Call to action
H. S. Teoh
hsteoh at quickfur.ath.cx
Thu Sep 24 19:00:15 UTC 2020
On Thu, Sep 24, 2020 at 06:09:42PM +0000, bachmeier via Digitalmars-d wrote:
> On Thursday, 24 September 2020 at 15:06:25 UTC, novice2 wrote:
[...]
> > But from other side, then read forum and see "ah, this this bug, i made
> > PR at 2009" and see then it still not merged... may be my language
> > barrier prevent me from understand this...
>
> Those are typically bigger PRs. When I submit small things like
> documentation or an update to a unit test, they're merged within a day
> or a few days at the most.
Yes, I recently submitted a couple of small PRs that got merged pretty
quickly. The ones that get stuck generally have one or more of the
following challenges:
- The PR fails to pass the autotesters and CIs, and the submitter is
slow to respond and/or not responding, so there's not much we can do
about it;
- The changeset is large, therefore hard to review, so most reviewers
(who are volunteers with limited free time) are less inclined to
review them;
- The change requires expertise in a particular area to review correctly
(e.g., deals with tricky concurrency issues, or OS details, or tricky
pointer/memory manipulations or involves some complex spec like
parsing some DSL, or floating-point algorithms that are tricky to
write in a way that won't risk catastrophic loss of precision or wrong
behaviour in corner cases). If the area is too specialized, many
reviewers feel unqualified to review it (probably rightly so), so it
tends to get less attention that it should.
- The PR introduces some controversial bits of code that one or more
reviewers may not agree on, or it breaks existing code in some way, or
requires approval from the leadership (which generally takes a long
time, and unfortunately requires a lot of persistence on the part of
the submitter).
T
--
MAS = Mana Ada Sistem?
More information about the Digitalmars-d
mailing list