gavindiaz/pumpfun-bonkfun-bot_github
1
0
Fork 0
mirror of https://github.com/chainstacklabs/pumpfun-bonkfun-bot.git synced 2026-08-05 03:37:44 +00:00
Code Issues Packages Projects Releases Wiki Activity
280 Commits 2 Branches 0 Tags
678fa19fe477fbbfcddf6305343eba59ae79231e
Commit Graph

12 Commits

Author SHA1 Message Date
Anton Sauchyk 678fa19fe4 fix(learning-examples): replace getProgramAccounts scan in get_graduating_tokens (#182) 
* fix(learning-examples): replace getProgramAccounts scan in get_graduating_tokens

The pump.fun program owns over 10M accounts and no provider will scan it, so
get_graduating_tokens.py could not run at all (#178). getProgramAccountsV2 is
not a fix: it is a provider extension rather than core Agave, and its `limit`
is a scan budget, not a result count, so one filtered answer costs ~1000
sequential pages.

Rewrite discovery onto filtered programSubscribe, which applies dataSize and
memcmp server-side and is accepted even by public api.mainnet-beta.solana.com.
Every write to a curve pushes the full 151-byte account, so progress is
computed per update with no accumulated state. Add a Geyser sibling that
reports the same thing with the slot and signature behind each update.

Also fix two bugs that would have survived the rewrite: the mint lookup
queried SPL Token, which returns nothing for the Token-2022 ATAs that every
create_v2 coin uses, and the threshold was a hardcoded constant rather than
Global.initial_real_token_reserves.

Closes #178

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

* fix(learning-examples): drop the graduation cutoff that filtered nothing

zero_prefix_gate offered a cutoff so high that no coin could fail it, so for any
--min-progress below 64.5% the subscription was unfiltered while the banner
reported a filter as active. Offer only the three cutoffs that actually narrow,
and say plainly when none applies.

Rewrite the threshold notes in both scripts in plain English: which cutoffs
exist, whether a given threshold gets one, and the part that matters — the
pre-filter saves bandwidth but does not decide the answer, so the requested
percentage is honoured either way. The banner now names the cutoff as a
percentage instead of byte offsets.

Both directions were checked against mainnet by running the filtered and
unfiltered subscriptions side by side for a minute, on both transports: the
filtered stream matched the below-cutoff set exactly, with 143 of 168 curves
above the cutoff on WebSocket and 128 of 154 on Geyser.

Also document the two scripts in the README example table, and record under
throughput that getProgramAccounts over the whole pump.fun program is no longer
served by any provider.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-07-29 16:52:04 +02:00
Anton Sauchyk dcf66ec1b3 docs(agents): correct stale layer chain, ruff baseline, and fund-spending script rule 
Verified every claim in CLAUDE.md against the code after the recent debugging,
IDL, bot, and learning-example updates. Four were wrong:

- The dependency chain listed `utils` before `interfaces`, but
  utils/idl_manager.py imports interfaces.core and interfaces imports nothing
  internal. It also made trading and monitoring peers when trading imports
  monitoring, and omitted cleanup and geyser entirely.
- Ruff baseline said ~2400 errors; `uv run ruff check` reports 1676.
- `live_*` was presented as the marker for scripts that spend real funds. Only
  2 of the 13 scripts that submit transactions carry that prefix — manual_*,
  mint_and_buy* and cleanup_accounts.py spend too.
- compare_listeners.py was cited without its path, ambiguous against
  listen-migrations/compare_migration_listeners.py.

The protocol section (v2 account counts, BondingCurve offsets, PumpSwap i128,
PDA seeds, fee-recipient sets, tx-status rules) checks out against
idl/pump_fun_idl.json and src/; both offline verifiers still pass.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-07-29 15:38:47 +02:00
Anton Sauchyk a33f4fd035 fix(learning-examples): repair crashing listeners and decoders, prune obsolete paths, unify naming (#181) 
Audited every script in learning-examples/ against mainnet.

