Problem was that we were not globally replaceing `\\` with `\`.
Also, this PR adds some new tests to sarif-utils.ts. In doing so, we
have fixed a small bug in getPathRelativeToSourceLocationPrefix.
Previously, we were uri decoding the sarifRelativeUri. However, this is
no longer correct because the result is another URI and it should
remain encoded if it originally was.
Resolves#585
When retrieving paginated results, need to make sure we are getting
page offsets from the correct results file.
Previously, we were incorrectly extracting page offsets from the default
(unsorted) file. With this change, we ensure that we get offsets from
the proper results file when there is a request for a page of results.
* It is now more general and the logic is simplified
* Also, add more comments
* Rename `adaptBqrs` to `transformBqrsResultSet`
* Remove a react error for missing a key attribute in a list
This refactoring combines the types in `bqrs-types.ts` and
`bqrs-cli-types.ts`. Historically, the former was used for BQRS files
parsed by the extension and the latter for BQRS files parsed by the cli.
They describe the same file types, but using different property and type
names.
We have moved to parsing all BQRS files by the cli. This refactoring
removes the `bqrs-types.ts` file and replaces all BQRS references to
use types in `bqrs-cli-types.ts`.
Additionally, the `adapt.ts` file has been deleted since its purpose
was to convert between extension and cli BQRS types. Some one type and
one function from `adapt.ts` has been moved from `adapt.ts` to
`bqrs-types.ts`. It's possible that we want to do a further refactoring
to simply remove them both.
A common situation when a file is not relevant for a particular result
is to return an empty file path location.
Currently, we are displaying this situation as a hyperlink in the
results, but when clicking on the link, there is an error.
To mirror the behaviour of Eclipse, we should avoid showing a link here.
This commit changes that behaviour.
Fixes#543
```ql
newtype TA = TB()
private predicate foo() { any() }
```
Our TextMate grammar didn't realize that the newtype declaration ended after the closing paren of the branch's parameter list, so the `private` modifier was highlighted incorrectly.
It's surprisingly tricky to get TextMate to handle this correctly, so I wound up just treating the IPA declaration head (`newtype TA`), the branch head (`= TB`), the branch parameter list, and the branch body as directly children of the module body. This is kind of hacky, but it does fix the bug without introducing any new cases where we have incorrect highlighting of valid code.
This leads to less sharing of codepaths which is a little bad (slightly more
repetition and rendundancy) but a lot good (can independently fix the way
raw results are redisplayed so as to be actually correct).
