GHC Weekly News - 2015/03/24
austin at well-typed.com
Wed Mar 25 02:21:39 UTC 2015
(This post is available online with hyperlinks at
It's time for the GHC weekly news. We've had an absence of the last
one, mostly due to a lot of hustle to try and get 7.10 out the door
(more on that shortly throughout this post). But now we're back and
everything seems to be taken care of.
This week, in the wake of the GHC 7.10 release (which is occuring EOD,
hopefully), GHC HQ met up for a brief chat and caught up:
- This week GHC HQ met for only a very short time to discuss the
pending release - it looks like all the blocking bugs have been fixed,
and we've got everything triaged appropriately. You'll hopefully see
the 7.10 announcement shortly after reading this.
We've also had small amounts of list activity (the past 6-8 weeks have
been very, very quiet it seems):
- Your editor sent out a call for developers to fill information in
on the Status page about what they plan to do. If you're working on
something, please add it there!
- Herbert Valerio Riedel asked about a possible regression regarding
identifiers containing unicode subscripts - but nobody has replied!
- Doug Burke chimed in as a new contributor and wrote down some
notes on what it took to write his first patch and submit it to us -
and we really appreciate the feedback, Doug!
- Yitzchak Gale revived a thread he started a while back, which
puttered out: bootstrapping GHC 7.8 with GHC 7.10. The long and short
of it is, it should just about work - although we still haven't
committed to this policy, it looks like Yitz and some others are quite
adamant about it.
- Neil Mitchell uncovered a nasty bug in GHC 7.10.1 RC3, submitted
it to us. He also wrote a fantastic
post explaining the issue]. And it was promply diagnosed, fixed, and
taken care of by our own Joachim Breitner. Thank you for the fast
response Joachim and Neil!
- Mark Lentczner has announced Alpha releases of the Haskell
Platform 2015.2.0.0, containing GHC 7.10.1 RC3:
- Simon Peyton Jones asks: what's the current state about having
simultaneous installations of a package? Simon is a bit confused as to
why this is still a problem when we have all the tools to solve it, it
looks like! (But read on for more):
- Michael Snoyman asks: can we get a new feature patch in GHC
7.10.2? The answer seems to be an unfortunate 'no', but with some
tips, Michael may end up backporting the changes from HEAD to GHC 7.10
Some noteworthy commits that went into `ghc.git` in the past two weeks include:
- Commit 71fcc4c096ec0b575522e4c2d0104ef7a71a13c5 - GHC defaults to
using the `gold` linker on ARM/Android and ARM/Linux targets.
- Commit 9dfdd16a61e79cb03c633d442190a81fe5c0b6b8 - Bump `ghc-prim`
to version 0.4.0.0.
- Commit 42448e3757f25735a0a5b5e2b7ee456b5e8b0039 - GHC HEAD now
always looks for LLVM 3.6 specifically.
Closed tickets this past week include: #9122, #10099, #10081, #9886,
#9722, #9619, #9920, #9691, #8976, #9873, #9541, #9619, #9799, #9823,
#10156, #1820, #6079, #9056, #9963, #10164, #10138, #10166, #10115,
#9921, #9873, #9956, #9609, #7191, #10165, #10011, #8379, #10177,
#9261, #10176, #10151, #9839, #8078, #8727, #9849, #10146, #9194,
#10158, #7788, #9554, #8550, #10079, #10139, #10180, #10181, #10170,
#10186, #10038, #10164, and #8976.
Austin Seipp, Haskell Consultant
Well-Typed LLP, http://www.well-typed.com/
More information about the ghc-devs