Broken — found by running them, all invisible offline:

- listen_geyser.py crashed with IndexError after ~11 coins: it never resolved
  v0 address-lookup-table accounts, which geyser reports in
  meta.loaded_writable_addresses / loaded_readonly_addresses. Resolving them
  removes the crash and brings detections level with the WebSocket listeners,
  35 coins each per 150 s.
- compare_listeners.py logged 13,090,862 error lines / 888 MB in 150 s and never
  printed its own 30-second report: the inner recv() loop caught ConnectionClosed
  in a broad handler that only logged, so every following recv() raised at once
  and the outer reconnect handler was unreachable. Now 12 KB and exit 0. Same
  shape fixed in compare_migration_listeners.py, listen_blocksubscribe.py and
  extract_blocksubscribe_transactions.py; the last two also gained the reconnect
  loop their siblings already had.
- decode_from_gettransaction.py matched instructions on account count instead of
  discriminator, reporting a real 19-account create_v2 as claim_cashback with
  every account under the wrong name. It also walked only top-level
  instructions, and in 40 consecutive pump.fun transactions there was 1
  top-level pump instruction against 8 inner ones.
- decode_from_blocksubscribe.py crashed on every real create_v2: on chain the
  trailing args are variable length, 0001 in one tx and 00 in another, so
  is_cashback_enabled can be absent entirely.
- poll_bonding_curve_progress.py polled a hardcoded dead mint and took no argv.

Obsolete:

- Delete listen_blocksubscribe_old_raydium.py. Seven minutes on mainnet produced
  0 initialize2 events while the wrapper listener caught 3 real migrations.
- Delete the duplicate geyser stubs and protos under listen-new-tokens/. The
  protos were byte-identical to src/geyser/proto and the stubs had drifted; both
  geyser examples now import src.geyser.generated.
- Recapture all four fixtures. The old ones were from Aug 2024 and included a
  49-byte pre-creator bonding curve.

Behind the protocol:

- fetch_price.py, get_bonding_curve_status.py, poll_bonding_curve_progress.py
  and decode_from_getaccountinfo.py never read quote_mint and scaled by a
  hardcoded 1e9. Against a live USDC-paired curve the price was off by 1000x.
- get_pumpswap_pools.py stopped parsing at coin_creator and missed the i128
  virtual_quote_reserves. Live pools carry 17.5845 SOL of them, which
  under-prices by 3.5-23.9% when ignored.

Duplication and naming:

- Merge manual_buy_cu_optimized.py into manual_buy.py --cu-optimized. The
  deleted file's docstring said 512 KB while its code used 16 MB; simulation
  confirms 512 KB and 4 MB both fail MaxLoadedAccountsDataSizeExceeded on
  Token-2022 mints, so 16 MB is the correct value.
- Merge listen_logsubscribe_abc.py into listen_logsubscribe.py. Its ATA
  derivation hardcoded the legacy token program, so every Associated BC it
  printed for a Token2022 coin was an address that does not exist on chain.
  Fixed on merge and cross-checked 59/59 against on-chain accounts.
- Remove 19 dead symbols. BREAKING_FEE_RECIPIENTS is still live in the PumpSwap
  scripts and stays there.
- Normalize naming: kebab-case directories, RPC method names as one lowercase
  token, scripts verb-first. Rules documented in CLAUDE.md.

get_graduating_tokens.py is knowingly left broken: getProgramAccounts over the
whole pump program is now rejected by providers and it needs a
getProgramAccountsV2 rewrite, which belongs in its own PR.

Verified: both offline gates pass, all 41 examples parse, every read-only script
exercised on mainnet against SOL- and USDC-paired coins, no new ruff findings
(427 -> 413). No script that spends real funds was run.

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-07-29 15:26:26 +02:00
Anton Sauchyk 7727015548 docs: consolidate agent docs, rewrite README, prune deps (#179) 
Rewrite README.md around setup and configuration: fix the clone URL,
document the actual .env variable names, add tables for bots/*.yaml and
the learning-examples directories, and drop the empty changelog, the
2025 roadmap, and the protocol deep-dives that duplicated CLAUDE.md.

