simple/docs/decoded-editing.md

Decoded editing — file-format survey and design

Status: design note for the Edit decoded feature (v3.33.0). It classifies the EFs the workbench knows (the client-side decoder registry) by the spec's own structure and content coding, assigns each an editing shape, and sets the release waves. The reference is the pinned specs (projects/docs/uicc/): 3GPP TS 31.102 v18.10 (USIM), TS 51.011 v4.15 (SIM), TS 24.008 v18.8 (network name / PLMN / address coding), TS 102 221 (file structure / Annex A text). pySim is a cross-check only.

The editor is client-side, like the decoders: the file is read raw, decoded into its schema's values, edited, re-encoded and written with /api/write. Only the files listed with a shape have an encoder; everything else keeps the raw editor.

1. Editing shapes

Each shape is one renderer + one encoder contract. fields are the flat form values, list is a transparent file holding N items, items is a record holding a TLV list; the record-scoped variants reuse the same field/item renderers inside a per-record block with a "write this record" checkbox (auto-checked when an input changes).

Shape	File / record content	Editor
S0 read-only	card- or network-managed values (location, keys, counters)	no editor; the button says why
S1 scalar/flag TR	one or a few bytes / flags	one form
S2 text TR	Annex A (TS 102 221) or NAI text	one form, text field
S3 BCD TR	ICCID / IMSI (nibble-swapped BCD, parity/type nibble)	one form, digits
S4 PLMN list TR	n × 3-byte PLMN (TS 24.008)	item rows, add/remove
S5 PLMN+ACT list TR	n × (3-byte PLMN + 2-byte ACT)	item rows, add/remove, tech multi-select
S6 bit/service list TR	service or identifier bit lists	item list with labels (checkbox per service/id)
S7 fixed-record fields	alpha + BCD address + ids (ADN family), LAI+id (OPL), code+alpha+category (ECC), TP parameters (SMSP)	per-record form + checkbox
S8 record TLV list	PNN (name TLVs), DIR (application templates), ARR (access rules), PBR (A8/A9/AA containers)	per-record item rows, add/remove/edit; unknown TLVs preserved
S9 nested structures	PBR containers, DIR templates, EXT chains	later: a tree, or read-only with a structured view
S10 dedicated builders	SMS (TPDU), SMSP (TPDU parameter set), key material	later: a dedicated form, not a generic encoder

Rules that apply to every encoder (and are what the tests pin):

• encode only the edited values, from the original record/file bytes: length, padding and anything the schema does not expose are preserved;
• never change the file/record length; refuse when the new value does not fit;
• preserve the original text coding when the text is unchanged, and choose a spec-legal coding when it changed (Annex A: unpacked GSM 7-bit / UCS-2; PNN network name: packed GSM 7-bit with the spare-bit count / UCS-2);
• refuse (not silently truncate) when an edit would invalidate a cross-file reference (EF.EXT1 chain, PBR/IAP, OPL→PNN record id, CMI pointer).

2. Survey of the known files

Wave is the release that adds the encoder. Files we do not decode yet are listed only where they matter for the plan; the spec's ~120 USIM EFs include many specialized ones (MMS, VGCS/VBS, ePDG, IMS configuration, …) that are not editing candidates.

