[swift-lldb-dev] Failed to build OS X toolchain package broken due to commit b08610b

Todd Fiala todd.fiala at gmail.com
Fri Dec 4 00:56:32 CST 2015

Okay I think I see the issue.  It is that we can still pass backdoor
combinations of build modes after the '--' in build-script that will
control the final build variant used by lldb.

The logic for this check doesn't belong in build-script, but rather
build-script-impl.  build-script-impl is the only place to find the final
build mode after considering both the values passed to build-script, and
other bits that can be passed along to build-script-impl.

I'm going to revert my change and add a proper fix in the morning.  I don't
have enough brain power to do it right and test it now.

On Thu, Dec 3, 2015 at 10:51 PM, Todd Fiala <todd.fiala at gmail.com> wrote:

> Yep I finally added enough gunk to the command line to reproduce.
> I'm looking at how it's getting the Release value now.  That's the issue
> (I'm not seeing it come in as Release).
> On Thu, Dec 3, 2015 at 10:44 PM, Dmitri Gribenko <gribozavr at gmail.com>
> wrote:
>> On Thu, Dec 3, 2015 at 10:41 PM, Todd Fiala <todd.fiala at gmail.com> wrote:
>> > I see.
>> >
>> > And I think I see the issue.  It's trying to do a release-debuginfo
>> build
>> > across the board.  LLDB chokes on those.  So we'll need to explicitly
>> set
>> > the lldb build variant to something different (either release or debug,
>> > likely release).
>> The preset is trying to do that (lldb-build-type=Release), but
>> apparently is getting ignored.
>> Dmitri
>> --
>> main(i,j){for(i=2;;i++){for(j=2;j<i;j++){if(!(i%j)){j=0;break;}}if
>> (j){printf("%d\n",i);}}} /*Dmitri Gribenko <gribozavr at gmail.com>*/
> --
> -Todd

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-lldb-dev/attachments/20151203/d7350c31/attachment-0001.html>

More information about the swift-lldb-dev mailing list