We have a dapp canister registered with sns which handles upgrades for other canisters. There is generic function in the canister to take wasm and upgrade respective canisters. While generating the payload for ExecuteGenericFunction didc encode fails to encode the wasm and results into argument too long. Is there a way to provide a file to didc encode.
didc command used:
didc encode --defs ".dfx/ic/canisters/platform_orchestrator/service.did" --method platform_orchestrator_generic_function "(variant {UpgradeSubnetCanisters = $(cat argument)})" --format blob'
argument file contains the wasm file encode in blob.
error:
Argument list too long
This is the whole script that I use to upgrade one of our canisters post_cache using platform_orchestrator which is controlled by sns.
CANISTER_NAME=post_cache
mkdir -p "proposals/${CANISTER_NAME}"
touch "proposals/${CANISTER_NAME}/upgrade.json"
WASM=".dfx/ic/canisters/${CANISTER_NAME}/${CANISTER_NAME}.wasm.gz"
CHAR=$(hexdump -ve '1/1 "%.2x"' "$WASM")
CHAR_ESCAPED=$(printf "%s" "$CHAR" | sed 's/../\\&/g')
touch argument
printf "(record {canister = variant {PostCacheWasm}; version = \"${{ github.ref_name }}\"; wasm_blob = blob \"%s\"})" "$CHAR_ESCAPED" > argument
PAYLOAD=$(nix-shell --run 'didc encode --defs ".dfx/ic/canisters/platform_orchestrator/service.did" --method platform_orchestrator_generic_function "(variant {UpgradeSubnetCanisters = $(cat argument)})" --format blob')
./quill sns \
--canister-ids-file ./sns/sns_canister_ids.json \
--pem-file actions_identity.pem \
make-proposal \
--proposal "(
record {
title=\"Upgrade ${CANISTER_NAME}\";
url=\"https://yral.com\";
summary=\"# Upgrade ${CANISTER_NAME}
${{ steps.changelog.outputs.changes}}
\";
action=opt variant {
ExecuteGenericNervousSystemFunction = record {
function_id = 4002: nat64;
payload = $PAYLOAD;
}
}
}
)" $NEURON_ID > "proposals/${CANISTER_NAME}/upgrade.json"
./quill send proposals/${CANISTER_NAME}/upgrade.json --yes
1 Like
I would recommend you to try ic-repl. In there, you can use file("argument") to load a blob from file.
The whole didc command can be simplified to
encode <platform_orchestrator_canister_id>.platform_orchestrator_generic_function(variant {UpgradeSubnetCanisters = file("argument")})
Furuthermore, the whole bash script can be rewritten in ic-repl, and with the offline mode in ic-repl, you don’t even need quill.
1 Like
Hey @chenyan, Thanks for response. Will it be possible to submit an Execute generic function proposal using ic-repl? Could you direct me to any documentation that I can refer to for this?
Certainly possible. We don’t have a lot documentations. The syntax and built-in functions are all listed in the readme file. There is also a few example scripts: ic-repl/examples at master · dfinity/ic-repl · GitHub.
Do you want to just replicate the bash script you posted? It seems it’s just a canister call to the SNS canister. The syntax for the Candid argument is the same as in quill. So roughly, it looks like this:
identity action "actions_identity.pem";
call "sns-canister-id".manage_neuron
The candid assist feature should guide you through to fill in the arguments for manage_neuron method.
Yes I just wanted to replicate the bash script. I wrote the script using ic-repl. Thanks for the suggestion is much cleaner approach. But I am running into the issue when using CANISTER_NAME from env variable. The script seems to work fine when CANISTER_NAME = post_cache | user_index but gives error when CANISTER_NAME = individual_user_template. Let me know if I am doing something wrong in the script.
