I am still skeptical about fsck. Can you time it on a large repo? The
Allura repo might be a good start but there are much much bigger ones out
there too.
In principle I don't like doing extra work to handle a corner case when
try/except could too. But if it is fast all the time it could be ok.
--
Dave Brondsema
Principal Software Engineer - sourceforge.net
---
** [tickets:#7873] Git branch & tag speedups**
**Status:** in-progress
**Milestone:** unreleased
**Labels:** performance sf-current sf-4
**Created:** Fri Apr 17, 2015 09:23 PM UTC by Dave Brondsema
**Last Updated:** Wed Jun 10, 2015 05:20 PM UTC
**Owner:** Heith Seewald
I saw that git pages on forge-allura.apache.org were slow, so I looked at
stats.log and saw that the sidebar was the slowest part. I did some additional
digging and found 2 specific areas for improvement:
* in `git_main.py` change `default_branch_name` to a `@LazyProperty` since it
is called many times inside a loop in `RepositoryApp.sidebar_menu`
* Since `sidebar_menu` only requests a certain number of branches, pass that
"limit" all the way through to `git_repo.py`'s `branches` method so that
`is_valid()` is only called a minimum number of times needed.
* make sure the default branch logic to put it at the top of the list still
works (e.g. always put it at the top first)
In addition to those changes, generalize and apply the same approach to the
tags. And also check ForgeHg to see if mercurial can benefit the same way.
---
Sent from forge-allura.apache.org because [email protected] is subscribed
to https://forge-allura.apache.org/p/allura/tickets/
To unsubscribe from further messages, a project admin can change settings at
https://forge-allura.apache.org/p/allura/admin/tickets/options. Or, if this is
a mailing list, you can unsubscribe from the mailing list.