- cabal-devel - Haskell.org

Re: [Hackage] #626: Package build error due to unicode
by Hackage 18 Oct '11

18 Oct '11

#626: Package build error due to unicode --------------------------------+------------------------------------------- Reporter: basvandijk | Owner: Type: defect | Status: reopened Priority: normal | Milestone: Component: hackageDB website | Version: 1.6.0.1 Severity: normal | Resolution: Keywords: | Difficulty: unknown Ghcversion: | Platform: --------------------------------+------------------------------------------- Changes (by nomeata): * status: closed => reopened * resolution: fixed => Comment: I think this is a bug; the behaviour of "Setup haddock" should not depend on the locale of the user; this did bite the Debian package maintainers: http://bugs.debian.org/645573 I guess this can be fixed by a simple call to hSetEncoding. -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/626#comment:2> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

patch applied (/home/srv/code/hackage-server): "Allow import of users with .htpasswd hashes, and upgrading of those hashes" and 1 others
by Max Bolingbroke 18 Oct '11

18 Oct '11

Tue Oct 18 18:03:05 BST 2011 Max Bolingbroke <batterseapower(a)hotmail.com> * Allow import of users with .htpasswd hashes, and upgrading of those hashes Ignore-this: f741f0a7cc379f4bf26dccee10401fb0 M ./Distribution/Server.hs -3 +3 M ./Distribution/Server/Features/Html.hs -13 +4 M ./Distribution/Server/Features/Users.hs -28 +30 M ./Distribution/Server/Framework/Auth.hs -14 +42 M ./Distribution/Server/Framework/AuthCrypt.hs +23 M ./Distribution/Server/Framework/AuthTypes.hs -1 +7 M ./Distribution/Server/LegacyImport/BulkImport.hs -10 +4 M ./Distribution/Server/LegacyImport/HtPasswdDb.hs -5 +1 M ./Distribution/Server/Pages/Util.hs -1 +14 M ./Distribution/Server/Users/Backup.hs -9 +23 M ./Distribution/Server/Users/State.hs -17 +22 M ./Distribution/Server/Users/Types.hs -2 +4 M ./Distribution/Server/Users/Users.hs -39 +60 Tue Oct 18 20:32:35 BST 2011 Max Bolingbroke <batterseapower(a)hotmail.com> * Fix authentication with PUT/DELETE requests Ignore-this: 7800c854d91273e94144bfa40adea8c3 M ./Distribution/Server.hs -2 +1 M ./Distribution/Server/Framework/Auth.hs -1 +2 M ./Distribution/Server/Util/Happstack.hs +41

1 0

Upgrading .htpasswd passwords from old Hackage
by Max Bolingbroke 18 Oct '11

18 Oct '11

Hi Duncan, Please find attached another patch for you to look at. (I'm pushing smaller changes to the server without asking, but this is a larger change you will probably want to think carefully about.) This patch: 1. Allows us to import users with .htpasswd password hashes into the Hackage database 2. Prompts users with such an "old-style" password to upgrade their passwords when they first try to log in Is this the right thing to do, in your opinion? Do you have any comments about the implementation? In the course of developing this I noticed that the "change password" functionality seems to be broken for me: I can create a user just fine, but attempting to *change* the password constantly fails with "401 Unauthorized". (No, it is not the case that the password change worked and it is just asking me to authenticate with new credentials). Is this the experience of others as well? To reproduce this: 1. Create a user (e.g. foo/foo) at http://localhost:8080/users/register 2. Go to http://localhost:8080/users/register 3. Authenticate as the new user in the box that pops up 4. Attempt to change the password by a new password in the boxes (e.g. bar/bar) 5. Enter the new credentials in the box (foo/bar). It should authenticate you but it doesn't work. The old credentials (foo/foo) don't work either. It's really weird! Max

1 0

Re: Hackers hacking on hackage: introductions
by David Laing 17 Oct '11

17 Oct '11

Hi all, After Duncan's mail I figured I should say hello and mention what I'm keen to work on etc... I'm planning on splitting my free-Haskell work about 50/50 between my side projects and doing pretty much anything in the Hacakge / Cabal space. That might only end up being 4-8 hours a week, but every little bit helps, right? I'm currently an approximation to an intermediate Haskell programmer and am steadily converging, so if anyone wants a minion for any work they have planned feel free to drop me a line. My current thinking is that unless someone snags me for minion-work I'll start on some of the backup / restore work mentioned in the TODO file, hopefully sometime in the next week or so. Is there some kind of test plan or general level of testing that's hoped for? Should I add Arbitrary instances for the data being backed up so we can use QuickCheck to test that (more or less) backup . restore = id? Cheers, Dave

2 3

Re: [Hackage] #403: Bad error message when running ./configure fails to find sh
by Hackage 16 Oct '11

