August 2013 - ghc-devs

SV: ./configure not using system MinGW
by Niklas Larsson 12 Aug '13

12 Aug '13

Hi! It's perfectly normal, to use that gcc. What it should do is to check out the ghc-tarballs repo and unpack a gcc from there. Now that I think of it, I had to manually get that repo last time. Maybe sync-all is broken? . Niklas ----- Ursprungligt meddelande ----- Från: "Kyle Van Berendonck" <kvanberendonck(a)gmail.com> Skickat: ‎2013-‎08-‎12 10:04 Till: "ghc-devs(a)haskell.org" <ghc-devs(a)haskell.org> Ämne: ./configure not using system MinGW Hi, I tried to build GHC on windows recently and for some reason it tries to use a MinGW in repopath/inplace/mingw. I had to copy my entire mingw installation there to get it to configure. This isn't normal, right? Regards.

1 0

./configure not using system MinGW
by Kyle Van Berendonck 12 Aug '13

12 Aug '13

Hi, I tried to build GHC on windows recently and for some reason it tries to use a MinGW in repopath/inplace/mingw. I had to copy my entire mingw installation there to get it to configure. This isn't normal, right? Regards.

1 0

Machine-readable nightly output
by Edward Z. Yang 11 Aug '13

11 Aug '13

Hey Geoffrey, The nightly builds are running on your box, correct? How hard would it be to modify the nightly scripts so that they also output machine-readable information about the build? For example, we could then setup a fingerprint database, where given some revision of head, you could find out if it built, and what the hashes of all the repositories it built with were. Edward

2 2

Successful build of GHC HEAD on OS X 10.9 Mavericks
by Luke Iannini 11 Aug '13

11 Aug '13

Hi guys, Just wanted to report on the steps/hacks to a successful build of GHC HEAD on OS X 10.9 thanks to help from Carter Schonwald and Austin Seipp. 1) I installed LLVM 3.4 brew install llvm --with-clang --HEAD 2) I used this build.mk: HADDOCK_DOCS = NO BUILD_DOCBOOK_HTML = NO BUILD_DOCBOOK_PS = NO BUILD_DOCBOOK_PDF = NO SPLIT_OBJS = NO INTEGER_LIBRARY = integer-simple 3) Austin let me know I needed to pass "-x assembler-with-cpp" to clang when in preprocessor mode, so I wrote a wrapper clang.hs, compiled it, put it in my path, and changed /usr/local/lib/ghc-7.6.2/settings to use it. (it also passes some flags Carter suggested, along with extra include flags, which were necessary as of a couple weeks ago but might have been fixed in the newer DPs) https://gist.github.com/lukexi/e11668ced04524564a8b 4) I used ./configure --with-gcc=clang (where clang is the wrapper script above) 5) "inplace/bin/mkdirhier" bootstrapping/. failed mysteriously, so I manually created bootstrapping/ in the ghc build dir. 6) DTrace failed with an "unsupported compiler" error during compilation, so I killed that section in configure.ac # if test -n "$DtraceCmd"; then # if test "x$TargetOS_CPP-$TargetVendor_CPP" = "xdarwin-apple" -o "x$TargetOS_CPP-$TargetVendor_CPP" = "xsolaris2-unknown"; then # HaveDtrace=YES # fi # fi 7) I got the error: libraries/integer-simple/GHC/Integer/Type.hs:836:50: Not in scope: '.' libraries/integer-simple/GHC/Integer/Type.hs:847:67: Not in scope: '.' I hacked through this by manually find-and-replacing WORD_SIZE_IN_BITS with 64 in Type.hs. make && make install then ran to completion and some simple test programs seem to compile and run fine. Cheers Luke

