Aurora Labs Ltd.

Disclosed herein are techniques for creating and using tokens representing portions of programming code. Techniques include identifying a body of programming code; associating a plurality of tokens with respective portions of the body of programming code; configuring model input data for a code language processing model, the model input data comprising the plurality of tokens; and analyzing at least a part of the body of programming code using the code language processing model influenced by the model input data.

Disclosed herein are techniques for automatically reconfiguring code based on function and subfunction analysis. Techniques include determining, by parsing a code structure representing a plurality of functions, that at least one first function has a threshold degree of complexity; identifying, in response to the determination, a plurality of subfunctions based on the at least one first function, the plurality of subfunctions having a commonality with the at least one first function; and forming a second function by reconfiguring the first function to call at least one of the identified plurality of subfunctions.

Disclosed herein are techniques for shrinking security patches. Techniques include accessing executable code; scanning the executable code for an indicator of 3rd-party code associated with a software vulnerability; identifying, based on the scanning, the indicator of 3rd-party code; determining, based on the scanning, that the executable code includes a local fix patching the software vulnerability or that the executable code is not configured to rely on the 3rd-party code; and based on the determination that the executable code includes a local fix patching the software vulnerability or that the executable code is not configured to rely on the 3rd-party code, performing at least one of: generating a security patch file that does not patch the software vulnerability; or removing, from a security patch file, a patch associated with the software vulnerability, thereby reducing a size of the security patch file.

Disclosed herein are techniques for analyzing software build changes. Techniques include accessing first executable code associated with a first version; accessing second executable code associated with a second version; determining a code delta between the first executable code and the second executable code, the code delta being based on a change of at least one first element of code in the first executable code to at least one second element of code in the second executable code; determining a software vulnerability associated with at least one of the at least one first element of code or the at least one second element of code; and generating a report including a pairing of an indicator of the software vulnerability with an indicator of at least one of the at least one first element of code or the at least one second element of code.

Disclosed herein are techniques for reducing processing load for software testing. Techniques include accessing code for testing; performing functional analysis of the code to construct a functional behavior representation of the code; determining, based on the functional behavior representation, a first testing interaction between a first test and the code; determining, based on the functional behavior representation, a second testing interaction between a second test and the code; determining that the first testing interaction is stronger than the second testing interaction; and based on the determination that the first testing interaction is stronger than the second testing interaction, applying the first test to the code.

Disclosed herein are techniques for generating a linker script file. Techniques include accessing user definition code; accessing user configuration code; based on the user definition code and the user configuration code, identifying at least one linker script syntax; and generating a linker script file configured for generating executable code, the linker script file being based on the user definition code and the user configuration code.

Disclosed herein are techniques for training a model to predict data size. Techniques include initializing a model having model parameters; training the model to predict source code data size by: inputting first model input data to the model, the first model input data including a first set of source code parameters associated with a data size parameter associated with a first source code; and modifying at least one of the model parameters to improve prediction of source code data size by the model; and validating the model by inputting second model input data to the trained model, the second model input data including a second set of source code parameters associated with a data size parameter of a second source code.

Disclosed herein are techniques for training code language models. Techniques include making a plurality of programming code segments available to a code language processing model; providing an output of the code language processing model to one or more regression layers; determining, based on the one or more regression layers, a degree of functional similarity between two portions of the output; providing the degree of functional similarity to the code language processing model; and updating, based on the degree of functional similarity, the code language processing model.

Disclosed herein are techniques for creating and using tokens representing portions of programming code. Techniques include identifying a body of programming code; associating a plurality of tokens with respective portions of the body of programming code to generate a token-based representation of the body of programming code, wherein the associating comprises determining at least one canonical representation of at least one of the respective portions of the body of programming code; providing the token-based representation of the body of programming code to an emulator, the emulator being configured to interpret token-based representations; and receiving, from the emulator, an emulation result.

Disclosed herein are techniques for creating and using tokens representing portions of programming code. Techniques include identifying a first body of programming code associated with a hardware or software source attribute; associating a plurality of tokens with respective portions of the first body of programming code; configuring model input data for training a code language processing model customized in accordance with the hardware or software source attribute, the model input data comprising the plurality of tokens; and training, using the model input data, the code language processing model to analyze at least a part of the first body of programming code or a part of a second body of programming code, thus producing a customized and trained code language processing model in accordance with the hardware or software source attribute.