mbox series

[0/4] Fix library testsuite compilation for build sysroot

Message ID alpine.LFD.2.21.1911081714120.13542@redsun52.ssa.fujisawa.hgst.com
Headers show
Series Fix library testsuite compilation for build sysroot | expand

Message

Maciej W. Rozycki Nov. 11, 2019, 6:11 p.m. UTC
Hi,

 This patch series addresses a problem with the testsuite compiler being 
set up across libatomic, libffi, libgo, libgomp with no correlation 
whatsoever to the target compiler being used in GCC compilation.  
Consequently there in no arrangement made to set up the compilation 
sysroot according to the build sysroot specified for GCC compilation, 
causing a catastrophic failure across the testsuites affected from the 
inability to link executables.

 The fix is based on a similar arrangement already made for passing 
autoconf output variables in libgomp, and uses the GCC_UNDER_TEST (or 
GOC_UNDER_TEST, as applicable) TCL variable already used across several 
GCC library testsuites.

 Verified with a cross-compiler configured for the `riscv-linux-gnu' 
target and the `x86_64-linux-gnu' host and using RISC-V/Linux QEMU in the 
user emulation mode as the target board.  Also no change in results with 
`x86_64-linux-gnu' native regression testing.

 See individual change descriptions for details.

 OK to apply to the GCC repo (for libraries maintained externally I'll be 
happy to assist with any merging required, although given that these 
changes are confined to autoconf/automake scriptery they should be 
straightforward to apply, barring any conflicts in generated files)?

  Maciej

Comments

Ulderico Cirello Nov. 11, 2019, 6:15 p.m. UTC | #1
Hi Maciej,


Go's project doesn't take mail patches for changes. Please use gerrit (
https://go-review.googlesource.com/ ).


Thanks





        - ccf


On Mon, Nov 11, 2019 at 10:12 AM Maciej W. Rozycki <macro@wdc.com> wrote:

> Hi,
>
>  This patch series addresses a problem with the testsuite compiler being
> set up across libatomic, libffi, libgo, libgomp with no correlation
> whatsoever to the target compiler being used in GCC compilation.
> Consequently there in no arrangement made to set up the compilation
> sysroot according to the build sysroot specified for GCC compilation,
> causing a catastrophic failure across the testsuites affected from the
> inability to link executables.
>
>  The fix is based on a similar arrangement already made for passing
> autoconf output variables in libgomp, and uses the GCC_UNDER_TEST (or
> GOC_UNDER_TEST, as applicable) TCL variable already used across several
> GCC library testsuites.
>
>  Verified with a cross-compiler configured for the `riscv-linux-gnu'
> target and the `x86_64-linux-gnu' host and using RISC-V/Linux QEMU in the
> user emulation mode as the target board.  Also no change in results with
> `x86_64-linux-gnu' native regression testing.
>
>  See individual change descriptions for details.
>
>  OK to apply to the GCC repo (for libraries maintained externally I'll be
> happy to assist with any merging required, although given that these
> changes are confined to autoconf/automake scriptery they should be
> straightforward to apply, barring any conflicts in generated files)?
>
>   Maciej
>
> --
> You received this message because you are subscribed to the Google Groups
> "golang-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to golang-dev+unsubscribe@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/golang-dev/alpine.LFD.2.21.1911081714120.13542%40redsun52.ssa.fujisawa.hgst.com
> .
>
Kaz Kylheku (libffi) Nov. 11, 2019, 6:23 p.m. UTC | #2
On 2019-11-11 10:15, Ulderico Cirello wrote:
> Hi Maciej,
> 
> Go's project doesn't take mail patches for changes.

Is it that they'd have to read man pages and learn how to use common 
utilities?

Or that nobody has written a "patch in Go" yet?
Ian Lance Taylor Nov. 11, 2019, 6:33 p.m. UTC | #3
On Mon, Nov 11, 2019 at 10:15 AM Ulderico Cirello
<uldericofilho@gmail.com> wrote:
>
> Go's project doesn't take mail patches for changes. Please use gerrit ( https://go-review.googlesource.com/ ).

