Problematic file names

[danielmlow]

@sanuann , the files that submitted at the end of a protocol are currently saved like this (bottom files in white):

having characters after .zip can cause problems for some people when trying to unzip. could we change it to

test_debbie_01-001.zip
test_debbie_01-002.zip
test_debbie_01-003.zip
test_debbie_01-004.zip
etc?

So that they end in .zip and are easy to unzip?

[sanuann]

i don't remember why exactly it was done that way. @satra is it okay to change it to the format daniel mentioned above?

[satra]

isn't that the multi part upload? in which case the zip file should be constructed by concatenating the parts.

[satra]

one should not be trying to unzip those files, just concatenating them. and the multipart upload should also apply potentially to activity uploads no?

[sanuann]

yes, they are the multipart uploads. for the activity uploads, I chose not to use multipart upload. do we need it by default for any uploads?

[satra]

do we need it by default for any uploads?

yes, since any activity could generate a lot of data, especially if those are voice recordings. it should only be a single upload if the size is smaller than the multipart upload threshold.

[danielmlow]

other people using reproschema in the future may not be as tech savvy, so that’s why i think it’s worth changing it so that it is easy to unzip or combine, but fine as is if we provide users with instructions on how to combine

[satra]

@danielmlow - they are chunked uploads, they cannot be unzipped on their own. they need to be combined to create a zip file. this has nothing to do with tech savviness just that information is being uploaded in parts.

@sanuann - if you know the number of chunks in advance, may be useful to add that .001-010 for example if that's chunk 1 of 10 chunks.

[danielmlow]

What's the correct way of merging?

from https://unix.stackexchange.com/questions/40480/how-to-unzip-a-multipart-spanned-zip-on-linux :

"you need to first concatenate the pieces, then repair the result. cat test.zip.* concatenates all the files called test.zip.* where the wildcard * stands for any sequence of characters; the files are enumerated in lexicographic order, which is the same as numerical order thanks to the leadings zeroes. >test.zip directs the output into the file test.zip."

cat test.zip.* >test.zip
zip -FF test.zip --out test-full.zip
unzip test-full.zip

"If you created the pieces by directly splitting the zip file, as opposed to creating a multi-part zip with the official Pkzip utility, all you need to do is join the parts."

cat test.zip.* >test.zip
unzip test.zip