Also, don't fetch a github user or their token when both are known. This
currently only affects the Github Token auth type. Github App
installations will continually fetch tokens every time we clone a repo.
In the future we should check the `ExpiresAt` field of the Github App
token and determine if we need to fetch a new one at that point.
* Add SSH config option for the git source
The auth message is empty since we use the git binary underneath to
handle the SSH authentication.
* Import digitaloceanv2
I'm not sure I fully understand why this issue exists. But I think the
short version is this: When we attempted to paginate users, we would set
a variable's Page value. But that variable appears to not actually be a
pointer, despite being added as one. It probably has to do with how
struct embedding works. Either way, if we make the overall options
variable the whole thing, and update its embedded struct with our page
variable, everything works out.
* Handle errors w/ github source.
* Fix loop var captured by func literal.
* Fix loop var captured by func literal.
* Set completed progress if the scan completes with no errors.
* Set progress to 100% if the scope and iteration are both 0.
* Fix commentary.
* Fix test.
* Return after the defer to os.RemoveAll.
* Fix unauth scan.
* Inline range loop.
* update tests for partial scan completion with errors. Ensure correct progress is set.
* Update progress for all sources.
* Update github test.
* Address comments.
* Scan binary files for git sources
* Create data chunks in for loop
* Linter feedback and newline commit result
* Use disk buffered reader and chunker function
* Use a config struct when scanning and engine source.
* fix tests.
* Move test_helpers to the sources pkg.
* Handle ScanGit error in tests.
* adderss comments.
* Use functional options.
* Remove temp var.
* Add better var names for the setup functions for each config.
* Remove unused var.
* fix error logs.
* fix error logs.
* single line.
* remove blank lines.
The fragment trace was a bit too verbose even at the trace level. We may
want to trace the file being chunked or something like that, but not the
entire diff.