2.1 Transparent files

EF	FID	Spec	Size	Content	Shape	Wave
EF.ICCID	2FE2	TS 51.011 10.3.1	10 B	BCD (nibble-swapped), F padding	S3	3.34
EF.IMSI	6F07	TS 31.102 4.2.2	9 B	length byte + '9'-prefixed BCD	S3	3.34
EF.IST	6F07	TS 31.102 4.2.2 (SAIP)	1–2 B	service bits	S6	3.34
EF.UST	6F38	TS 31.102 4.2.8	X ≥ 1	service bits	S6	3.34
EF.SST	6F38	TS 51.011 10.3.7	X ≥ 1	service bits	S6	3.34
EF.EST	6F56	TS 31.102 4.2.47	X ≥ 1	enabled-service bits	S6	3.34
EF.LI	6F05	TS 31.102 4.2.1	2n	ISO 639 language codes	S6	3.34
EF.LP / EF.PL	6F05 / 2F05	TS 51.011 10.3.1/10.3.2	1–n	language codes	S6	3.34
EF.SPN	6F46	TS 31.102 4.2.12	17 B	condition byte + Annex A text	S2	3.33
EF.PLMNsel	6F30	TS 51.011 10.3.4	3n (n ≥ 8)	PLMN list	S4	3.34
EF.FPLMN	6F7B	TS 31.102 4.2.16	3n (n ≥ 4)	PLMN list	S4	3.34
EF.EHPLMN	6FD9	TS 31.102 4.2.84	3n	PLMN list	S4	3.34
EF.PLMNwAcT	6F60	TS 31.102 4.2.5	5n (n ≥ 8)	PLMN + ACT	S5	3.33
EF.OPLMNwAcT	6F61	TS 31.102 4.2.53	5n (n ≥ 8)	PLMN + ACT	S5	3.33
EF.HPLMNwAcT	6F62	TS 31.102 4.2.54	5n (n ≥ 1)	PLMN + ACT	S5	3.33
EF.SPDI	6FCD	TS 31.102 4.2.66	x	TLV: A3 > 80 > PLMN list	S4	3.34
EF.GID1 / GID2	6F3E / 6F3F	TS 31.102 4.2.10/4.2.11	n	text (Annex A)	S2	3.34
EF.SUUME	6F54	TS 51.011 10.3.34	X+Y	alpha (Annex A)	S2	3.35
EF.IMPI/DOMAIN/IMPU	6F02/6F03/6F04	TS 31.102 (ISIM)	X	NAI text	S2	3.35
EF.ACC	6F78	TS 31.102 4.2.15	2 B	class bits	S1/S6	3.34
EF.PHASE	6FAE	TS 51.011 10.3.19	1 B	phase id	S1	3.34
EF.HPPLMN	6F31	TS 31.102 4.2.6	1 B	search period	S1	3.34
EF.AD	6FAD	TS 31.102 4.2.18	4+X	op mode, info, MNC length	S1	3.34
EF.CBMI	6F45	TS 31.102 4.2.14	2n	CB message ids	S6	3.34
EF.CBMID	6F48	TS 31.102 4.2.20	2n	CB message ids	S6	3.34
EF.CBMIR	6F50	TS 31.102 4.2.22	4n	CB id ranges	S6	3.34
EF.ACL	6F57	TS 31.102 4.2.48	X	version byte + TLV list (DD APNs)	S8-TR	3.35
EF.LOCI	6F7E	TS 31.102 4.2.17	11 B	TMSI/LAI/status	S0	—
EF.PSLOCI	6F73	TS 31.102 4.2.23	14 B	P-TMSI/RAI/status	S0	—
EF.EPSLOCI	6FE3	TS 31.102 4.2.91	18 B	GUTI/TAI/status	S0	—
EF.LOCIGPRS	6F53	TS 51.011 10.3.33	14 B	GPRS location	S0	—
EF.Kc / KcGPRS	6F20 / 6F52	TS 31.102 4.4.3.1/4.4.3.2	9 B	ciphering keys	S0	—
EF.Keys / KeysPS	6F08 / 6F09	TS 31.102 4.2.3/4.2.4	33 B	keys + status	S0	—

2.2 Linear fixed / cyclic files

EF	FID	Spec	Record	Content	Shape	Wave
EF.ADN	6F3A	TS 31.102 4.4.2.3	X+14	alpha + len + TON/NPI + 10 B BCD + CCP + EXT1	S7	3.33
EF.FDN	6F3B	TS 31.102 4.2.24	X+14	ADN layout (EXT2)	S7	3.33
EF.SDN	6F49	TS 31.102 4.2.29	X+14	ADN layout (EXT3)	S7	3.33
EF.MSISDN	6F40	TS 31.102 4.2.26	X+14	ADN layout (EXT5)	S7	3.33
EF.LND	6F44	TS 51.011 10.5.8	X+14	ADN layout, cyclic	S7	3.33
EF.BDN	6F4D	TS 31.102 4.2.44	X+15	ADN layout + Comparison Method Pointer	S7	3.33 (fix)
EF.PNN	6FC5	TS 31.102 4.2.58	X ≥ 3	TLVs: 43 full / 45 short / 80 PLMN info	S8	3.33
EF.OPL	6FC6	TS 31.102 4.2.59	8	LAI (PLMN+LAC) + PNN record id	S7	3.35
EF.ECC (SIM)	6FB7	TS 51.011 10.3.27	3n	emergency codes (TR)	S6	3.34
EF.ECC (USIM)	6FB7	TS 31.102 4.2.21	X+4	code(3) + alpha(X) + category(1)	S7	3.35
EF.EXT1–EXT5	6F4A–6F55	TS 31.102 4.4.2.4 / 4.2.30/31/45/37	13	type-dependent: additional data / subaddress	S9	3.36
EF.SMSP	6F42	TS 31.102 4.2.27	28+Y	alpha + parameter indicators + TP-DA/SCA/PID/DCS/VP	S10	3.36
EF.SMS	6F3C	TS 31.102 4.2.25	176	status byte + TPDU	S10	3.36
EF.ARR	2F06/6F06	TS 31.102 4.2.55	X	access-rule TLVs (ISO 7816-4)	S8/S9	3.36
EF.DIR	2F00	TS 51.011 10.3.3	X	application template TLVs (61 > 4F/50/73…)	S8/S9	3.36
EF.PBR	4F30/4FXX	TS 31.102 4.4.2.1	X	BER-TLV: A8/A9/AA containers, C0–CC links	S9	later
EF.EPSNSC	6FE4	TS 31.102 4.2.92	—	NAS security context	S0	—
EF.ICI/OCI (call logs)	6F80/6F81	TS 31.102 4.2.33/4.2.34	X+28/X+27	ADN-like + time/date → cyclic, ME-written	S0 (plan)	—