4 4

NOTICE: The Big Gitolite Switch
by Austin Seipp 09 Aug '13

09 Aug '13

Hello all, Following our proposal from about a week ago, Herbert and I plan to make the Gitolite migration happen Real Soon Now. If possible, we would actually like to have it happen tomorrow, although that's understandably a bit too short of a window of warning perhaps. However, commit activity has been low the past two weeks, so sometime in the next few days is probably the best window (Simon gets back August 15th, so we would like something in place by then.) We already have just about everything in place I believe. This is a final call for anyone who might need shell access or something of the sorts. I envision the plan will go like this: * We will turn off push access to the repositories, AND turn off shell access, by revoking SSH keys temporarily from everyone in the 'darcs' group (don't worry, there will be backups.) * I'll take everyone who has committed to any of the repositories in the last 1yr and add them to the new setup. That'll cover most of the current group, I think. There may be some people who were recently added but haven't committed yet. I'll try to remember everyone, but please yell if there is a problem. * We'll make the switch to gitolite by reinstating the SSH keys that are already there, and reopen the tree. We WILL NOT be deleting the existing user accounts. I don't want to possibly destroy something left there. Instead, if you would like, you can just email me and I'll re-add your key if necessary to get your stuff back. I expect this will be a very short and easy migration. Herbert has already given details of changing your existing trees via ./sync-all as well: http://www.mail-archive.com/ghc-devs@haskell.org/msg01781.html I will follow up with a reply to this email containing precise details on the switch, including the expected downtime window, and what time it will be happening. Please let me know if you have questions or concerns. -- Regards, Austin - PGP: 4096R/0x91384671

1 2

Ian leaving and the glorious future
by Simon Peyton-Jones 09 Aug '13

09 Aug '13

Dear ghc-dev friends, Ian has now officially left Well Typed, so there now isn't anyone whose day job it is to respond to GHC bug mail, Trac tickets etc. Well Typed are hiring, and there will be someone new in due course, but until then there will be a hiatus. In general, this is part of what I see as a long, slow movement: as time goes on, GHC is becoming more and more public property. Simon M's move to Facebook is another part of that movement. Increasingly, GHC is going to rely more on the community to maintain and develop it, and less on GHC HQ. And when I say "the community", I think firstly of the ghc-devs list :-). If the community is to be expected to do more of the work, they should also get more influence about decisions. I don't think there's any rush, but I'm open to forming a GHC steering committee (if you would like) to broaden the sense of "ownership". Don't worry: I'm not bowing out! But I am very much trying to encourage you to feel that GHC is *your* compiler, not just SPJ/SM's compiler that you can contribute to. I'm open to suggestions for how best to nurture broad participation. And meanwhile, in the current hiatus, any support that any of you can give is hugely welcome! Thanks. Simon

8 10

5144: Pattern synonyms -- help needed to sort out exports
by Dr. ERDI Gergo 08 Aug '13

08 Aug '13

Hi, Last week, I started implementing http://ghc.haskell.org/trac/ghc/ticket/5144 "Pattern synonyms". I'm using http://ghc.haskell.org/trac/ghc/wiki/PatternSynonyms for the details, and so far I have "pattern-only synonyms" working, with most of the infrastructure also laid down for the "simple pattern synonyms". My work is avalable at https://github.com/gergoerdi/ghc.git in the pattern-synonyms branch. My plan is to get the simple pattern synonyms stuff working as well, and then submit it as a first cut. Before I formally submit my patches for review to this mailing list, though, there is one thing that's got me stumped: allowing pattern synonyms to be exported. I thought it would just work if I make sure the pattern synonym names (which are Conids) end up in the tcg_exports field of the TcGblEnv. However, this is not the case. Could someone point me in the right direction please on how to make these new declarations exportable? Thanks, Gergo -- .--= ULLA! =-----------------. \ http://gergo.erdi.hu \ `---= gergo(a)erdi.hu =-------' All C programs are derived from `Hello World'.

