Skip to contents

Ideally, I would have preferred to further develop loggit, as I consider the concept highly commendable and worthy of recognition. Unfortunately, this was not possible because the maintainer was unresponsive to messages or pull requests at the time of the fork (the last release was over three years ago). Additionally, CRAN’s rules do not permit the takeover of the package without the maintainer’s consent as long as the package passes the checks.

Therefore, I ultimately decided to create a fork (loggit2) to address some fundamental issues and implement necessary improvements. In the following, I will discuss the most pressing points.

Call of conditions

In loggit, all “condition log handlers” (loggit::message(), loggit::warning(), loggit::stop()) presented themselves as the call of the condition. This, combined with the fact that loggit does not allow passing a condition to the “condition log handlers” (e.g., in tryCatch()), made it significantly more difficult to trace the origin of a condition.

loggit::message("This is a message")
#> {"timestamp": "2022-04-12T10:55:02-0500", "log_lvl": "INFO", "log_msg": "This is a message"}
#> This is a message
loggit::warning("This is a warning")
#> {"timestamp": "2022-04-12T10:55:02-0500", "log_lvl": "WARN", "log_msg": "This is a warning"}
#> Warning in loggit::warning("This is a warning"): This is a warning
loggit::stop("This is an error")
#> {"timestamp": "2020-05-31T20:59:33-0500", "log_lvl": "ERROR", "log_msg": "This is an error"}
#> Error in loggit::stop("This is an error"): This is an error

In loggit2, each condition handler shows the same call as its base equivalent.

base::message("This is another message")
#> This is another message
loggit2::message("This is another message")
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "INFO", "log_msg": "This is another message\n"}
#> This is another message

base::warning("This is another warning")
#> Warning: This is another warning
loggit2::warning("This is another warning")
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "WARN", "log_msg": "This is another warning"}
#> Warning: This is another warning

base::stop("This is another error")
#> Error in eval(expr, envir, enclos): This is another error
loggit2::stop("This is another error")
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "ERROR", "log_msg": "This is another error"}
#> Error in eval(expr, envir, enclos): This is another error

For further information and comparisons look here.

Missing stopifnot()

In loggit, there are no “condition log handlers” (i.e., functions that work like their base equivalents but also write to the log) for base::stopifnot(). This makes the migration of existing code (without logs) unnecessarily complicated.

In loggit2, this has been added:

base::stopifnot("TRUE is not true" = TRUE, "This is an error" = 3L < 1L, "This is another error" = FALSE)
#> Error: This is an error
loggit2::stopifnot("TRUE is not true" = TRUE, "This is an error" = 3L < 1L, "This is another error" = FALSE)
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "ERROR", "log_msg": "This is an error"}
#> Error: This is an error

Log entries containing :

In loggit, log entries containing : are not correctly parsed by loggit::read_logs(). In the best case, the log messages are truncated after the :. See also here. This bug is fixed in loggit2.

Clearing the log

In loggit, clearing the log with loggit::rotate_logs(0L) results in the log being unreadable by loggit::read_logs(), even with new entries. See also here. This bug is fixed in loggit2.

Conclusion

In addition to these issues, there are several others, but these are the most pressing and can all be easily fixed.

Apart from the aforementioned bugs and missing features, there is significant untapped potential in the control settings (e.g., global settings regarding log levels) that can be implemented without complicating loggit.

All of these aspects will be addressed in loggit2. I welcome any suggestions for new features, bug reports, or even a direct pull request on either topic.

Further comparisons

In loggit, each condition handler shows itself as the call of the condition.

f <- function() {
  loggit::message("This is another message")
  loggit::warning("This is another warning")
  loggit::stop("This is another error")
}

f()
#> {"timestamp": "2022-04-12T10:55:02-0500", "log_lvl": "INFO", "log_msg": "This is a message"}
#> This is a message
#> {"timestamp": "2022-04-12T10:55:02-0500", "log_lvl": "WARN", "log_msg": "This is a warning"}
#> Warning in loggit::warning("This is a warning"): This is a warning
#> {"timestamp": "2020-05-31T20:59:33-0500", "log_lvl": "ERROR", "log_msg": "This is an error"}
#> Error in loggit::stop("This is an error"): This is an error

In loggit2, each condition handler shows the same call as its {base} equivalent.

f <- function() {
  base::message("This is another message")
  base::warning("This is another warning")
  base::stop("This is another error")
}

f()
#> This is another message
#> Warning in f(): This is another warning
#> Error in f(): This is another error

f <- function() {
  loggit2::message("This is another message")
  loggit2::warning("This is another warning")
  loggit2::stop("This is another error")
}

f()
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "INFO", "log_msg": "This is another message\n"}
#> This is another message
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "WARN", "log_msg": "This is another warning"}
#> Warning in f(): This is another warning
#> {"timestamp": "2024-06-08T10:46:21+0000", "log_lvl": "ERROR", "log_msg": "This is another error"}
#> Error in f(): This is another error