Make CLAUDE.md the single agent guide and symlink AGENTS.md to it.
AGENTS.md carried wrong env var names, a stale Python floor, and a
config key that does not exist; its safety rules move into CLAUDE.md.
Document that `uv pip install -e .` puts src/ on sys.path, so imports
are `from utils.logger import ...` rather than `from src.utils...`.

Delete .cursor/rules/, .kiro/steering/, and .windsurf/rules/ - three
byte-identical copies of rules referencing APIs that do not exist in
src/. All three tools read AGENTS.md natively.

Fix pyproject.toml:
- requires-python >=3.9 -> >=3.11; the code uses `X | None` (3.10+) and
  ruff already targets py311
- drop borsh-construct and construct-typing, neither of which is
  imported anywhere (construct-typing still resolves via solana)
- move grpcio-tools to the dev group; it is protoc, needed only to
  regenerate the geyser_pb2 stubs, never at runtime
- move dev deps from [project.optional-dependencies] to
  [dependency-groups] so `uv sync` installs ruff, making the documented
  `ruff check` / `ruff format` commands actually available

Also gitignore .claude/settings.local.json, which is per-developer.

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-07-29 14:07:17 +02:00
Anton Sauchyk 836d873d27 fix: report on-chain trade outcomes correctly, and repair every broken learning example (#177) 
* fix(tx-status): read meta.err before reporting a trade succeeded

confirm_transaction answers "did this signature land in a block?", never
"did it succeed". A landed transaction can have reverted, and RPC reports
that only in meta.err. Reading whether a call threw instead of reading
meta.err produced false results in both directions.

False success — a reverted trade reported as confirmed:

- ten examples (pump.fun, pumpswap, letsbonk) printed "Transaction
  confirmed" without checking meta.err
- src/cleanup/manager.py discarded confirm_transaction's boolean and
  logged "Closed successfully" unconditionally, so a reverted close
  reported rent as reclaimed while the account stayed open
- learning-examples/cleanup_accounts.py did the same

False failure — a good trade reported as unconfirmed:

- _get_transaction_result omitted maxSupportedTransactionVersion, so the
  RPC answered -32015 for every versioned (v0) transaction. meta.err was
  unreadable and a successful trade read back as failed. The bot sends
  legacy transactions, which is the only reason this was survivable.
- confirm_transaction raised TypeError on a base58 str (solana-py wants a
  Signature) while _get_transaction_result raised on a Signature (not
  JSON serializable). Both were swallowed by a broad except into "not
  confirmed". The annotations pointed the wrong way too:
  build_and_send_transaction returns Signature, not str.

Changes:

- add learning-examples/tx_status.py — assert_transaction_succeeded and
  confirm_and_assert, replacing the copy duplicated in mint_and_buy{,_v2}
- wire it into the ten examples that confirmed without checking
- read the boolean in both cleanup paths
- normalize str/Signature at the client boundary; correct the annotations
- send maxSupportedTransactionVersion: 0 on getTransaction
- split verify_transaction_succeeded out of confirm_transaction so the
  meta.err check can run against a transaction that landed earlier

Two reporting bugs found while testing the above:

- live_v2_round_trip read balances at solana-py's default (finalized)
  commitment while confirming trades at "confirmed", so the end read saw
  pre-trade state and it printed "net change: +0.000000000 SOL" after a
  real round trip. Verified: on a busy account finalized trails confirmed
  by ~263k lamports.
- cleanup_accounts produced no output at all, success or failure, because
  get_logger attaches no handler and only the bot installs one. httpx is
  pinned to WARNING alongside it — the RPC endpoint carries an API key.

