir/bin/mono` which is not the actual prefix (prefix here being the final destination, `/Library/Frameworks/Mono.framework/...`) but a staging area that bockbuild uses (this is how we build on system prefixes without installing on CI) For the F# build, bockbuild expects things to end up in `/fsharp-staging-dir/$(prefix)` (because that's what most `make install`s do when given `DESTDIR=/fsharp-staging-dir/`). If things end up in _any other_ directory in `/fsharp-staging-dir/` (like `/fsharp-staging-dir/macsdk-staging-dir` in our case) that would cause us to miss packaging those files (this has happened before, with packages that have broken or different use of `DESTDIR`). So there is an added write-protection for anywhere in `/fsharp-staging-dir` that *isn't* `/fsharp-staging-dir/$(prefix)` , because a write-protection error, while obtuse, is much preferable to missing the fact that some files are not packaged (as would have happened here) So the fix is to stop trying to infer the prefix, and return to providing it explicitly to the build system (this used to be the case via the `configure` step) * Also, if we are doing an override of the 'make' command, we need to take care of passing DESTDIR ourselves too * Since now we perform 'make install' on the make step, make the 'install' step into a no-operation. * Look for sn where mono is, not in the final destination (which again, may not be the same)

alexischr · 2019-04-15T17:02:55Z

@monojenkins backport to 2019-04

monojenkins · 2019-04-15T17:03:58Z

@alexischr backporting to 2019-04 failed, the patch results in conflicts:

Applying: attempt F# update
Using index info to reconstruct a base tree...
M	packaging/MacSDK/fsharp.py
.git/rebase-apply/patch:96: space before tab in indent.
 	@echo "Installing $(ASSEMBLY)"
.git/rebase-apply/patch:102: space before tab in indent.
 	@mkdir -p $(DESTDIR)$(monodir)/fsharp
.git/rebase-apply/patch:104: space before tab in indent.
 	@if test "x$(DELAY_SIGN)" = "x1"; then \
.git/rebase-apply/patch:105: space before tab in indent.
 	    echo "Signing $(outdir)$(ASSEMBLY) with Mono key"; \
.git/rebase-apply/patch:124: trailing whitespace.
 
warning: squelched 1 whitespace error
warning: 6 lines add whitespace errors.
Falling back to patching base and 3-way merge...
Removing packaging/MacSDK/patches/fsharp-string-switchName.patch
Removing packaging/MacSDK/patches/fsharp-path-overloads.patch
Removing packaging/MacSDK/patches/fsharp-debug-pinvoke-fix.patch
Auto-merging packaging/MacSDK/fsharp.py
CONFLICT (content): Merge conflict in packaging/MacSDK/fsharp.py
error: Failed to merge in the changes.
Patch failed at 0001 attempt F# update

Please backport manually!

alexischr · 2019-04-15T17:23:13Z

@auduchinok I am attempting backports from master to 6.0 and 6.2 (see PRs referring to this one). Otherwise this will make the next bi-monthly branch (6.4).

These are unreleased/unstable Mono versions and have not finished being integrated with Xamarin products, but 6.0 is on preview as a standalone Mac package: https://www.mono-project.com/download/preview/ . If this merges on 6.0, we can update the 6.0 preview to include it. Do note - this package is not guaranteed or even expected to work with any releases of Visual Studio for Mac.

As we were not targeting this for a specific Mono/Xamarin release (and this is a big change), I did not end up backporting to stable branches.

* Use a custom prefix instead of trying to infer it, which doesn't work under bockbuild. The following line is problematic in the scope of building the Mac SDK package (using bockbuild): https://github.com/fsharp/fsharp/blob/ccfccbacfdb16b2ab143cf83f5ece20f738fd864/mono/config.make#L5 It's problematic because where `mono` lives in this case would be something like `/macsdk-staging-dir/bin/mono` which is not the actual prefix (prefix here being the final destination, `/Library/Frameworks/Mono.framework/...`) but a staging area that bockbuild uses (this is how we build on system prefixes without installing on CI) For the F# build, bockbuild expects things to end up in `/fsharp-staging-dir/$(prefix)` (because that's what most `make install`s do when given `DESTDIR=/fsharp-staging-dir/`). If things end up in _any other_ directory in `/fsharp-staging-dir/` (like `/fsharp-staging-dir/macsdk-staging-dir` in our case) that would cause us to miss packaging those files (this has happened before, with packages that have broken or different use of `DESTDIR`). So there is an added write-protection for anywhere in `/fsharp-staging-dir` that *isn't* `/fsharp-staging-dir/$(prefix)` , because a write-protection error, while obtuse, is much preferable to missing the fact that some files are not packaged (as would have happened here) So the fix is to stop trying to infer the prefix, and return to providing it explicitly to the build system (this used to be the case via the `configure` step) * Also, if we are doing an override of the 'make' command, we need to take care of passing DESTDIR ourselves too * Since now we perform 'make install' on the make step, make the 'install' step into a no-operation. * Look for sn where mono is, not in the final destination (which again, may not be the same)

