Infer Action Dimensions From Restart File Information

[bska]

PRs #2651 and #2663 added a Runspec argument to the restart loader as a means of inferring the maximum number of actions, the number of conditions per action and other essential pieces of information needed to correctly interpret the *ACT and *ACN restart file arrays. While this practice has served us well in the interim, it's strictly speaking somewhat fragile. A restarted simulation run is allowed to increase these dimensions compared to the base run and in that case we'd be basing our offsets on out-of-bounds values.

This commit switches to loading the base run's action dimensions directly from the pertinent INTEHEAD array items which, in turn, enables removing the Runspec argument to RstState::load(). We still need the run's number of tabulated PVT regions in order to correctly load the oil vaporisation properties, but I want to push the scattering to the client side at some point in the future. For now we add a new overload of RstState::load() function that takes the number of PVT regions as an optional<int>–defaulting to nullopt which we treat as a single PVT region–and make the overload taking the Runspec argument call the other overload. The Runspec overload will be marked "deprecated" once all downstream users have been converted to the new overload.

The one compromise that we have to add here is a floating-point based derivation of each actions "start" time. We would previously infer this from the Runspec argument's START value, but this is no longer available. Instead we use a calculation based on the RstHeader's sim_time() and the SimTim element of DOUBHEAD. Other derivations may be possible, but this appears sufficiently accurate for now.

[bska]

jenkins build this please

[bska]

jenkins build this please

[Copilot]

Pull request overview

This PR makes restart loading more robust by deriving ACTIONX-related dimensions (and some other restart metadata) directly from restart header arrays instead of relying on Runspec, and it introduces a new RstState::load() overload that only needs an optional PVT-table count.

Changes:

• Update restart header/state loading to infer ACTIONX dimensions and start-time information from INTEHEAD/DOUBHEAD, enabling removal of the Runspec dependency in the primary RstState::load() path.
• Extend INTEHEAD/LOGIHEAD vector item definitions and header parsing to include ACTIONX sizing fields and a temperature flag.
• Adjust restart-related tests and minor related refactors/cleanup (including initialization and formatting tweaks).

Reviewed changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/test_rst.cpp	Update RstHeader construction to no longer require Runspec.
tests/test_InteHEAD.cpp	Update RstHeader construction in tests to match new API.
tests/test_AggregateWellData.cpp	Update RstHeader construction in tests to match new API.
opm/output/eclipse/VectorItems/logihead.hpp	Add explicit LOGIHEAD indices for temperature flags.
opm/output/eclipse/VectorItems/intehead.hpp	Add explicit INTEHEAD indices for ACTIONX sizing fields.
opm/output/eclipse/LogiHEAD.cpp	Use named LOGIHEAD indices instead of numeric literals for TEMP flags.
opm/output/eclipse/InteHEAD.hpp	Remove variousUDQ_ACTIONXParam() from public interface.
opm/output/eclipse/InteHEAD.cpp	Fold ACTIONX-related INTEHEAD sizing values into actionParam() and simplify MAXSPRLINE calculation.
opm/output/eclipse/CreateInteHead.cpp	Stop calling removed variousUDQ_ACTIONXParam().
opm/io/eclipse/rst/well.cpp	Switch tracer/temperature handling to be restart-header driven and modernize extraction logic.
opm/io/eclipse/rst/state.hpp	Add new RstState::load() overload and remove Runspec dependency from action restoration helpers.
opm/io/eclipse/rst/state.cpp	Implement new load() overload and refactor ACTIONX restore logic to use header-derived sizing.
opm/io/eclipse/rst/header.hpp	Remove stored Runspec from RstHeader; add ACTIONX sizing + inferred start-time fields.
opm/io/eclipse/rst/header.cpp	Populate new header fields, including inferred start times from DOUBHEAD.
opm/input/eclipse/Schedule/OilVaporizationProperties.hpp	Minor formatting and default-initialization cleanup for members.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 15 out of 15 changed files in this pull request and generated 2 comments.

[bska]

jenkins build this please

[Copilot]

Pull request overview

Copilot reviewed 17 out of 17 changed files in this pull request and generated 2 comments.

Suppressed comments (1)

opm/io/eclipse/rst/state.cpp:784

• Action/condition array dimensions are now taken from restart header fields (action_*_size, max_action_conditions) and immediately used to compute subspan() sizes and indexed in RstAction::Condition::valid() / RstAction construction. If any of these header values are 0/negative or inconsistent with the expected condition sizes, this can lead to UB/out-of-bounds reads before any exception is thrown (e.g. Condition::valid() indexes iacn[...] without a size check). Add sanity checks (>= expected minimums and consistent with VectorItems::{IACN,SACN,ZACN}::ConditionSize) and fail fast with a clear error if the restart header is malformed/unsupported.

    const auto iacn_cond_size = static_cast<std::size_t>(this->header.action_iacn_size);
    const auto sacn_cond_size = static_cast<std::size_t>(this->header.action_sacn_size);
    const auto zacn_cond_size = static_cast<std::size_t>(this->header.action_zacn_size);

    const auto iacn_action_size = this->header.max_action_conditions * iacn_cond_size;
    const auto sacn_action_size = this->header.max_action_conditions * sacn_cond_size;
    const auto zacn_action_size = this->header.max_action_conditions * zacn_cond_size;

[Copilot]

Pull request overview

Copilot reviewed 17 out of 17 changed files in this pull request and generated no new comments.

[bska]

Copilot reviewed 17 out of 17 changed files in this pull request and generated no new comments.

Very good. In that case, I'll mark this PR as "ready for review" by a human reviewer.

[bska]

jenkins build this please

[akva2]

LGTM

[bska]

jenkins build this please

[bska]

PR approved and build check is green. I'll merge into master.