Adds learning-examples/verify_tx_status_checks.py: offline stub checks, a
scan that fails if an example confirms without checking meta.err, an AST
check that nothing in src/ discards the boolean, a guard that
getTransaction opts into v0, and --live, which replays issue #175's three
signatures against mainnet and requires both layers to reject them on
meta.err (Custom: 6062) rather than on a failed fetch. The two src/ guards
were mutation-tested: reintroducing each bug makes them fail.

The BuybackFeeRecipientMissing (6062) half of #175 was already fixed by
the buy_v2/sell_v2 migration in 02343b7.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

* fix(client): retry RPC timeouts instead of letting them escape post_rpc

aiohttp signals a request timeout with asyncio.TimeoutError, which is not
an aiohttp.ClientError. post_rpc caught only the latter, so every RPC
timeout propagated to the caller unretried — and str() on it is empty, so
whatever logged it printed a blank reason.

Found while running learning-examples/live_listener_matrix.py: three of
the four listeners died mid-run with "CRASHED: " and no message. The
endpoint was answering getHealth in ~100ms while getAccountInfo hung past
60s, and every caller that touches it (sol_balance, ata_is_closed,
AccountCleanupManager.cleanup_ata) went down with it. With the retry in
place the same run degrades to a logged failure and completes.

Also makes the verifier report a raising check as a failure rather than
aborting the run — several checks assert that a call does NOT raise, so
the raise is the finding and the remaining checks still need to report.

Covered by "RPC timeouts are retried, not raised" in
verify_tx_status_checks.py, mutation-tested against the one-exception
version.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

* fix(examples): raise websocket frame limit, load .env, drop dead code

Running every learning example turned up three separate reasons a shipped
script could not work at all.

blockSubscribe examples were completely non-functional. websockets
defaults max_size to 1 MiB and Solana's blockSubscribe frames run well
past that, so the connection died with a 1009 close on the first real
block and the retry loop spun. listen_blocksubscribe.py logged 158,613
"message too big" errors in 35s and decoded zero tokens; compare_listeners
produced 10.1M error lines and its block column never reported anything.
Pass max_size=WEBSOCKET_MAX_MESSAGE_BYTES (32 MiB, the value the bot's own
listeners already use) at all 12 example connect sites — logsSubscribe and
programSubscribe included, since they have the same latent ceiling.

After: listen_blocksubscribe decodes tokens with 0 errors, and
compare_listeners reports provider_1_block alongside geyser and logs.

Seven examples read SOLANA_* from the environment but never called
load_dotenv(), so they only ran with variables already exported —
manual_buy, manual_buy_cu_optimized, manual_buy_geyser, manual_sell,
fetch_price, blockSubscribe_extract_transactions and
sample_cashback_pumpswap. manual_buy died on
"None isn't a valid URI: scheme isn't ws or wss" against a normal .env
checkout.

Dead code and artifacts:

- unused datetime import in listen_pumpportal
- initial_real_token_reserves computed and never read in mint_and_buy and
  mint_and_buy_v2
- gitignore blockSubscribe-transactions/, which a shipped example writes
  into 1200+ files deep and which nothing ignored
- stop tracking trades/trades.log; .gitignore has listed it all along, but
  a tracked file ignores .gitignore

ruff over src/ and learning-examples/ goes 2396 -> 2393 findings: the three
removals, nothing new.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

* fix: read balances at confirmed, derive letsbonk platform_config per pool

Running every example against mainnet surfaced four more bugs, three of
them the same root cause as the commitment bug already in this branch:
solana-py defaults to "finalized", but trades confirm at "confirmed", and
finalization lags far enough behind that anything read in between is
pre-trade state.

get_token_account_balance defaulted to finalized. Cleanup reads it to
decide whether to burn before closing, so right after a sell it saw the
pre-sell amount and built a burn for tokens the account no longer held —
the burn + close reverted with InsufficientFunds and the rent stayed
locked. Observed live: "Burning 35766666 tokens" on an account the sell
had already emptied, then Custom(1). Defaults to confirmed now, which is
also what confirm_transaction uses.

manual_sell_pumpswap read the user's base balance at finalized, which
failed two ways in one session: "could not find account" when the ATA had
been created by a buy moments earlier, and a stale non-zero balance whose
transfer then reverted with insufficient funds. Pool vault reads move to
confirmed too, so quotes are not priced off stale reserves.