1 0

Warnings when pushing to testsuite
by Jan Stolarek 08 Aug '13

08 Aug '13

When I'm pushing to testsuite repo I get lots of warnings: [t-jastol@cam-05-unx : ~/ghc-working/testsuite] git push origin master Counting objects: 13, done. Delta compression using up to 32 threads. Compressing objects: 100% (7/7), done. Writing objects: 100% (7/7), 783 bytes, done. Total 7 (delta 5), reused 0 (delta 0) remote: To git@github.com:ghc/testsuite.git remote: ccc29fb..67d6a19 master -> master remote: chmod: changing permissions of `/home/darcs/testsuite.git': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/packed-refs': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/applypatch-msg': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/pre-rebase': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/post-update': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/post-receive': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/git-notifier': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/pre-commit': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/git-notifier-OLD': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/samples': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/prepare-commit-msg': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/commit-msg': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/post-commit': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/update': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/pre-applypatch': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/hooks/post-receive-SAVE': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/HEAD': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/objects': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/objects/27': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/objects/27/d4d6e8924d3dabf66308378536c89cb8a9a4f8': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/objects/27/a92422f56493b0d98793f6ce02bf83b981c09b': Operation not permitted -- Lots more remote: chmod: changing permissions of `/home/darcs/testsuite.git/logs/refs/heads/at-defaults': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/config': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/branches': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/.git-notifier.dat': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/description': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/git-notifier.log': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/refs': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/refs/heads': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/refs/tags': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/info': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/info/exclude': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/info/refs': Operation not permitted remote: chmod: changing permissions of `/home/darcs/testsuite.git/ghc-notifier.log': Operation not permitted To ssh://jstolarek@darcs.haskell.org/srv/darcs/testsuite.git 9ba44e3..67d6a19 master -> master Notice that warnings refer to github: remote: To git@github.com:ghc/testsuite.git remote: ccc29fb..67d6a19 master -> master Any idea why this happens? I don't get warnings when pushing to main GHC repo. Janek

1 0

s/libraries/packages/g in ghc.git/.gitmodules
by Herbert Valerio Riedel 07 Aug '13

07 Aug '13

Hello GHC devs, in the current ghc.git's .gitmodules file, the entries look like ,---- | [submodule "libraries/binary"] | path = libraries/binary | url = http://darcs.haskell.org/libraries/binary.git/ `---- I'd like to normalize the urls above to use the http://darcs.haskell.org/packages/ scheme instead to match what's used in the 'packages' file as remotepaths. (the way I understand it, the /libraries/ folder is an historical artifact, and should have been obsolete for a couple of years by now -- so the new gitolite setup won't support the /libraries/ alias anymore) ...is there any reason you can think of, I should not perform this modification to the .gitmodules file? cheers, hvr

1 0

Error on push to GHC git repo due to Trac?
by Erik de Castro Lopo 07 Aug '13

07 Aug '13

Hi all, Got the error below. Erik bash: warning: setlocale: LC_ALL: cannot change locale (en_AU.UTF-8) Counting objects: 5, done. Delta compression using up to 4 threads. Compressing objects: 100% (3/3), done. Writing objects: 100% (3/3), 412 bytes | 0 bytes/s, done. Total 3 (delta 2), reused 0 (delta 0) remote: /bin/bash: warning: setlocale: LC_ALL: cannot change locale (en_AU.UTF-8) remote: To git@github.com:ghc/ghc.git remote: 9082111..0daee29 master -> master remote: Traceback (most recent call last): remote: File "./hooks/post-receive-trac", line 202, in <module> remote: handle_ref(env=env, *line.split()) remote: File "./hooks/post-receive-trac", line 163, in handle_ref remote: db = env.get_db_cnx() remote: File "/usr/lib/python2.7/dist-packages/trac/env.py", line 329, in get_db_cnx remote: return get_read_db(self) remote: File "/usr/lib/python2.7/dist-packages/trac/db/api.py", line 90, in get_read_db remote: return _transaction_local.db or DatabaseManager(env).get_connection() remote: File "/usr/lib/python2.7/dist-packages/trac/db/api.py", line 152, in get_connection remote: return self._cnx_pool.get_cnx(self.timeout or None) remote: File "/usr/lib/python2.7/dist-packages/trac/db/pool.py", line 227, in get_cnx remote: return _backend.get_cnx(self._connector, self._kwargs, timeout) remote: File "/usr/lib/python2.7/dist-packages/trac/db/pool.py", line 147, in get_cnx remote: raise TimeoutError(errmsg) remote: trac.db.pool.TimeoutError: Unable to get database connection within 0 seconds. (TracError: The user erikd requires read _and_ write permissions to the database file /home/trac/ghc/db/trac.db and the directory it is located in.) To darcs.haskell.org:/srv/darcs/ghc.git 9082111..0daee29 master -> master -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/

2 1