Closed irees closed 4 years ago
A couple other notes: 1. it runs the delete before checking if the -zip
path is present or valid gtfs 2. it's a recursive delete of the entire tree.
Thanks for the observation @irees, MobilityData will be exploring this.
Hey @irees ,
-input
path will exist if the validator has been run previously. Not sure we should fail in that case?
@fabrice-v I would suggest considering re-naming the -input
to something different, as I think it's confusing to users. I think most end users would view what it does (serve as unzipped location of a GTFS zip file) as a cache or working directory, not as input to the application. The user-facing "input" is really the -zip
param.
That is a good point indeed. Shall we keep the possibility to specify where said cache is created which was the original intent or simplify it to something like 'an input
folder will be created with the content of the provided dataset' ?
@fabrice-v I think it's a good idea to keep the cache location configurable via command line, as these files can get huge and users may not want to unzip to the same drive as where they are running the application.
Alright here is how I propose to move forward
-zip
to -inputzip
-input
to -extract
Thoughts?SGTM!
Alright here is how I propose to move forward
- rename
-zip
to-inputzip
- rename
-input
to-extract
Thoughts?
This had been implemented in d55683b78aa905bf0b8ad186ae66a4483898928a, see linked PR
@lionel-nj Could you update the release docs in that PR too? https://github.com/MobilityData/gtfs-validator/blob/master/RELEASE.md
Sure, working on it.
Note that in the final version of the fix -zip
was renamed to -zipinput
Describe the bug I mistakenly used
-input
to refer to my zip file, instead of-zip
, which results in the file specified by-input
to be immediately deleted.To Reproduce
Expected behavior Fail when
-input
path exists.Witnessed behavior See above
Screenshots See above
Environment used
openjdk 14 2020-03-17