draft proposal on provenance-qualified dependencies
Simon Peyton Jones
simonpj at microsoft.com
Fri Feb 23 16:03:52 UTC 2018
| current existing "search path-like" functionality where a chain of
| overlays may live over a repo. The proposal does not address that,
| because it is how things work already -- although arguably, the way in
| which this works may be insufficiently understood among existing cabal
| users?
Well I can say with certainty that it's insufficiently understood by /this/ cabal user.
I had no idea there could be more than one repo, which 'cabal update' caches locally.
Simon
| -----Original Message-----
| From: Gershom B [mailto:gershomb at gmail.com]
| Sent: 23 February 2018 15:02
| To: Simon Peyton Jones <simonpj at microsoft.com>
| Cc: cabal-devel <cabal-devel at haskell.org>
| Subject: Re: draft proposal on provenance-qualified dependencies
|
| Thanks for these comments Simon. It is good to have a sanity-check on
| these proposals before they go before a broad audience. I'll try to
| take them into account and submit this properly as a proposal
| (including creating the associated discussion thread) in the next few
| days. If anyone else has any thoughts (bear in mind this involves
| cross-cutting changes across cabal files and a bit of ghc) please send
| them on.
|
| On "Does a particular build have to use packages from one repo only?"
| -- the idea is that _per package_ a provenance may be specified to a
| specific repo. If no provenance is specified, then there is the
| current existing "search path-like" functionality where a chain of
| overlays may live over a repo. The proposal does not address that,
| because it is how things work already -- although arguably, the way in
| which this works may be insufficiently understood among existing cabal
| users? (In fact, looking at the cabal documentation, I see that the
| description of multiple remote repos doesn't specify the manner in
| which they are combined, which it should).
|
| Cheers,
| Gershom
|
| On Fri, Feb 23, 2018 at 6:26 AM, Simon Peyton Jones
| <simonpj at microsoft.com> wrote:
| > Gershom
| >
| > Looks like a great idea.
| >
| > Lots of questions though:
| >
| >
| > - I think more motivation would be helpful. E.g. "You are in change
| > of the GHC 8.6 release. Package authors don't want to upload a
| new
| > version of their packages until 8.6 is out, but you still want to
| > smoke-test 8.6 against their packages. Doing so requires some
| minor
| > changes (version bounds, changes in base-library APIs etc); so you
| > want to be able to make these changes in a sandbox that won't hurt
| > anyone". Or something like that.
| >
| > Maybe describe other use-cases. It's *much* easier to evaluate
| > a proposal when I'm totally clear what it's for.
| >
| > - Does a particular build have to use packages from one repo only?
| > Or is there something like a "search path"?
| >
| > Thanks!
| >
| > Simon
| >
| > | -----Original Message-----
| > | From: cabal-devel [mailto:cabal-devel-bounces at haskell.org] On
| > | Behalf Of Gershom B
| > | Sent: 19 February 2018 00:26
| > | To: cabal-devel <cabal-devel at haskell.org>
| > | Subject: draft proposal on provenance-qualified dependencies
| > |
| > | Hey all, I mentioned (on the long SLURP thread) that I was
| thinking
| > | about a general proposal for provenance-qualified dependencies to
| > | reduce coupling in the haskell ecosystem. Having worked it out a
| > | bit, I think the bigger win is it also provides a way to specify
| > | dependencies on git repos, etc., which has been an oft-requested
| > | feature.
| > |
| > | I don't want to submit it as an ecosystem proposal proper without
| > | further polish, and I held off on bugging a larger audience of
| cabal
| > | folks until the 2.2 branch was cut. So now I'm passing this along
| > | for further comment and polish before I make a real proposal:
| > |
| > |
| https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgit
| > | hu
| > | b.com%2Fgbaz%2Fghc-proposals%2Fblob%2Fpatch-1%2Fproposals%2F0000-
| > | provenance-qualified-
| > |
| > |
| imports.rst&data=04%7C01%7Csimonpj%40microsoft.com%7C64fd20012b9a4b2
| > | 4d
| > |
| > |
| 28508d5772f6cf2%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6365459
| > | 67
| > |
| 936143539%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMz
| > | Ii
| > | LCJBTiI6Ik1haWwifQ%3D%3D%7C-
| > |
| > |
| 1&sdata=cH0gNADmzA%2BTkmXZEDY6lLYUTx2D2KX%2B3T8KO%2FvU86s%3D&reserve
| > | d=
| > | 0
| > |
| > | There's no urgency, but it would be good to get some feedback in
| > | the next few weeks if possible.
| > |
| > | Cheers,
| > | Gershom
| > | _______________________________________________
| > | cabal-devel mailing list
| > | cabal-devel at haskell.org
| > |
| > |
| https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail
| > | .h
| > | askell.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fcabal-
| > |
| > |
| devel&data=04%7C01%7Csimonpj%40microsoft.com%7C64fd20012b9a4b24d2850
| > | 8d
| > |
| > |
| 5772f6cf2%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6365459679361
| > | 43
| > |
| 539%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJB
| > | Ti
| > | I6Ik1haWwifQ%3D%3D%7C-
| > |
| 1&sdata=fgfMNTNt%2BwEQ5PaTKxtl0bmO7wDv9sBiMUnWSbJhcnE%3D&reserved=0
More information about the cabal-devel
mailing list