Why the first is kept
The budget is read back by the figure that caps a fill before the fill is counted, so the totals standing against a command were measured by whichever budget was there at the time. The recording revises no total it finds already counted: it either writes a budget or keeps one. So the figure those totals were measured by is the figure that stays.
What is offered and what is heard
A budget naming neither a size nor a spend is not recorded at all. One offered where none is held is written under the command's name.
A second one differing from the first is reported only where the caller asked for conflicts to be reported. Of the four roads that offer a budget, one asks: the road installing a command. The replay of a scheme's start, the replay of a submitted intent, and the budget of a closing order all offer theirs silently.
So a replay that offers a second budget differing from the first keeps the first without a word, and the record of the disagreement exists nowhere.
Related
- The part of a fill a command counts
- What a queue reservation is for
- How a restart reads its record back
Source evidence
---
row: P165-02
baseline: 49aa659
created: 2026-11-09 12:39 UTC
modified: 2026-11-09 13:28 UTC
evidence:
- claim: "A budget naming neither a size nor a spend is not recorded at all"
source: "`execution_manager.py` 2383, then 2384"
- claim: "Where a budget is already held for the command, the one offered is dropped and the first kept"
source: "`execution_manager.py` 2386, then 2388, then 2407"
- claim: "A second budget differing from the first is reported only where the caller asked for conflicts to be reported"
source: "`execution_manager.py` 2389-2391"
- claim: "Otherwise the budget offered is written under the command's name"
source: "`execution_manager.py` 2409"
- claim: "It is read back by the figure that caps a fill before the fill is counted"
source: "`execution_manager.py` 2477, read from 2428"
- claim: "Nothing in the recording revises the totals already counted"
source: "`execution_manager.py` 2386, then 2407 against 2409"
- claim: "One road of the four asks for conflicts, the one installing a command"
source: "`execution_manager.py` 2345, then 2346, then 2350, against 1129, and 1158, and 4927"
---