refactor: table-driven DIMM format parsing with tests (#654)

* refactor: table-driven DIMM format parsing with tests

Replace the cascading if-else in getDIMMParseInfo and switch in
getDIMMSocketSlot with a unified dimmFormats table. Each entry holds
pre-compiled regexes, match predicates, and extraction functions.

- Fix regex bugs: [\d+] matched single char, not multi-digit numbers
- Add descriptive dimmType names (dimmTypeInspurICX, dimmTypeQuantaGNR, etc.)
- Add 66 characterization tests covering all 17 formats, ordering
  edge cases, and multi-digit regression cases

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* formatting

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* chore: use descriptive dimmType names in tests, remove old aliases

Update all test references from dimmType0-16 to descriptive names
(dimmTypeInspurICX, dimmTypeQuantaGNR, etc.) and remove the backward
compatibility aliases from dimm.go.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* test: add test case for nil-result path in deriveDIMMInfoOther

Exercise the return nil, nil path (line 597) where the first DIMM
identifies a format but a subsequent DIMM fails to match it.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix: prevent nil panic in getDIMMSocketSlot for matchBoth formats

For formats requiring both bankLocPat and locPat to match (matchBoth),
getDIMMSocketSlot now requires both matches to be non-nil before calling
extractFunc. Previously, a partial match (one nil, one non-nil) would
pass a nil slice to extractFunc, causing an index-out-of-range panic.

This is realistic in deriveDIMMInfoOther where the format is identified
from dimms[0] but applied to all subsequent DIMM rows.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

Author: harp-intel