A filename with a space within, for example "spaced filename.txt" (contains a tab character in the filename),
and a hash digest, for example, "12345", would be recorded in the
file-list as:
spaced filename.txt 12345
The current parser would determine that the <filename> is "spaced"
and the <sha> is "filename.txt". While this would not halt the
process, it would make the diffs-generation assume that the file has
been changed, which is incorrect.
We need to fix this bug:
Approach A:
An approach would be to quote the filename in the file-list. For example,
"spaced filename.txt" 12345
But how to handle the " character in the filename? For example,
"comma\\"filename.txt" 12345
Approach B:
We could change the format, from <filename> <hash> to <hash> <filename>.
This is quite easier to implement. We are sure the <hash> can not contain
a space/tab character.
How to handle new input, with embedded space characters, being added to the
format.
A filename with a space within, for example
"spaced filename.txt"(contains a tab character in the filename),and a hash digest, for example,
"12345", would be recorded in thefile-list as:
The current parser would determine that the
<filename>is"spaced"and the
<sha>is"filename.txt". While this would not halt theprocess, it would make the diffs-generation assume that the file has
been changed, which is incorrect.
We need to fix this bug:
Approach A:
An approach would be to quote the filename in the file-list. For example,
But how to handle the
"character in the filename? For example,Approach B:
We could change the format, from
<filename> <hash>to<hash> <filename>.This is quite easier to implement. We are sure the
<hash>can not containa space/tab character.
How to handle new input, with embedded space characters, being added to the
format.