fix(import): preserve paths containing colons when parsing primary su…

[hazelmayank]

Summary

microcks import previously parsed the optional :true / :false suffix with strings.Split(input, ":")[0], which splits on the first colon. Any input containing more than one colon — including Windows-absolute paths such as C:\Temp\api.yaml — had its path truncated to the part before the first colon, and the rest was passed to strconv.ParseBool. The eventual upload failed with open C: no such file or directory, with no hint that the parser had silently eaten the path.

This PR fixes the parser to detect the boolean suffix from the right using strings.LastIndex + strconv.ParseBool, so paths containing colons (Windows-absolute paths, paths with multiple colons) are preserved while the documented :true / :false suffix continues to work exactly as before.

---

Why

• Windows users are blocked from using absolute paths with microcks import today.
• The failure mode (open C: no such file or directory) gives no hint that the parser truncated the path.
• All currently-documented inputs (./api.yaml, ./api.yaml:true, ./api.yaml:false, ./api.yaml:1, …) keep working identically after the fix — this is a strict superset of previous behavior.

---

What changed

• Extracted a small pure helper parseImportEntry(input string) (string, bool) in cmd/import.go.
• The helper inspects the substring after the last : and only treats it as mainArtifact when strconv.ParseBool succeeds; otherwise the entire input is treated as the path with mainArtifact = true.
• Replaced the inline if strings.Contains(f, ":") { ... } block in the import loop with a single call to the helper.
• Removed the now-redundant Cannot parse '...' as Bool warning, because the new logic treats a non-bool trailing segment as part of the path rather than a parse failure.
• Added a new test file cmd/import_test.go with table-driven cases for the parser.

---

Test plan

• go build ./...
• go vet ./...
• gofmt -l cmd/import.go cmd/import_test.go — no output
• go test ./cmd/... -run TestParseImportEntry -v — all table rows pass
• go test ./... — full sweep passes
• go test ./cmd/... -count=3 — passes (no state leakage)
• Manually verified on Windows: microcks import C:\Temp\api.yaml no longer fails with open C: no such file or directory
• Manually verified that ./api.yaml, ./api.yaml:true, ./api.yaml:false, ./api.yaml:1 all parse as before

---

Test cases covered

The table-driven test exercises:

Input	Path	mainArtifact
./api.yaml	./api.yaml	true
./api.yaml:true	./api.yaml	true
./api.yaml:false	./api.yaml	false
./api.yaml:TRUE	./api.yaml	true
./api.yaml:0	./api.yaml	false
C:\Temp\api.yaml	C:\Temp\api.yaml	true
C:\Temp\api.yaml:false	C:\Temp\api.yaml	false
C:\Temp\api.yaml:true	C:\Temp\api.yaml	true
path:with:colon.yaml	path:with:colon.yaml	true
path:with:colon.yaml:false	path:with:colon.yaml	false
./api.yaml:notabool	./api.yaml:notabool	true

---

Files changed

• cmd/import.go — parser refactored; strconv import retained for ParseBool.
• cmd/import_test.go — new file; first test in this package's import-command path.

---

Notes

• No flag changes; no behavior change for currently-working inputs.
• A similar bug exists in cmd/importURL.go for URLs that include a port (e.g. https://localhost:8080/api.yaml). It is intentionally left out of this PR to keep scope tight; happy to send a follow-up if maintainers prefer that shape.
• This is one of my early contributions to the repo; I am still learning Go and welcome any feedback on idiom or structure.

Closes #352

[Harishrs2006]

the LastIndex + ParseBool approach is the right way to handle this.

cmd/importURL.go has the same colon-splitting bug and isn't touched by this PR. Also, PR #291 is already open and covers both import.go and importURL.go together