Figure - uploaded by Antonio Mastropaolo
Content may be subject to copyright.
Source publication
Logging is a practice widely adopted in several phases of the software lifecycle. For example, during software development log statements allow engineers to verify and debug the system by exposing fine-grained information of the running software. While the benefits of logging are undisputed, taking proper decisions about where to inject log stateme...
Contexts in source publication
Context 1
... simplify the discussion of the results, we answer our research questions together through a quantitative and qualitative analysis. Table 3 reports the results achieved by the four experimented models output of different pre-training strategies (i.e., Multi-Task, LogStmt-Task, Denoising-Task, and No pre-training) in terms of correct predictions. Table 3 shows the correct predictions for all combinations of the three "components" to predict (i.e., log level, position where to inject it, and log message). ...Context 2
... 3 reports the results achieved by the four experimented models output of different pre-training strategies (i.e., Multi-Task, LogStmt-Task, Denoising-Task, and No pre-training) in terms of correct predictions. Table 3 shows the correct predictions for all combinations of the three "components" to predict (i.e., log level, position where to inject it, and log message). In other words, we analyze cases in which (i) at least one of the three components to predict was correct (e.g., at least the level), (ii) at least two were correct (e.g., level and location), and (iii) the entire log statement was correctly synthesized (level, position, and message). ...Context 3
... other words, we analyze cases in which (i) at least one of the three components to predict was correct (e.g., at least the level), (ii) at least two were correct (e.g., level and location), and (iii) the entire log statement was correctly synthesized (level, position, and message). Table 3 can be read as follows. Each row includes three symbols below the three components to predict. ...Context 4
... in general, concerning the role played by the pre-training, we observed a substantial boost of performance only in the case of the Denoising-task (+2.51% of perfectly predicted log statements). The other two pre-trainings only marginally improved the performance of the base model (see Table 3). In the following, we focus on the best-performing model. ...Context 5
... two-fold achievement (i.e., log level and position) suggests that LANCE can effectively support developers with logging activities. Looking at the third row in Table 3, it is instead clear that LANCE struggles to generate logging messages that are identical to the ones manually written by developers (success in 16.90% of cases). This difference in performance among the three log statement "components" to predict is kind of expected. ...Context 6
... simplify the discussion of the results, we answer our research questions together through a quantitative and qualitative analysis. Table 3 reports the results achieved by the four experimented models output of different pre-training strategies (i.e., Multi-Task, LogStmt-Task, Denoising-Task, and No pre-training) in terms of correct predictions. Table 3 shows the correct predictions for all combinations of the three "components" to predict (i.e., log level, position where to inject it, and log message). ...Context 7
... 3 reports the results achieved by the four experimented models output of different pre-training strategies (i.e., Multi-Task, LogStmt-Task, Denoising-Task, and No pre-training) in terms of correct predictions. Table 3 shows the correct predictions for all combinations of the three "components" to predict (i.e., log level, position where to inject it, and log message). In other words, we analyze cases in which (i) at least one of the three components to predict was correct (e.g., at least the level), (ii) at least two were correct (e.g., level and location), and (iii) the entire log statement was correctly synthesized (level, position, and message). ...Context 8
... other words, we analyze cases in which (i) at least one of the three components to predict was correct (e.g., at least the level), (ii) at least two were correct (e.g., level and location), and (iii) the entire log statement was correctly synthesized (level, position, and message). Table 3 can be read as follows. Each row includes three symbols below the three components to predict. ...Context 9
... in general, concerning the role played by the pre-training, we observed a substantial boost of performance only in the case of the Denoising-task (+2.51% of perfectly predicted log statements). The other two pre-trainings only marginally improved the performance of the base model (see Table 3). In the following, we focus on the best-performing model. ...Context 10
... two-fold achievement (i.e., log level and position) suggests that LANCE can effectively support developers with logging activities. Looking at the third row in Table 3, it is instead clear that LANCE struggles to generate logging messages that are identical to the ones manually written by developers (success in 16.90% of cases). This difference in performance among the three log statement "components" to predict is kind of expected. ...Similar publications
Developing a software service requires a strict software development life cycle and process. This process demands controlling all application code through source control management as well as a rigorous versioning and branching strategy. However, the platform and infrastructure also benefit from this rigor. Software services must be deployed to a t...
