| Summary: | [patch] `emerge -u portage` restart fails with "emerge: please tell me what to do." | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Zac Medico <zmedico> |
| Component: | [OLD] Core system | Assignee: | Portage team <dev-portage> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | betelgeuse, gentoo |
| Priority: | High | Keywords: | InVCS |
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | |||
| Bug Blocks: | 115839 | ||
| Attachments: | emerge-restart-resume.patch | ||
|
Description
Zac Medico
2006-01-05 22:07:30 UTC
Created attachment 76306 [details, diff] emerge-restart-resume.patch This patch uses the list of long options stored in "myopts" to generate a new argument list for the os.execv call. As before, selected "badlongopts" are filtered out. In addition, "--resume" is added to the arguments and the first item in the mtimedb mergelist (portage itself) is removed. Hopefully this will have the desired behavior while being less error prone than the previous approach. The patch applies to all versions of portage currently in the tree (2.0.51.22-r3 through 2.1_pre3-r1). Note that this bug is actually unrelated to bug 107865 (though symptoms initially led me to suspect they were related). Thinking about it, --skipfirst should be added to badlongopts (or equivalently, leave portage in the mergelist and force --skipfirst intentionally). One drawback of my suggested approach is that it assumes the new version of portage will be able to read the mtimedb that was created by the old version (good assumption?). This is now in trunk for release in portage-2.1_pre5. *** Bug 124163 has been marked as a duplicate of this bug. *** Released in 2.1_pre5. |