The letsbonk examples hardcoded platform_config. LaunchLab pools do not
share one — partner launches carry their own — so every buy/sell against
such a pool failed with ConstraintAddress (2012):

  AnchorError caused by account: platform_config
  Left:  5thqcDwKp5QQ8US4XRMoseGeGbmLKMmoKZmS6zHrQAsA
  Right: FfYek5vEz23cMkWsdJwG2oa6EphsvXSHrGpdALN4g6W1

PoolState carries platform_config at field 18 and these scripts already
parse the pool, so they now take the pool's own value and keep the
constant only as a fallback. src/platforms/letsbonk already documented
this; only the examples were stale. All four scripts now simulate and land
(buy_exact_in, sell_exact_in, buy_exact_out, sell_exact_out).

They also discarded simulation logs on failure, printing an error number
with no indication of which account or constraint broke. They print the
program logs now — that is how the above was diagnosed.

Runnability and credentials:

- fetch_price, cleanup_accounts and the four letsbonk scripts hardcoded a
  placeholder ("...", "YOUR_TOKEN_MINT_ADDRESS_HERE") and could not run at
  all without editing the source. They take argv[1] now, matching
  manual_sell and the pumpswap scripts. cleanup_accounts takes "2022" as
  argv[2] for Token-2022 mints, which every pump.fun coin is.
- six scripts printed the RPC/WSS endpoint, which carries an API key, into
  stdout. They print only the host now.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

* fix(cleanup): never burn wrapped SOL, and label what actually happened

learning-examples/cleanup_accounts.py burned any non-zero balance before
closing. The token program rejects a burn of native (wrapped) SOL with
NativeNotSupported (error 10), so the burn + close transaction reverted and
a WSOL account could never be cleaned up — the pumpswap examples leave one
behind on every sell. Closing a WSOL account already returns both the
wrapped lamports and the rent, so there is nothing to burn first.

src/cleanup/manager.py already had this guard; only the example was
missing it. Observed live: "Burning 2206381 tokens" then Custom(10), with
2.2m lamports stranded until the guard went in.

The success line also claimed "Burned and closed" for the unwrap path,
which burns nothing. It now reports Unwrapped/Burned/Closed to match the
instructions actually built.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

* chore: drop captured-output artifacts and a spent one-off investigation script

learning-examples/decoded_buy_tx_from_getTransaction.json and
decoded_create_tx_from_getTransaction.json are not JSON — they are captured
stdout from the decode scripts, several objects concatenated, so json.load
raises "Extra data" on both. Nothing reads them and nothing can. The
raw_*.json fixtures stay: all three are valid input and were re-checked
through decode_from_getTransaction / decode_from_blockSubscribe /
decode_from_getAccountInfo.

learning-examples/pumpswap/sample_cashback_pumpswap.py was added by #168 to
find the position of the extra account that cashback PumpSwap pools
require. That question is answered — the layouts are in
platforms/pumpfun/instruction_builder.py and machine-checked against the
IDL by verify_v2_account_layout.py. It is undocumented, unlike every other
pumpswap example in the README, and nothing imports it.

Not touched: logs/ (run history), .cursor/.kiro/.windsurf (deliberate
mirrors of the same rules for other editors), and the raw_*.json fixtures.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

* fix(tx-status): fail closed on missing metadata, stop retrying landed reverts

Addresses the must-fix findings from review. Two of them were wrong in code
this branch added.

assert_transaction_succeeded folded meta=None into err=None and returned
success. Missing execution metadata means the outcome is unknown, not that
it succeeded — which is the exact mistake this module exists to prevent. It
raises now.

It also always read at "confirmed" while confirm_and_assert accepted a
commitment argument, so confirming at "finalized" and then reading status at
"confirmed" could report success before the finalization the caller asked
for. The commitment is threaded through.

