[tpm] This month: LWP hack night
olaf.alders at gmail.com
Mon Mar 13 09:55:31 PDT 2017
> On Mar 12, 2017, at 4:01 PM, arocker at Vex.Net wrote:
>> to LWP::UserAgent, WWW::Mechanize, URI, etc. I now have co-maint on all
>> of these modules,
> I've just had a look at the 3 you mentioned. 2 have tickets in
> rt.cpan.org, while the other appears to have only github items. If that's
> not important, I can wait for an explanation. :-)*
That's just because internally we haven't committed to moving everything over to Github yet. It's a bit confusing. I'd like to see everything on Github. Maybe I'll make another push and see if we can get that done before our hack night.
> It strikes me that the first necessary step with such overwhelming lists
> is a spot of triage, possibly along axes of importance and ease of
> implementation? Maybe even also a third dimension of age? E.g. I can't see
> a lot of point investigating a failure on Windows XP. If a ticket is very
> old, does it mean it really needs to be fixed, or nobody's cared for so
> long that it can be ignored, (and preferably, deleted)?
> How can we get the swamp(s) down to manageable dimensions?
You raise a good point. Some kind of plan of attack would be good. Personally, I've just been looking at low hanging fruit, like stuff that can obviously be closed or that has an easy fix. Most of these modules are used so widely that adding new dependencies is done with extreme caution.
A lot of the outstanding pull requests are quite old, so they need someone to rebase them against the current master and, in a lot of cases, add appropriate tests. What happened in the past was that the pull requests were sent but nobody every replied to them (this was before my time). Now it's too late to get the author to provide a test because s/he has in many cases moved on to something else.
As far as which OS is or is not supported any longer, I'd need to get buy-in from others as to where support could be dropped.
We only have a couple of hours at this month's meeting, so I think we could probably focus on getting rid of more low hanging fruit, like reworking existing pull requests or adding tests etc. In some cases the bugs may have already been fixed, but the ticket still needs to be closed. Some are wish list requests which may or may not need to be closed.
Basically, there's a lot of housekeeping that still needs to be done, so even making a bit of progress here would be amazing.
More information about the toronto-pm