On Thu, Oct 21, 2010 at 5:58 PM, Dun Peal <[email protected]> wrote:

> > I won't ask...
>
> We tag each release. We follow "release early, release often". So, we
> have many tags.
>

Ah, okay, thanks for telling even if I didn't ask then :-)

> I suppose the right thing to do is to only display a subset of the tags,
> > which would require a UI to search/browse the tags.
>
> Yes. Would you accept a patch that does that?
>

Sure. One way to solve it could be to use a folder-like convention for tags.
For instance I could create the tag releases/v1 - and we could treat these
as "folders" in the UI and add functionality for listing all tags within one
of these "folders". Just an idea I got, does it make any sense?


> Hm, can you clarify?  The tags are physically stored in one long text
> file: `.git/packed-refs`. So for us, that file does contain several
> tens of thousands of lines, but I'm not sure that would be a
> performance problem, seeing that it's just reading a file with a few
> tens of thousands of lines, which shouldn't tax modern hardway.
>

>From the top of my mind, Grit (the Ruby/Git library we're using) will create
a Tag object for each tag that's fetched, which would mean a few thousand
object instantiations for each request - this may or may not have a
performance penalty. If we choose to go with the folder convention, we could
have the tag fetching code accept a parameter that filters which Tag objects
get instantiated - but it may not be a problem at all. At least one that
caching the tags wouldn't solve.

Cheers,
- Marius

-- 
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]

Reply via email to