These patches are for gccgo, not the gc toolchain.  They should
probably have been sent to gofrontend-dev rather than golang-dev.  The
gccgo repo does take patches via e-mail; I route them through Gerrit
as needed.

Ian
Ian Lance Taylor Nov. 11, 2019, 6:34 p.m. UTC | #4
On Mon, Nov 11, 2019 at 10:31 AM Kaz Kylheku (libffi)
<382-725-6798@kylheku.com> wrote:
>
> On 2019-11-11 10:15, Ulderico Cirello wrote:
> > Hi Maciej,
> >
> > Go's project doesn't take mail patches for changes.
>
> Is it that they'd have to read man pages and learn how to use common
> utilities?
>
> Or that nobody has written a "patch in Go" yet?

Please be polite; thanks.  There are many advantages to using Gerrit
for code review.  It has nothing to do with reading man pages or using
the patch program.

Ian
Maciej W. Rozycki Nov. 11, 2019, 6:42 p.m. UTC | #5
On Mon, 11 Nov 2019, Ulderico Cirello wrote:

> Go's project doesn't take mail patches for changes. Please use gerrit (
> https://go-review.googlesource.com/ ).

 Thanks for your reply; this is however too much effort for my limited 
resources and a one-off change.

 The reason is I'm not actively working on Go and I have only enabled Go 
frontend compilation/verification for my RISC-V effort in case there is a 
regression caused by a machine backend change that happens to only trigger 
for the Go frontend so that it does not go unnoticed.

 I have provided this change in a hope it is useful to the community and 
in these circumstances hopefully someone actually interested in Go will 
pick up and merge this change; otherwise I will drop my local change and 
consequently Go verification once GCC 10 has been released.

 Thank you for your understanding.

  Maciej
Maciej W. Rozycki Nov. 11, 2019, 6:44 p.m. UTC | #6
On Mon, 11 Nov 2019, Ian Lance Taylor wrote:

> > Go's project doesn't take mail patches for changes. Please use gerrit ( https://go-review.googlesource.com/ ).
> 
> These patches are for gccgo, not the gc toolchain.  They should
> probably have been sent to gofrontend-dev rather than golang-dev.  The
> gccgo repo does take patches via e-mail; I route them through Gerrit
> as needed.

 I may have misinterpreted this paragraph[1]:

"Submitting Changes

   Changes to the Go frontend should follow the same process as for the
   main Go repository, only for the gofrontend project and the
   gofrontend-dev@googlegroups.com mailing list rather than the go project
   and the golang-dev@googlegroups.com mailing list. Those changes will
   then be merged into the GCC sources."

Sorry about that; I think it might benefit from a rewrite for clarity 
though.

References:

[1] "Contributing to the gccgo frontend - The Go Programming Language", 
    <https://golang.org/doc/gccgo_contribute.html>

  Maciej
Ian Lance Taylor Nov. 11, 2019, 11:16 p.m. UTC | #7
On Mon, Nov 11, 2019 at 10:44 AM Maciej W. Rozycki <macro@wdc.com> wrote:
>
> On Mon, 11 Nov 2019, Ian Lance Taylor wrote:
>
> > > Go's project doesn't take mail patches for changes. Please use gerrit ( https://go-review.googlesource.com/ ).
> >
> > These patches are for gccgo, not the gc toolchain.  They should
> > probably have been sent to gofrontend-dev rather than golang-dev.  The
> > gccgo repo does take patches via e-mail; I route them through Gerrit
> > as needed.
>
>  I may have misinterpreted this paragraph[1]:
>
> "Submitting Changes
>
>    Changes to the Go frontend should follow the same process as for the
>    main Go repository, only for the gofrontend project and the
>    gofrontend-dev@googlegroups.com mailing list rather than the go project
>    and the golang-dev@googlegroups.com mailing list. Those changes will
>    then be merged into the GCC sources."
>
> Sorry about that; I think it might benefit from a rewrite for clarity
> though.
>
> References:
>
> [1] "Contributing to the gccgo frontend - The Go Programming Language",
>     <https://golang.org/doc/gccgo_contribute.html>


The paragraph seems reasonable clear to me, so I'm obviously missing
something.  Can you suggest a clearer rewrite?  Thanks.

Ian