16 Oct '11

#403: Bad error message when running ./configure fails to find sh ----------------------------+----------------------------------------------- Reporter: duncan | Owner: Type: defect | Status: closed Priority: normal | Milestone: Cabal-1.8 Component: Cabal library | Version: HEAD Severity: minor | Resolution: fixed Keywords: | Difficulty: very easy (<1 hour) Ghcversion: 6.10.1 | Platform: Windows ----------------------------+----------------------------------------------- Comment(by mark): [http://www.jailbreakiphone.fr/ jailbreak] [http://www.jailbreakiphone.fr/tag/jailbreak-5/ jailbreak iOS 5] [http://www.jailbreakiphone.fr/jailbreak-5-untethered-1279/ jailbreak 5 untethered] [http://www.jailbreak-iphone4s.fr/ jailbreak iphone 4S] [http://www.applicationcydia.com/ application cydia] [http://www.freemobilemag.com/ free mobile] [http://www.freemobilemag.com/forfaits/ forfait free mobile] [http://www.forfaitfree.com/ free mobile] [http://www.forfaitfree.com/#forfait-free forfait free] [http://www.forfaitfree.com/#forfait-free-mobile forfait free mobile] [http://www.blogfps.com/ Battlefield 3] -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/403#comment:2> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

Documentation building client for cabal-install
by Max Bolingbroke 16 Oct '11

16 Oct '11

Hi Hackagers, I've written a rough-and-ready documentation building service for the Hackage 2.0 effort. It is functional enough to submit build reports and build documentation for a few example packages on my local Hackage instance. Duncan: please review+commit! I also had to make some small changes to cabal-install that should go into the Cabal repo. Comments/suggestions welcome. There is one major known flaw in that the invocation of "cabal install" performed by the builder is not sandboxed. So arbitrary user code runs with the same privileges as the documentation builder itself. I guess the best solution to this is a chroot (to restrict the writeable part of the file system)+timeout? Cheers, Max

3 7

Re: [Hackage] #447: build multiple packages in parallel
by Hackage 16 Oct '11

16 Oct '11

#447: build multiple packages in parallel ---------------------------------+------------------------------------------ Reporter: duncan | Owner: refold Type: enhancement | Status: assigned Priority: normal | Milestone: Component: cabal-install tool | Version: Severity: normal | Keywords: Difficulty: normal | Ghcversion: 6.8.3 Platform: | ---------------------------------+------------------------------------------ Comment(by refold): Current status (for those interested): Building multiple packages in parallel [http://dissocial.st/blog/2011-08-22-GSoC-Final was implemented], but the patches are not merged into the mainline as of yet; I'm now [https://github.com/23Skidoo/ghc-parmake working on parallelising 'cabal build']. -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/447#comment:4> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

Re: [Hackage] #48: improve c2hs support
by Hackage 15 Oct '11

15 Oct '11

#48: improve c2hs support -----------------------------+---------------------------------------------- Reporter: duncan | Owner: Type: enhancement | Status: new Priority: normal | Milestone: Cabal-2.0 Component: Cabal library | Version: Severity: normal | Keywords: Difficulty: hard (< 1 day) | Ghcversion: 6.2.1 Platform: | -----------------------------+---------------------------------------------- Changes (by mdybdal): * cc: dybber@… (added) -- Ticket URL: <http://hackage.haskell.org/trac/hackage/ticket/48#comment:7> Hackage <http://haskell.org/cabal/> Hackage: Cabal and related projects

1 0

patch applied (cabal): "Include benchmarks in product of 'setup sdist'" and 1 others
by Johan Tibell 12 Oct '11

12 Oct '11

Wed Oct 12 13:50:36 PDT 2011 Johan Tibell <johan.tibell(a)gmail.com> * Include benchmarks in product of 'setup sdist' M ./cabal/Distribution/Simple/SrcDist.hs -5 +26 View patch online: http://darcs.haskell.org/cgi-bin/darcsweb.cgi?r=cabal;a=darcs_commitdiff;h=… Wed Oct 12 14:00:35 PDT 2011 Johan Tibell <johan.tibell(a)gmail.com> * Install phase pulls in benchmark dependencies when necessary M ./cabal-install/Distribution/Client/Install.hs -3 +16 View patch online: http://darcs.haskell.org/cgi-bin/darcsweb.cgi?r=cabal;a=darcs_commitdiff;h=…

1 0

Section name clash checks in PackageDescription/Check.hs
by Johan Tibell 12 Oct '11

12 Oct '11

Hi, While adding some package checks for the new benchmark section I came up with some questions regarding section names. I think we need to clarify the design here. Which things live in the same namespace? The current checks seems inconsistent: * The library (i.e. package name) and executables live in different namespaces and thus there's no check that executables don't share the library name. * The executables and test suites live in the same namespace: , check (not (null testsThatAreExes)) $ PackageBuildImpossible $ "These test sections share names with executable sections: " ++ commaSep testsThatAreExes * The test suites and the library live in the same namespace: , check libNameClash $ PackageBuildImpossible $ "The test suite " ++ testName test ++ " has the same name as the package." I think the latter is wrong. I think the executable, test suite, and benchmark names should live in the same namespace and the package (and thus library) name should be in a different one. Do you agree? If we all agreed then I will remove the check for test suites/library name clashes above and add a check that checks all pairs of executable, test suite, and benchmark names. Cheers, Johan

2 2