The manual buy/sell retry loops built one message and blockhash before the
loop, so a landed revert was retried by resubmitting byte-identical signed
bytes — three attempts with backoff that could never succeed. Reverts now
raise TransactionRevertedError (a RuntimeError subclass, so existing
handlers keep working) and the loops treat it as terminal. Rebuilding and
re-signing per attempt is the fuller fix but a bigger change to these
scripts than this branch should carry.

cleanup_accounts guessed the mint's token program, defaulting to legacy SPL
unless argv[2] was "2022". The ATA address differs between programs, so a
wrong guess derives an address that does not exist and the script reports
"already closed" for an account it never looked at. It reads the owner off
the mint account instead, which is authoritative, and rejects anything not
owned by a token program. Verified live: pump.fun coins resolve to
Token-2022, letsbonk and USDC to legacy. argv[2] is gone — nothing to guess.

urlsplit(...).netloc keeps any user:pass@ userinfo, so the endpoint
redaction added earlier still printed credentials for providers that put the
key there. Uses .hostname in all five sites.

Four checks added to verify_tx_status_checks.py, each mutation-tested:
missing metadata, revert-is-terminal, commitment propagation, and a scan
that fails if any example goes back to netloc.

Skipped, both minor and pre-existing repo-wide rather than introduced here:
validating that RPC/WSS env vars are non-None before connecting (every
example has this shape; fixing two of them would just make it inconsistent),
and the cryptic base58 error when a letsbonk script runs with no argument,
which matches the "..." placeholder convention the pumpswap and manual_sell
examples already use.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QFrxcZfCf9C76voCb7M8Pa

---------

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-07-29 13:44:35 +02:00
Anton Sauchyk 02343b775b feat(pumpfun): migrate to buy_v2/sell_v2 and support non-SOL quote assets (#176) 
Refresh the vendored IDLs from pump-fun/pump-public-docs @ 9c82f61 and move all
pump.fun trading onto the v2 instruction interface. This is required, not
optional: legacy buy/sell cannot trade coins paired against a quote asset other
than SOL, and USDC is already whitelisted in the on-chain Global account.

Protocol changes absorbed:

- buy_v2 (27 accounts) / sell_v2 (26 accounts) replace the legacy instructions.
  Every account is mandatory and the order is identical for all coins, so the
  conditional cashback/mayhem account lists are gone. Legacy remains available
  via PumpFunInstructionBuilder(use_legacy_instructions=True).
- BondingCurve is 151 bytes: virtual_sol_reserves -> virtual_quote_reserves,
  real_sol_reserves -> real_quote_reserves, plus quote_mint at offset 83. Old
  field names are kept as aliases so existing callers keep working.
- v2 instruction data drops the track_volume OptionBool; amounts are in the
  quote mint's raw units rather than always lamports.
- create_v2 carries a non-SOL quote mint as optional remaining accounts 17-19,
  and CreateEvent gained quote_mint, so extreme_fast_mode can resolve the quote
  asset without an extra fetch.

USDC support: new trade.quote_amounts and filters.allowed_quote_mints config,
accepting "sol"/"usdc" aliases or raw mints. Amounts are per-quote-mint because
1 USDC and 1 SOL are not interchangeable. A coin whose quote mint has no
configured amount is skipped rather than traded at the wrong size, so SOL-only
configs are unaffected.

Bug fixes found while verifying:

- The logs and blocks listeners set no websocket max_size, so any frame over
  1 MiB closed the connection with 1009 and the token in it was lost. Raised
  to 32 MiB.
- PumpSwap priced against the raw quote vault balance, ignoring the new
  Pool.virtual_quote_reserves (i128 at offset 245; live pools are 301 bytes).
  Upstream's note that this field is 0 everywhere is out of date: a live pool
  carries 17.58 SOL against a 148 SOL vault, a 10.15% price error.
- The seller read curve state once at confirmed commitment and silently fell
  back to create-time values, risking a stale creator_vault and ConstraintSeeds.
  It now retries at processed, matching the buyer.
- Account cleanup would burn wrapped SOL when force_burn was set, destroying
  value that closing the account returns. WSOL is now closed without burning.
- The mint scripts treated a landed transaction as a successful one, so a
  reverted buy printed as success. They now assert the on-chain result.

