Posted by dww on October 11, 2009 at 6:57am
| Project: | CVS deploy |
| Version: | 7.x-1.x-dev |
| Component: | Code |
| Category: | task |
| Priority: | critical |
| Assigned: | dww |
| Status: | closed (fixed) |
Issue Summary
#597484: Use the Queue API to fetch available update data is going to break cvs_deploy attempting to get info about HEAD releases, since we no longer stash everything in a single cache entry in {cache_update}. Once that lands, we need to fix what we're doing in here.
The memory intensive solution is to grab everything, like we do now.
The query intensive solution is to only query individual projects that have a HEAD tag one by one as we need them.
Comments
#1
For people trying to test #597484, here's the simple patch for the memory-intensive option. ;) We can consider the query-intensive approach if/when #597484 lands.
#2
Dave Reid and I both think we should probably go fine-grained and solve this problem with more queries and less RAM. However, now that #597484: Use the Queue API to fetch available update data is in, cvs_deploy is currently broken with D7. So, getting this working again is critical. ;) Therefore, I committed #1 to HEAD as a stop-gap measure. We can revisit this as we work on #238950: Meta: update.module RAM consumption.
#3
Automatically closed -- issue fixed for 2 weeks with no activity.