Summary: | sys-apps/portage: emerge ignores optional build-time only dependencies when updating system | ||
---|---|---|---|
Product: | Portage Development | Reporter: | Chris Smith <chris> |
Component: | Unclassified | Assignee: | Portage team <dev-portage> |
Status: | CONFIRMED --- | ||
Severity: | normal | CC: | aditsu, adrin.jalali, esigra |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
See Also: |
https://bugs.gentoo.org/show_bug.cgi?id=475850 https://bugs.gentoo.org/show_bug.cgi?id=350230 https://bugs.gentoo.org/show_bug.cgi?id=635606 |
||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 300071, 598444, 926291 |
Description
Chris Smith
2012-07-24 19:30:17 UTC
The explanation is that the first command pulls in the bison-2.6 update because it traverses the dependencies in a different order. It's not an error because the <sys-devel/bison-2.6 dependency is build-time only, and therefore it's not really required unless you are going to rebuild chromium. So, it's behaving as intended, and I think we can close this bug with "WONTFIX" resolution. Never ran across that problem before. It's annoying as it is but what if there were 10 or 20 such packages? And they took a long time to build? Maybe it's the fault of the Chromium ebuild but wherever the fault lies it shouldn't work this way. (In reply to comment #1) > The explanation is that the first command pulls in the bison-2.6 update > because it traverses the dependencies in a different order. It's not an > error because the <sys-devel/bison-2.6 dependency is build-time only, and > therefore it's not really required unless you are going to rebuild chromium. > So, it's behaving as intended, and I think we can close this bug with > "WONTFIX" resolution. (In reply to comment #2) > Never ran across that problem before. It's annoying as it is but what if > there were 10 or 20 such packages? Normally, packages work with the highest available version of a given package, so it's not an issue most of the time. > And they took a long time to build? > Maybe it's the fault of the Chromium ebuild but wherever the fault lies it > shouldn't work this way. Considering how uncommon it is, I don't think it's a problem. Pretty soon, chromium will get fixed so that it works with newer bison. (In reply to comment #3) > Considering how uncommon it is, I don't think it's a problem. Pretty soon, > chromium will get fixed so that it works with newer bison. Isn't it up to portage to prevent such issues? Why doesn't a system update take world dependencies into consideration? Sorry - still seems like a portage flaw to me. Yeah, we can trigger backtracking in order to prevent the upgrade from occurring. *** Bug 486550 has been marked as a duplicate of this bug. *** *** Bug 658326 has been marked as a duplicate of this bug. *** |