RE: `partition'
On 15 October 2004 09:55, Josef Svenningsson wrote: No, sorry. It constitutes a change in library API, which you can't #ifdef your way around because there isn't a CPP symbol which changes with the patchlevel (on purpose). In what sense is it a change in the API? The only thing that is needed is to add a tilde to a pattern. This does not change the interface of the library. I see no reason to #ifdef around it. But if your policy is not to change the semantics of functions between patchlevels then I'll guess we'll survive until 6.4. As you say, it's a change in the semantics of partition. You might write code that depends on the lazy behaviour, and such code wouldn't work with earlier patchlevels. Yes, the no change to library APIs policy is quite restrictive. It was introduced as a way to try to provide stability between releases - we provide bugfixes without breaking your code. Ideally we'd also like to push out more frequent major releases too, but releases take time. Cheers, Simon ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
RE: `partition'
No, sorry. It constitutes a change in library API, which you can't #ifdef your way around because there isn't a CPP symbol which changes with the patchlevel (on purpose). Cheers, Simon On 14 October 2004 11:09, Serge D. Mechveliani wrote: Indeed, could it? For 6.2.2 is have not been released so far ... On Thu, Oct 14, 2004 at 12:02:20PM +0200, Josef Svenningsson wrote: A little browsing in the CVS reveals that partition is fixed in GHC's libraries but that it was done after the 6.2 branch. So the 6.2 branch is still using the old erroneous definition. Could this fix be merged into the 6.2 branch, please? /Josef -Original Message- From: [EMAIL PROTECTED] [mailto:glasgow-haskell- [EMAIL PROTECTED] On Behalf Of Johannes Waldmann Sent: den 14 oktober 2004 10:30 To: Serge D. Mechveliani Cc: [EMAIL PROTECTED] Subject: Re: `partition' the function let ns = [1 .. 10^6] :: [Int] p = partition even ns in take 2 $ fst p computes too long in ghc-6.2.2-September-26 and needs several megabyte of stack. this has been noticed a few times, for example http://www.mail-archive.com/[EMAIL PROTECTED]/msg07508.html -- -- Johannes Waldmann, Tel/Fax: (0341) 3076 6479 / 6480 -- -- http://www.imn.htwk-leipzig.de/~waldmann/ - ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
RE: `partition'
No, sorry. It constitutes a change in library API, which you can't #ifdef your way around because there isn't a CPP symbol which changes with the patchlevel (on purpose). In what sense is it a change in the API? The only thing that is needed is to add a tilde to a pattern. This does not change the interface of the library. I see no reason to #ifdef around it. But if your policy is not to change the semantics of functions between patchlevels then I'll guess we'll survive until 6.4. All the best, /Josef On 14 October 2004 11:09, Serge D. Mechveliani wrote: Indeed, could it? For 6.2.2 is have not been released so far ... On Thu, Oct 14, 2004 at 12:02:20PM +0200, Josef Svenningsson wrote: A little browsing in the CVS reveals that partition is fixed in GHC's libraries but that it was done after the 6.2 branch. So the 6.2 branch is still using the old erroneous definition. Could this fix be merged into the 6.2 branch, please? /Josef -Original Message- From: [EMAIL PROTECTED] [mailto:glasgow-haskell- [EMAIL PROTECTED] On Behalf Of Johannes Waldmann Sent: den 14 oktober 2004 10:30 To: Serge D. Mechveliani Cc: [EMAIL PROTECTED] Subject: Re: `partition' the function let ns = [1 .. 10^6] :: [Int] p = partition even ns in take 2 $ fst p computes too long in ghc-6.2.2-September-26 and needs several megabyte of stack. this has been noticed a few times, for example http://www.mail-archive.com/[EMAIL PROTECTED]/msg07508.html -- -- Johannes Waldmann, Tel/Fax: (0341) 3076 6479 / 6480 -- -- http://www.imn.htwk-leipzig.de/~waldmann/ - ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
`partition'
Dear GHC developers, the function let ns = [1 .. 10^6] :: [Int] p = partition even ns in take 2 $ fst p computes too long in ghc-6.2.2-September-26 and needs several megabyte of stack. What might be the reason? - Serge Mechveliani [EMAIL PROTECTED] ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
Re: `partition'
the function let ns = [1 .. 10^6] :: [Int] p = partition even ns in take 2 $ fst p computes too long in ghc-6.2.2-September-26 and needs several megabyte of stack. this has been noticed a few times, for example http://www.mail-archive.com/[EMAIL PROTECTED]/msg07508.html -- -- Johannes Waldmann, Tel/Fax: (0341) 3076 6479 / 6480 -- -- http://www.imn.htwk-leipzig.de/~waldmann/ - ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
RE: `partition'
A little browsing in the CVS reveals that partition is fixed in GHC's libraries but that it was done after the 6.2 branch. So the 6.2 branch is still using the old erroneous definition. Could this fix be merged into the 6.2 branch, please? /Josef -Original Message- From: [EMAIL PROTECTED] [mailto:glasgow-haskell- [EMAIL PROTECTED] On Behalf Of Johannes Waldmann Sent: den 14 oktober 2004 10:30 To: Serge D. Mechveliani Cc: [EMAIL PROTECTED] Subject: Re: `partition' the function let ns = [1 .. 10^6] :: [Int] p = partition even ns in take 2 $ fst p computes too long in ghc-6.2.2-September-26 and needs several megabyte of stack. this has been noticed a few times, for example http://www.mail-archive.com/[EMAIL PROTECTED]/msg07508.html -- -- Johannes Waldmann, Tel/Fax: (0341) 3076 6479 / 6480 -- -- http://www.imn.htwk-leipzig.de/~waldmann/ - ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
Re: `partition'
Indeed, could it? For 6.2.2 is have not been released so far ... On Thu, Oct 14, 2004 at 12:02:20PM +0200, Josef Svenningsson wrote: A little browsing in the CVS reveals that partition is fixed in GHC's libraries but that it was done after the 6.2 branch. So the 6.2 branch is still using the old erroneous definition. Could this fix be merged into the 6.2 branch, please? /Josef -Original Message- From: [EMAIL PROTECTED] [mailto:glasgow-haskell- [EMAIL PROTECTED] On Behalf Of Johannes Waldmann Sent: den 14 oktober 2004 10:30 To: Serge D. Mechveliani Cc: [EMAIL PROTECTED] Subject: Re: `partition' the function let ns = [1 .. 10^6] :: [Int] p = partition even ns in take 2 $ fst p computes too long in ghc-6.2.2-September-26 and needs several megabyte of stack. this has been noticed a few times, for example http://www.mail-archive.com/[EMAIL PROTECTED]/msg07508.html -- -- Johannes Waldmann, Tel/Fax: (0341) 3076 6479 / 6480 -- -- http://www.imn.htwk-leipzig.de/~waldmann/ - ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs ___ Glasgow-haskell-bugs mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs