You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Followup to #16 which introduced the --err-skip-line-length option to totally disable the "very long lines" error:
--err-skip-line-length is useful to avoid being spammed when many files ave too large lines to be searched.
However, it's dangerous to use this option: it completely silence the fact that some files were skipped because of their size.
To fix this, I suggest to print a one line error, at the end of the sift execution, when at least one file was skipped because of its size (maybe with a skipped file counter).
This would both avoid the spam (one line is not a spam, and if it is, we could have another option to really disable all of this, as it's currently done), and still inform the user that some files have been skipped: if the user wants full results, it could disable the option, or change the blocksize limit, as explained in the current error log.
The text was updated successfully, but these errors were encountered:
Thanks for your thoughts on this, I think this is a great idea!
I am just not sure whether this should get implemented as
a new option (something like --err-sum-line-length)
the new default, with a new option like --err-show-line-length to show all errors, one line per file
I got the impression over the last months that some users are not happy with the current behavior, so I might favor the second variant as that would still adhere to sift's "search everywhere" design principle and inform the user about possible problems and solutions.
Followup to #16 which introduced the
--err-skip-line-length
option to totally disable the "very long lines" error:--err-skip-line-length
is useful to avoid being spammed when many files ave too large lines to be searched.However, it's dangerous to use this option: it completely silence the fact that some files were skipped because of their size.
To fix this, I suggest to print a one line error, at the end of the
sift
execution, when at least one file was skipped because of its size (maybe with a skipped file counter).This would both avoid the spam (one line is not a spam, and if it is, we could have another option to really disable all of this, as it's currently done), and still inform the user that some files have been skipped: if the user wants full results, it could disable the option, or change the blocksize limit, as explained in the current error log.
The text was updated successfully, but these errors were encountered: