bpo-29243: Fix Makefile with respect to --enable-optimizations

[torsava]

When using the Profile Guided Optimization (./configure --enable-optimizations)
Python is built not only during make but rebuilt again during make test and
make install. This patch fixes the issue.

Note that this fix produces no change at all in the Makefile if configure is
run witout --enable-optimizations.

[vstinner]

I still see "all" in testall, testuniversal, quicktest, smelly, patchcheck. Please update these ones as well.

But thank you for working on this very annoying bug!

[torsava]

I still see "all" in testall, testuniversal, quicktest, smelly, patchcheck. Please update these ones as well.

Ah, you're right, thanks for the help! I've amended the commit accordingly.

[vstinner]

I ran the following test and I only see Python compiled twice, as expected:

• build with generate profile
• run the profile task
• build with the collected data

Command:

./configure --with-pydebug --prefix=/home/haypo/pgo --enable-optimizations
vim Makefile   # write "PROFILE_TASK=-c pass", just to get a quick test
make
make install
make quicktest

"all" target is no more used directly in Makefile

[vstinner]

Can you please document the change in the Build section in Misc/NEWS? With a NEWS entry, the change will be ready to be merged.

[torsava]

Amended the commit with a NEWS entry!

[louisom]

@torsava you don't need to rebase or squash the commit, at the end of the merge it will do it so :)

[vstinner]

Oh, I'm sorry, I missed the "First-time contributor" notice on GitHub. Would you mind to add your name in Misc/ACKS as well?

[torsava]

@torsava you don't need to rebase or squash the commit, at the end of the merge it will do it so :)

Ah, thank you! I wasn't sure what the merging policy was here, and I dislike simple changes drawn out over multiple commits. (I now realise it's addressed in the contributor guide, sorry for missing that.)

Oh, I'm sorry, I missed the "First-time contributor" notice on GitHub. Would you mind to add your name in Misc/ACKS as well?

Done, thanks!

[louisom]

@torsava yes, I did this before, I also like to squash to make it clean. But that is not good for others to reviewing on GitHub when doing squashing.

And there is a discussion (I dont' remember on python-dev or python-committers) about this, the final answer is: don't squash and leave to the end, you can see here for more information: devguide-pullrequest

[vstinner]

Thank you for your contrib Tomáš Orsava! This bug annoyed me, I opened a duplicate issue: http://bugs.python.org/issue29641

@torsava: Do you want to do the backports? https://docs.python.org/devguide/committing.html#backporting-changes-to-an-older-version

[torsava]

Thank you for your contrib Tomáš Orsava! This bug annoyed me, I opened a duplicate issue: http://bugs.python.org/issue29641

Glad to have helped!

@torsava: Do you want to do the backports?

I'm having some trouble with cherry_picker and I need to be going now. However, if it can wait a few days, I would like to do the backports!

[vstinner]

I'm having some trouble with cherry_picker and I need to be going now. However, if it can wait a few days, I would like to do the backports!

Take your time, there is no urgency ;-) You can ping me directly, we are colleagues ;-)

[gpshead]

thank you! :)