Speed-up bash completion by using sql cache again #1100
Conversation
m-blaha
force-pushed
the
completion_sqlite
branch
from
5827d39
to
d1b3eb3
Compare
|
I am sorry, but it looks like that it doesn't work. Please can you provide a test case? |
|
Doesn't work for me either. Nothing happens when pressing tab while completing either a command or a package name. Tried to debug it a bit, it may be returning after |
etc/bash_completion.d/dnf
Outdated
| summary-updateinfo swap system-upgrade update updateinfo update-minimal update-to | ||
| upgrade upgrade-minimal upgrade-to versionlock whatprovides | ||
| " | ||
|
|
There was a problem hiding this comment.
I don't think this is supposed to be added in the previous commit and then removed here.
|
I'd like to rework the bash completion completely from the scratch. Basically, all commands and options should be generated from the argument parsers and not read from the bunch of hardcoded lists. Probably the best solution would be to close this PR and start a new one (which can hopefully cover also plugins instead of #1102) |
This came across my mind 😄 really curious how you'd do it (or perhaps there already is some infra in the argument parser to pull these from it?). Up to you then, if you don't want to continue with this just close the PR. If you're confident you'll make the argparser way work and will have a better solution for the plugins too, I guess close the #1102 as well. Thanks! |
|
☔ The latest upstream changes (presumably f68dcb2) made this pull request unmergeable. Please resolve the merge conflicts. |
We already maintain this sqlite cache by generate_completion_cache plugin.
- removed abbreviation, they are not used by dnf any more - removed hardcoded list of commands
m-blaha
force-pushed
the
completion_sqlite
branch
from
d1b3eb3
to
66d91a9
Compare
|
Just wanted to chime in, having noticed very slow bash completions myself and as this seems to be a fairly commonly reported issue: @m-blaha had mentioned a possible rewrite of the bash completion which would warrant putting this off, but as that was some time ago I'm not sure of the status of it. I did pull the branch down and it seems to still merge with master and function as intended. Are there any outstanding considerations that need to be addressed? |
|
Closed in favor of #1817 |
We already maintain this sqlite cache by generate_completion_cache plugin.