auduchinok · 2019-04-16T15:53:17Z

Thanks. I was naively expecting we could get it in a 5.x release. 🙂

* Use a custom prefix instead of trying to infer it, which doesn't work under bockbuild. The following line is problematic in the scope of building the Mac SDK package (using bockbuild): https://github.com/fsharp/fsharp/blob/ccfccbacfdb16b2ab143cf83f5ece20f738fd864/mono/config.make#L5 It's problematic because where `mono` lives in this case would be something like `/macsdk-staging-dir/bin/mono` which is not the actual prefix (prefix here being the final destination, `/Library/Frameworks/Mono.framework/...`) but a staging area that bockbuild uses (this is how we build on system prefixes without installing on CI) For the F# build, bockbuild expects things to end up in `/fsharp-staging-dir/$(prefix)` (because that's what most `make install`s do when given `DESTDIR=/fsharp-staging-dir/`). If things end up in _any other_ directory in `/fsharp-staging-dir/` (like `/fsharp-staging-dir/macsdk-staging-dir` in our case) that would cause us to miss packaging those files (this has happened before, with packages that have broken or different use of `DESTDIR`). So there is an added write-protection for anywhere in `/fsharp-staging-dir` that *isn't* `/fsharp-staging-dir/$(prefix)` , because a write-protection error, while obtuse, is much preferable to missing the fact that some files are not packaged (as would have happened here) So the fix is to stop trying to infer the prefix, and return to providing it explicitly to the build system (this used to be the case via the `configure` step) * Also, if we are doing an override of the 'make' command, we need to take care of passing DESTDIR ourselves too * Since now we perform 'make install' on the make step, make the 'install' step into a no-operation. * Look for sn where mono is, not in the final destination (which again, may not be the same)

directhex · 2019-04-17T13:54:26Z

RPMs are done. Debs are blocked by #14103

nosami requested review from akoeplinger, alexischr, directhex and luhenry as code owners October 2, 2018 08:28

akoeplinger approved these changes Oct 2, 2018

View reviewed changes

akoeplinger reviewed Oct 2, 2018

View reviewed changes

alexischr merged commit c16b4cf into mono:master Apr 10, 2019

nosami deleted the fsharp-4.5 branch April 11, 2019 13:07

alexischr mentioned this pull request Apr 15, 2019

[2019-02][MacSDK] Bump F# to 4.5 (#10919) #14054

Merged

alexischr mentioned this pull request Apr 15, 2019

[2019-04][MacSDK] Bump F# to 4.5 (#10919) #14058

Merged

marek-safar added this to the 2019-02 (6.0.xx) milestone Apr 23, 2019

auduchinok mentioned this pull request Sep 10, 2019

Bump F# to 4.7 #16763

Closed

Bump F# to 4.5 #10919

Bump F# to 4.5 #10919

Uh oh!

Conversation

nosami commented Oct 2, 2018

Uh oh!

akoeplinger commented Oct 2, 2018

Uh oh!

akoeplinger Oct 2, 2018

Choose a reason for hiding this comment

Uh oh!

nosami Oct 2, 2018

Choose a reason for hiding this comment

Uh oh!

akoeplinger commented Oct 2, 2018

Uh oh!

akoeplinger commented Oct 3, 2018

Uh oh!

nosami commented Oct 3, 2018

Uh oh!

nosami commented Oct 4, 2018

Uh oh!

nosami commented Oct 5, 2018

Uh oh!

nosami commented Oct 7, 2018

Uh oh!

nosami commented Oct 8, 2018

Uh oh!

nosami commented Oct 8, 2018

Uh oh!

alexischr commented Oct 9, 2018

Uh oh!

dsyme commented Oct 18, 2018

Uh oh!

dsyme commented Oct 18, 2018

Uh oh!

dsyme commented Oct 18, 2018

Uh oh!

akoeplinger commented Oct 18, 2018

Uh oh!

nosami commented Oct 18, 2018

Uh oh!

akoeplinger commented Oct 18, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

akoeplinger commented Oct 18, 2018

Uh oh!

nosami commented Oct 19, 2018

Uh oh!

nosami commented Oct 19, 2018

Uh oh!

dsyme commented Oct 19, 2018

Uh oh!

akoeplinger commented Oct 19, 2018

Uh oh!

akoeplinger commented Oct 19, 2018

Uh oh!

alexischr commented Oct 19, 2018

Uh oh!

alexischr commented Oct 19, 2018

Uh oh!

alexischr commented Apr 10, 2019

Uh oh!

alexischr commented Apr 10, 2019

Uh oh!

directhex commented Apr 10, 2019

Uh oh!

alexischr commented Apr 10, 2019

Uh oh!

dsyme commented Apr 10, 2019

Uh oh!

dsyme commented Apr 10, 2019

Uh oh!

auduchinok commented Apr 10, 2019

Uh oh!

dsyme commented Apr 10, 2019

Uh oh!

alexischr commented Apr 10, 2019

Uh oh!

alexischr commented Apr 10, 2019

akoeplinger commented Oct 18, 2018 •

edited

Loading