Let edm::Wrapper<T> use the constructor T{kUninitialized}

[fwyzard]

PR description:

Let edm::Wrapper<T> use a special constructor T{kUninitialized} instead of the default constructor T{}, if the latter does not exist.

Add to DeviceProduct<T> an explicit constructor DeviceProduct{kUninitialized} that calls T{kUninitialized}, if the latter is valid.

Remove the default constructor from PortableCollection, PortableObject and related types, and introduce the special constructor taking kUninitialized.

PR validation:

Unit tests pass, and the full 2024 HLT runs.

[cmsbuild]

cms-bot internal usage

[fwyzard]

please test with #46876, #46875

[cmsbuild]

-code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46877/42905

• There are other open Pull requests which might conflict with changes you have proposed:
  • File DataFormats/Common/interface/DeviceProduct.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/Portable/interface/PortableDeviceCollection.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/Portable/interface/PortableHostCollection.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/Portable/interface/PortableHostObject.h modified in PR(s): Add {Copy,Move}ToDeviceCache<T> class templates and moveToDeviceAsync function template #43969
  • File DataFormats/SiPixelDigiSoA/interface/SiPixelDigiErrorsDevice.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/SiPixelDigiSoA/interface/SiPixelDigisDevice.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844

Code check has found code style and quality issues which could be resolved by applying following patch(s)

• code-format:
  https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46877/42905/code-format.patch
  e.g. curl -k https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46877/42905/code-format.patch | patch -p1
  You can also run scram build code-format to apply code format directly

[cmsbuild]

-code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46877/42907

• There are other open Pull requests which might conflict with changes you have proposed:
  • File DataFormats/Common/interface/DeviceProduct.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/Portable/interface/PortableDeviceCollection.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/Portable/interface/PortableHostCollection.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/Portable/interface/PortableHostObject.h modified in PR(s): Add {Copy,Move}ToDeviceCache<T> class templates and moveToDeviceAsync function template #43969
  • File DataFormats/SiPixelDigiSoA/interface/SiPixelDigiErrorsDevice.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844
  • File DataFormats/SiPixelDigiSoA/interface/SiPixelDigisDevice.h modified in PR(s): [RFC] Add tests and rudimentary protections for default-constructed PortableCollections #44844

Code check has found code style and quality issues which could be resolved by applying following patch(s)

• code-format:
  https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46877/42907/code-format.patch
  e.g. curl -k https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46877/42907/code-format.patch | patch -p1
  You can also run scram build code-format to apply code format directly

[fwyzard]

you pushed 3 commits few minutes ago and 4c8434c is touching non-core files

The commits touch those packages, but the comparison does not:

DataFormats/TestObjects/src/classes_def.xml

@@ -5,6 +5,7 @@
 <class name="edmtest::IntProduct" ClassVersion="10">
  <version ClassVersion="10" checksum="4286676158"/>
 </class>
+<class name="std::pair<edmtest::IntProduct, edm::RandomNumberGeneratorState>"/>
 <class name="edmtest::MaybeUninitializedIntProduct" ClassVersion="3">
  <version ClassVersion="3" checksum="3324182225"/>
 </class>

I though that the bot would consider the actual comparison, not the commits, because we want people to be able to replace the commits with a squashed version without requiring to re-sign the PRs ?

[smuzaffar]

I though that the bot would consider the actual comparison

No bot uses the commits and signatures are assigned to commits.

[fwyzard]

Ah, I see.

Could we consider to use the comparison instead of the commits as a future improvement ?

[smuzaffar]

I do not think it is possible specially when one force push changes and ref to old commits are gone

[cmsbuild]

+1

Size: This PR adds an extra 24KB to repository
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-2be352/43357/summary.html
COMMIT: 4c8434c
CMSSW: CMSSW_15_0_X_2024-12-10-1100/el8_amd64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/46877/43357/install.sh to create a dev area with all the needed externals and cmssw changes.

• DAS Queries: The DAS query tests failed, see the summary page for details.

Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 2 differences found in the comparisons
• DQMHistoTests: Total files compared: 47
• DQMHistoTests: Total histograms compared: 3623082
• DQMHistoTests: Total failures: 473
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3622589
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 46 files compared)
• Checked 206 log files, 177 edm output root files, 47 DQM output files
• TriggerResults: found differences in 1 / 45 workflows

[makortel]

I do not think it is possible specially when one force push changes and ref to old commits are gone

Darn. Although #43845 lists

• Not reset L2's signatures as long as the changes proposed by a force update are identical to changes before the force push. This includes

  • squashing multiple commits (c1, c2, c3, ..., cn) in to one (or more) new commit(s) (cx)
  • force pushing a different set of commits which have overall identical changes

So did that not work in the end?

[makortel]

+core

[smuzaffar]

So did that not work in the end?

@iarspider can check. I think squashing commits is working. I am not sure about adding new set of commits with identical changes though

[iarspider]

The changes were not identical: before and after squash.

[smuzaffar]

lets move this discussion to #43845

[fwyzard]

The changes were not identical: before and after squash.

As I wrote above, the changes between the two set of commits touch only the DataFormats/TestObjects package:

DataFormats/TestObjects/src/classes_def.xml

@@ -5,6 +5,7 @@
 <class name="edmtest::IntProduct" ClassVersion="10">
  <version ClassVersion="10" checksum="4286676158"/>
 </class>
+<class name="std::pair<edmtest::IntProduct, edm::RandomNumberGeneratorState>"/>
 <class name="edmtest::MaybeUninitializedIntProduct" ClassVersion="3">
  <version ClassVersion="3" checksum="3324182225"/>
 </class>

So I expected the bot to request a new signature only from the core category.

[jfernan2]

+1

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @rappoccio, @mandrenguyen, @antoniovilela, @sextonkennedy (and backports should be raised in the release meeting by the corresponding L2)

[mandrenguyen]

+1