launchpad-reviewers team mailing list archive

[Colin Watson <cjwatson@xxxxxxxxxxxxx>]

On Tue, Nov 13, 2012 at 05:49:20AM -0000, William Grant wrote:
> Unless I'm forgetting something, this eliminates the penultimate
> delayed copy callsite, the last being the API's Archive.syncSource(s).
> Do you have a strategy for the final excision of that callsite and the
> significant volume of called code?

Not only a strategy but a pending branch that I'll be ready to submit
shortly after this lands.

> This will probably double escape them; structured() escapes parameters
> itself. I'd also avoid talking about the escaping in the docstring, as
> one wouldn't normally expect a string to be passed through unescaped
> unless the docstring explicitly asked for HTML.

This mistake was because compose_synchronous_copy_feedback used
structured() in a slightly different way (with %-formatting rather than
passing separate replacement arguments, so structured() wouldn't have
handled escaping there).  Fixed.

> What does the logging around this look like? AFAICT the job doesn't
> log much else, so it may seem somewhat of a non sequitur in the
> output.

I added an extra line of logging to alleviate confusion.  (Of course, if
we actually want to see it routinely we'll need to change the PCJ runner
to log at DEBUG; should this be at INFO instead, or is it fine as it
is?)

-- 
https://code.launchpad.net/~cjwatson/launchpad/always-copy-packages-asynchronously-2/+merge/131928
Your team Launchpad code reviewers is subscribed to branch lp:launchpad.