Code review comment for lp://staging/~tseaver/loggerhead/sphinxify

Matt Nordhoff wrote:
> Tres Seaver wrote:
> > I was trying to document what effect the option actualy has, by
> > spelunking the code. From my exploration, it appears that the option
> > has a somewhat misleading name: it doesn't appear to make the
> > Loggerhead server willing to accept any changes from clients (in fact,
> > in my experience, I have to serve a "raw" bzr tree from another URL to
> > allow even a checkout).
> >
> > At any rate, I might be completely wrong about the intended (or actual)
> > semantics. I do know that passing '--allow-writes' on the command line
> > was the only way I found to get the branch UI to show the block of HTML
> > which gives directions for checking out the branch.
>
> --allow-writes is supposed to simply allow pushing to the smart server,
> and have no impact on the "bzr branch" example.
>
> This wouldn't be the first time it's gotten broken, though...
>
> The "bzr branch" example should always be displayed, unless you're
> serving a remote location (e.g. "serve-branches bzr+ssh://...") and the
> branch has no public_branch setting. Hmm, looking at the code, I may
> have already spotted an issue, though.

Following up on this, I sent a merge proposal to fix the "To get this branch" bit, and I tested --allow-writes and it's working as expected.