depends: Use CC_FOR_BUILD for config.guess

[hebasto]

On the master branch @ 3c88eac, consider the following commands in the depends subdirectory:

$ make print-build HOST=i686-pc-linux-gnu CC="clang -m32"
build=x86_64-pc-linux-gnu
$ make print-host HOST=i686-pc-linux-gnu CC="clang -m32"
host=i686-pc-linux-gnu

The printed variable values are expected.

However, switching the CC variable context from Makefile to the shell environment breaks expectations:

$ CC="clang -m32" make print-build HOST=i686-pc-linux-gnu
build=i686-pc-linux-gnu
$ CC="clang -m32" make print-host HOST=i686-pc-linux-gnu
host=i686-pc-linux-gnu

This PR fixes this issue.

[DrahtBot]

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

Code Coverage

For detailed information about the code coverage, see the test coverage report.

Reviews

See the guideline for information on the review process.
A summary of reviews will appear here.

[theuni]

Where/why is this an issue?

[hebasto]

Where/why is this an issue?

I faced this issue during my work on the CMake staging branch when the CC environment variable was defined by the CI -- https://github.com/hebasto/bitcoin/actions/runs/8822538616/job/24220973382.

The log shows no cross-compiling, but it is cross-compiling to i686-pc-linux-gnu.

[theuni]

Surely this isn't the only place env vars like this would cause issues? Feels like a cat-and-mouse game.

[maflcko]

The log shows no cross-compiling, but it is cross-compiling to i686-pc-linux-gnu.

Can you link to the exact lines in the log that show "no cross-compiling".

[hebasto]

The log shows no cross-compiling, but it is cross-compiling to i686-pc-linux-gnu.

Can you link to the exact lines in the log that show "no cross-compiling".

https://github.com/hebasto/bitcoin/actions/runs/8822538616/job/24220973382#step:14:245:

Cross compiling ....................... FALSE

Please note that that was a PR branch from the CMake migration project. That branch detects cross-compiling basing on host and build values when building with depends.

[fanquake]

Surely this isn't the only place env vars like this would cause issues? Feels like a cat-and-mouse game.

Yea, as-is this doesn't seem like a great fix, and may just break other things?

A better diff might be something like:

diff --git a/depends/Makefile b/depends/Makefile
index 005d9696fb..091511758d 100644
--- a/depends/Makefile
+++ b/depends/Makefile
@@ -51,7 +51,7 @@ FALLBACK_DOWNLOAD_PATH ?= https://bitcoincore.org/depends-sources
 C_STANDARD ?= c11
 CXX_STANDARD ?= c++20
 
-BUILD = $(shell ./config.guess)
+BUILD = $(shell CC_FOR_BUILD=$CC ./config.guess)
 HOST ?= $(BUILD)
 PATCHES_PATH = $(BASEDIR)/patches
 BASEDIR = $(CURDIR)

which would at least be using the option that is meant to be used for this.

[hebasto]

A better diff might be something like:
...
which would at least be using the option that is meant to be used for this.

I agree. Implemented.

Thanks!

[theuni]

However, switching the CC variable context from Makefile to the shell environment breaks expectations

Arguably that's because this wasn't intended to be supported :)

I suppose this fix is reasonable, though supporting env vars like this seems quite brittle.