Bugzilla – Full Text Bug Listing
| Summary: | PathMap service fixes for common usecases | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [Tools] TCF | Reporter: | Peder Andersen <Peder.Andersen> | ||||||
| Component: | Agent | Assignee: | Project Inbox <tcf.agent-inbox> | ||||||
| Status: | RESOLVED FIXED | QA Contact: | Eugene Tarassov <eugene> | ||||||
| Severity: | normal | ||||||||
| Priority: | P3 | CC: | cdtdoug, mober.at+eclipse | ||||||
| Version: | unspecified | ||||||||
| Target Milestone: | 0.6.0 | ||||||||
| Hardware: | PC | ||||||||
| OS: | Linux | ||||||||
| Whiteboard: | |||||||||
| Attachments: |
|
||||||||
Created attachment 208413 [details]
Proposed changes for identifier remap
I have committed the first patch. Thanks! I don't understand second patch. It looks incorrect to me. It checks if strlen(src) == strlen(dst), but it looks like intention was to check strlen(src) == strlen(fnm), which is same as fnm[k] == 0, which is already properly handled without the patch. You're right; I must have been confused between dst and fnm, *and* that the second patch generally seems unnecessary. So far as I'm concerned, this bug is now fixed, thanks! Changing to resolved/fixed. |
Created attachment 208412 [details] Proposed changes for rootfs remap Two usecases for the PathMap service don't work at the moment. First, users have a complete copy of the target's rootfs accessible via a host path, so we effectively want to map "target:/" to "host:/path/to/rootfs". Second, clients would like the ability to map a non-path identifier to an individual file, for example, to locate a target's kernel image (which typically does not reside on the target's filesystem).