2.3 Coverage vs the spec

The decoder registry covers ~45 EFs. The spec defines many more; the remaining ones are either specialized protocols (MMS, VGCS/VBS, ePDG, IMS configuration, GBA, 5G slice configuration) or administrative/key files that must not be edited from a workbench without a dedicated flow (Keys, DCK, network parameters, START-HFN). Nothing in the survey forces a new shape beyond S9/S10.

3. Findings (defects and constraints the survey exposed)

1. EF.BDN is X+15, not X+14. The shared ADN codec assumes the last two bytes are CCP1/EXT1; BDN adds a trailing Comparison Method Pointer (X+15, TS 31.102 4.2.44). The read-only decoded view currently mis-parses BDN records, and the v3.33.0 encoder would corrupt them. The codec must take the layout (ADN vs BDN) and preserve the CMI pointer.
2. EF.ECC has two layouts. SIM: transparent 3n (TS 51.011 10.3.27); USIM: linear fixed X+4 (TS 31.102 4.2.21). The decoder implements the SIM layout only; on a USIM the alpha bytes are read as more codes. A structure-aware decode is needed before an ECC editor.
3. S8 is not one shape. PNN (flat tag+length+value TLVs) is the simple case. DIR/ARR/PBR nest TLVs and cross-reference other files; they need either a tree editor or a structured read-only view first.
4. References are part of the model. EXT1–5 chains (ADN/MSISDN/BDN), OPL→PNN record ids, PBR→(ADN/PBC/EXT/IAP) by FID/SFI. An editor that ignores them can leave a card in a contradictory state: encoders must preserve untouched references and refuse edits that would break one.
5. Cyclic files edit like linear ones for our purposes (write by record number), but the card owns the order (LND/ICI/OCI/ACM); the editor notes it and the write is still a plain UPDATE.
6. FID collisions across DFs (e.g. 6F02 = EF.IMPI in ISIM and EF.OCST in USIM) make the fid-only decoder fallback unsafe; the name-based lookup is the primary one and encoders should not be fid-only where a collision is possible.
7. Text coding is per file. Annex A (TS 102 221) for SPN/ADN/SUUME; the Network Name IE (TS 24.008 10.5.3.5a) for PNN, where the first byte is 0x80 | spare for packed GSM 7-bit and 0x90 for UCS-2 — the real-card vector 43 0A 82 D4 32 BB BC 7E B7 5D E4 32 … confirms it.

4. Release waves

• v3.33.0 (current) — framework: encoder registry, field/list/items schemas, record checkboxes + auto-check, save/validation; shapes S2, S5, S7, S8 for SPN, PLMNwAcT/OPLMNwAcT/HPLMNwAcT, ADN/FDN/SDN/MSISDN/LND, PNN; the BDN layout fix; the efPlmnAct EC-GSM-IoT-only (0x88) decode fix.
• v3.34.0 — S1/S3/S6 wave: UST/SST/IST/EST service toggles, LI/LP/PL, ACC, PHASE, HPPLMN, AD, CBMI/CBMID/CBMIR, ECC (SIM), ICCID, IMSI, PLMNsel/FPLMN/EHPLMN, SPDI, GID1/GID2.
• v3.35.0 — remaining S7 records: OPL (PNN record select), ECC (USIM), SUUME, IMPI/DOMAIN/IMPU; EXT1–5 as a read-only structured view.
• v3.36.0+ — S10/complex: SMSP (parameter form), SMS (TPDU editor, reusing the existing PDU builders), ARR (rule editor — the decoder already renders rule lines), DIR (application templates), PBR (nested tree).
• Read-only forever (S0) unless a dedicated flow appears: LOCI/PSLOCI/ EPSLOCI/EPSNSC/LOCIGPRS, Kc/KcGPRS, Keys/KeysPS, ACM, START-HFN/THRESHOLD, call logs (ME-written), DCK/NETPAR.

5. Testing strategy

• Round-trip: for every encoder, encode(decode(original), original) must reproduce the original bytes for spec-canonical records (the vectors come from the spec's own examples, pySim's test data and real-card dumps).
• Edit-one-field: changing a field changes only that field's bytes (padding, coding of untouched fields, ids and extension pointers stay).
• Refusals: over-capacity values, broken references, unsupported layouts (BDN with the CMI pointer, EXT1 chains) throw an actionable error.
• No capability by accident: a file without an encoder must never show the editor (gating test), and a schema/label drift guard keeps the editors and decoders aligned.