#!/usr/bin/ic-repl -o
import platform_orchestrator_canister = "74zq4-iqaaa-aaaam-ab53a-cai";
import governance_canister="6wcax-haaaa-aaaaq-aaava-cai";
function generate_payload() {
let canister_type = variant {IndividualUserWasm};
if eq("${CANISTER_NAME}", "post_cache") {
let canister_type = variant {PostCacheWasm};
} else {
};
if eq("${CANISTER_NAME}", "user_index") {
let canister_type = variant {SubnetOrchestratorWasm};
} else {
};
if eq("${CANISTER_NAME}", "individual_user_template") {
let canister_type = variant {IndividualUserWasm};
} else {
};
encode platform_orchestrator_canister.platform_orchestrator_generic_function(
variant {
UpgradeSubnetCanisters = record {
version = "v1.0.0";
canister = canister_type;
wasm_blob = file("../.dfx/ic/canisters/$CANISTER_NAME/$CANISTER_NAME.wasm.gz");
}
}
)
};
let summary = "${SUMMARY}";
let res = encode governance_canister.manage_neuron(
record {
subaccount = encode ("a990d55e65186d5b4f38912c9a10748bf2e6626fb582a2b1d6442e087db56b10");
command = opt variant {MakeProposal = record {
title = "upgrade canister";
url = "yral.com";
summary = summary;
action = opt variant {ExecuteGenericNervousSystemFunction = record {
function_id = 4002;
payload = generate_payload();
}}
}}
}
)
What errors do you get? The script seems fine. At the end of generate_payload function, wasm_blob = file(...), I think you need to use ${CANISTER_NAME} as well?
This line is also strange. I think it should be blob "\a9\90\d5\5e..."
@chenyan if I set CANISTER_NAME = individual_user_template I get this error:
thread 'main' panicked at src/command.rs:181:45:
byte index 1591 is out of bounds of `
import platform_orchestrator_canister = "74zq4-iqaaa-aaaam-ab53a-cai";
import governance_canister="6wcax-haaaa-aaaaq-aaava-cai";
function generate_payload() {
let canister_type = variant {IndividualUserWasm};
if eq("${CANISTER_NAME}", "post_ca`[...]
stack backtrace:
0: 0x1032f9385 - <std::sys_common::backtrace::_print::DisplayBacktrace as core::fmt::Display>::fmt::h46bc8e86d021398c
1: 0x10331fd53 - core::fmt::write::h1c24e945c7ef8357
2: 0x1032f52ae - std::io::Write::write_fmt::hd9fe01a40900d808
3: 0x1032f9159 - std::sys_common::backtrace::print::hdbfcabb7d363540a
4: 0x1032fa735 - std::panicking::default_hook::{{closure}}::hecb24382dfeeea98
5: 0x1032fa4ae - std::panicking::default_hook::h78dca947ab9e89eb
6: 0x1032faba3 - std::panicking::rust_panic_with_hook::hb83cee65df04957a
7: 0x1032faab4 - std::panicking::begin_panic_handler::{{closure}}::hbe6e44d88f82bc77
8: 0x1032f9879 - std::sys_common::backtrace::__rust_end_short_backtrace::h6560cb9c1f1224c1
9: 0x1032fa7f6 - _rust_begin_unwind
10: 0x103359935 - core::panicking::panic_fmt::h2aac8cf45f7ae617
11: 0x103323636 - core::str::slice_error_fail_rt::h4d9a57249966a0ef
12: 0x10335a139 - core::str::slice_error_fail::h82e5c6802f33005a
13: 0x102b1995e - ic_repl::command::Command::run::h2d4c66b3d365e35d
14: 0x102ae5947 - ic_repl::main::hb06971eee4e8f42a
15: 0x102ac7dd6 - std::sys_common::backtrace::__rust_begin_short_backtrace::h1fa2fb319d8f789b
16: 0x102ab22a1 - std::rt::lang_start::{{closure}}::h0d70c3b73be83c25
17: 0x1032ecd10 - std::rt::lang_start_internal::h963c626e0d7ddd6e
18: 0x102aebbdc - _main
Hey @chenyan, I made all the suggested changes still facing the same issue 
I see. It’s a bug on my end for resolving environment variables. Will fix it soon.
Thank you so much @chenyan. This fixed the issue 
Hey, is there way in ic-repl to programmatically convert neuron-id text to blob format as you mentioned in this comment. I tried ("a990d55e65186d5b4f38912c9a10748bf2e6626fb582a2b1d6442e087db56b10" : blob); but it didn’t work
No, you need to insert the escape character, e.g., blob "\a9\90\d5...".
Just curious, where do you get the neuron-id text, instead of a blob? I could add a hex function to do the conversion, but want to see if there are other workaround.
1 Like