GN Mac/iOS target platform detection

[Marshall Greenblatt]

Hi All,

I'm attempting a desktop build of Chromium on Mac and GN is implicitly (and incorrectly) setting is_ios=true. I think there may be something wrong or missing in my Xcode.app (version 7.3.1). Can anyone point me to the code in GN that performs this target platform detection?

Thanks,
Marshall

[Nico Weber]

I don't think anything does that automatically. What's your args.gn?

--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
    http://groups.google.com/a/chromium.org/group/chromium-dev

[Marshall Greenblatt]

The relevant part is:

is_debug=false

is_official_build=true

target_cpu="x64"

[Nico Weber]

You're not setting target_os? Why are you sure is_ios is true? Did you put a `print(is_ios)` somewhere?

[Marshall Greenblatt]

On Fri, Aug 26, 2016 at 11:24 PM, Nico Weber <tha...@chromium.org> wrote:

You're not setting target_os?

Correct.

 

Why are you sure is_ios is true? Did you put a `print(is_ios)` somewhere?

Because I'm getting these build erors: http://pastebin.com/KczKpFTM

This is the part that tells me is_ios=true:

See //build/toolchain/mac/BUILD.gn:11:1: whence it was imported.
import("//build/config/ios/ios_sdk.gni")

Also, I can't reproduce the problem locally with an Xcode.app bundle that I downloaded and installed in the usual way via App Store.

[Dirk Pranke]

I would scatter some print() statements in //build/config/BUILDCONFIG.gn; that's the only place where is_ios should be getting set to true. I'd be surprised if that had something to do w/ your xcode install.

-- Dirk

[Marshall Greenblatt]

On Fri, Aug 26, 2016 at 11:59 PM, Dirk Pranke <dpr...@chromium.org> wrote:

I would scatter some print() statements in //build/config/BUILDCONFIG.gn; that's the only place where is_ios should be getting set to true. I'd be surprised if that had something to do w/ your xcode install.

Thanks for the suggestion. After debugging I see that the is_ios flag is not actually true in my case. The confusion resulted from M53, unlike master, including //build/config/ios/ios_sdk.gni without checking the is_ios flag [1] (and I was looking at the master sources instead of the M53 sources).

There also appears to be some non-deterministic behavior (or perhaps GN is intentionally running things in parallel) which added to the confusion. Sometimes `gn gen` fails with:

ERROR at //build/config/mac/mac_sdk.gni:42:5: Script returned non-zero exit code.

    exec_script("//build/mac/find_sdk.py", find_sdk_args, "list lines")

    ^----------

Current dir: /Users/buildagent/chromium/src/out/Debug_GN_x64/

Command: python -- /Users/buildagent/chromium/src/build/mac/find_sdk.py --print_sdk_path 10.10

Returned 1.

And other times it fails with:

ERROR at //build/config/ios/ios_sdk.gni:38:7: Script returned non-zero exit code.

      exec_script("//build/config/mac/sdk_info.py", [ ios_sdk_name ], "scope")

      ^----------

Current dir: /Users/buildagent/chromium/src/out/Debug_GN_x64/

Command: python -- /Users/buildagent/chromium/src/build/config/mac/sdk_info.py iphonesimulator

Returned 1.

In any case, I think this strongly indicates a problem with my Xcode.app bundle. I'll re-create it and hopefully that will resolve the problem.

[1] https://chromium.googlesource.com/chromium/src/+/53.0.2785.86/build/toolchain/mac/BUILD.gn

[Nico Weber]

On Mon, Aug 29, 2016 at 5:48 AM, Marshall Greenblatt <magree...@gmail.com> wrote:

On Fri, Aug 26, 2016 at 11:59 PM, Dirk Pranke <dpr...@chromium.org> wrote:

I would scatter some print() statements in //build/config/BUILDCONFIG.gn; that's the only place where is_ios should be getting set to true. I'd be surprised if that had something to do w/ your xcode install.

Thanks for the suggestion. After debugging I see that the is_ios flag is not actually true in my case. The confusion resulted from M53, unlike master, including //build/config/ios/ios_sdk.gni without checking the is_ios flag [1] (and I was looking at the master sources instead of the M53 sources).

There also appears to be some non-deterministic behavior (or perhaps GN is intentionally running things in parallel) which added to the confusion. Sometimes `gn gen` fails with:

ERROR at //build/config/mac/mac_sdk.gni:42:5: Script returned non-zero exit code.

    exec_script("//build/mac/find_sdk.py", find_sdk_args, "list lines")

    ^----------

Current dir: /Users/buildagent/chromium/src/out/Debug_GN_x64/

Command: python -- /Users/buildagent/chromium/src/build/mac/find_sdk.py --print_sdk_path 10.10

You can manually run `build/mac/find_sdk.py --print_sdk_path 10.10` to see what's going wrong. find_sdk.py is a pretty simple script and just calls xcrun two ways.

[Brett Wilson]

On Mon, Aug 29, 2016 at 2:48 AM, Marshall Greenblatt <magree...@gmail.com> wrote:

On Fri, Aug 26, 2016 at 11:59 PM, Dirk Pranke <dpr...@chromium.org> wrote:

I would scatter some print() statements in //build/config/BUILDCONFIG.gn; that's the only place where is_ios should be getting set to true. I'd be surprised if that had something to do w/ your xcode install.

Thanks for the suggestion. After debugging I see that the is_ios flag is not actually true in my case. The confusion resulted from M53, unlike master, including //build/config/ios/ios_sdk.gni without checking the is_ios flag [1] (and I was looking at the master sources instead of the M53 sources).

There also appears to be some non-deterministic behavior (or perhaps GN is intentionally running things in parallel) which added to the confusion. Sometimes `gn gen` fails with:

GN is extremely parallel so if there are two errors, I would expect that they might come in a different order on different runs.

Brett