Incorrect Handling of HEP3 Chunks ends in Incorrect Payload strings (from Buffer)

[Andreas]

Hi,

first thx for your work in HEP. We tried to use HepOp together with OpenSips, but HepOps fails on Sending our Messages to Postgres.

After going down to the root-problem we found a Problem with Handling of Chunks in HEP-JS:

hep-js/index.js

Lines 385 to 390 in 51b3221

 var hepParse = new Parser() 
   .endianess("big") 
   .uint16("vendor") 
   .uint16("type") 
   .uint16("length") 
   .buffer("chunk", { length: "length" }); 

incorrectly implements the length of the chunk (line 390)

Correct would be

var hepParse = new Parser() 
   .endianess("big") 
   .uint16("vendor") 
   .uint16("type") 
   .uint16("length") 
   .buffer("chunk", { length: function() {return this.length - 6} })  ; // Vendor Typ and Length are part oft the length 

after this change the Patches at

hep-js/index.js

Line 422 in 51b3221

 return { rcinfo: { capturePass: data.chunk.slice(0,data.chunk.length-6).toString() } }; 

hep-js/index.js

Line 426 in 51b3221

 return { rcinfo: { correlation_id: data.chunk.slice(0,data.chunk.length-6).toString() } }; 

hep-js/index.js

Line 428 in 51b3221

 return { rcinfo: { hepNodeName: data.chunk.slice(0,data.chunk.length-6).toString() } }; 

correcting the "too long" chunk must be removed!

The Root cause for our problem was the missing correction (-6) at line

hep-js/index.js

Line 424 in 51b3221

 return { payload: data.chunk.toString() }; 

Without the correction the next Vendor ID (0x00) will be part of the string, and this crashes pg-promise during execution of SQL containing 0x00

[Lorenzo Mangani]

Thanks for your report and in-depth analysis! Could you submit a PR with your changes for review and integration?

[Andreas]

Jup, we will do this tomorrow.

Have a nice Evening,

Cu

Andreas