Compute unit limits retuned from mainnet measurements: buy 100k -> 180k,
sell 60k -> 120k. Mint-and-buy is no longer atomic, because create_v2 plus
buy_v2 exceeds the 1232-byte transaction limit; both mint scripts send two
transactions.

Adds learning-examples/pump_v2.py as one shared, standalone v2 toolkit for the
example scripts, and three verification scripts: an offline layout check
against the IDL, a no-funds mainnet simulation, and a live listener matrix that
buys, sells and closes the ATA per listener.

Verified on mainnet: all four listeners (geyser, logs, blocks, pumpportal) and
all eight example scripts completed a real buy, sell and ATA close, each
confirmed by reading the transaction result back rather than trusting
confirmation alone. The USDC path is verified structurally only; no USDC-paired
coin could be found on-chain to exercise it.

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
 2026-07-28 17:58:33 +02:00
Anton Sauchyk cc85daaeb3 chore(pumpswap): post-2026-04-28 cutover + cashback layout (#168) 
* chore(pumpswap+docs): post-2026-04-28-cutover PumpSwap unlock + IDL docs

Drops the `INCLUDE_BREAKING_FEE_ACCOUNTS = False` gate in the PumpSwap
learning examples so the +2 breaking-fee accounts (fee recipient
readonly + its quote-mint ATA mutable) are always appended after
`pool-v2`. Mainnet pump-amm rejects this format pre-cutover (verified
6023 Overflow), so this PR is intentionally **draft until 2026-04-28
16:00 UTC** when the cutover happens; mark it ready-for-review and live-
test then.

Also captures the protocol gotchas we learned during this migration:

- README: a new "2026-04-28 program upgrade" section + an explicit note
  that the vendored IDL is incomplete (missing `bonding-curve-v2` and
  `pool-v2` PDAs that the on-chain program actually requires) with
  pointers to cross-check against on-chain txs.
- CLAUDE.md: a new "Pump.fun protocol notes" subsection summarising
  the same gotchas plus the BC/Pool/CreateEvent layout details and the
  extreme_fast_mode gotcha.

Open question (call out at review time, resolve post-cutover):
- BREAKING_FEE_RECIPIENT.md shows PumpSwap cashback account counts of
  27 buy / 26 sell vs 26 / 24 non-cashback — the extra cashback account
  seed/position isn't documented. Need to sample a real successful
  cashback PumpSwap tx after cutover and add the branch.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* feat(pumpswap): wire cashback layout for post-2026-04-28 program upgrade

Cashback PumpSwap pools require extra writable accounts inserted before
pool-v2. Identified layouts by sampling on-chain post-cutover txs:

- buy: insert user_volume_accumulator_quote_ata (27 accounts vs 26)
  ref tx 4JaWdExj…fvjK
- sell: insert user_volume_accumulator_quote_ata + user_volume_accumulator
  (26 accounts vs 24)
  ref tx 4ei1cJV7…NP3

Detect via pool account byte 244 (is_cashback_coin). Adds the standalone
sample_cashback_pumpswap.py used to reverse-engineer the layouts.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-04-28 18:20:53 +02:00
smypmsa 8ab8932168 docs(claude): add claude code rules 2025-08-11 05:35:25 +00:00
smypmsa bb35101a8f wip(core): platform aware trading 2025-08-02 09:25:39 +00:00
smypmsa 060a69da6e docs(claude): update uv commands 2025-07-18 12:05:30 +00:00
smypmsa 4d9864f936 docs: remove Raydium 2025-07-14 19:58:22 +00:00
smypmsa 51eae4f5a3 docs: add CLAUDE.md 2025-07-14 19:56:46 +00:00