Re: msys2 64 bit: help help!
Hi Simon, Andrey > > 3. After this step, starting a shell failed altogether with > "c:/msys64/mingw64_shell.bat is > > not recognised as an internal or external command". And sure enough, > there is no such file. > > Presumably it existed in step 1. So perhaps step 2 deleted it? > > [...] > > 4. As you mention, I then tried msys2_shell.cmd. It worked -- with a > noticeable delay of 5 > > seconds or so. > > I've also just got a new Win10 laptop and had the same issue with missing > mingw64_shell.bat during msys2 install. I solved it by creating mingw64.bat > with the following contents: > > msys2_shell.cmd -mingw64 -mintty > > I deleted all old shortcuts and use this script instead. Everything seems > to work fine -- can build GHC. > > Yes this is correct, the msys2 team has decided to "streamline" all their different batch files to launch msys from 4 to 1, hence the only remaining one is msys2_shell.cmd which accepts arguments of which shell to open and using which console host. Unfortunately this is done via their upgrade-core script and doesn't know how to remove the installer shortcuts, so you end up with dead shortcuts. > 1. I just left the machine for 10-15 mins and lo! the shell windows opened up. It just took a lng time. > > At this point, starting a new shell no longer took a long time. It all seemed to be working. First launch should be finishing setting up the environment so will take slightly longer, but shouldn't have taken that long. Could this be AV related? I always add an exception to the AV (or the build in windows defender) for the msys2 folder to prevent it from scanning files continuously. Especially building GHC this can slow things down considerably depending on the AV. > 2. I then ran pacman -Syuu as instructed on the installation page: https://sourceforge.net/p/msys2/wiki/MSYS2%20installation/ > > The log of what happened is below. There are numerous failures involving Cygwin, which I do not have installed, at least not so far as I know. I do not know if these failures matter. These instructions are basically telling it to upgrade the world. They are however a bit wrong, https://github.com/Alexpux/MSYS2-packages/issues/373 msys2 is derived from Cygwin so it inherits much of the problems of Cygwin. The msys2 runtime is the Cygwin runtime with patches added which is why the errors mention Cygwin. The issue is that the msys2-runtime has been upgraded by "pacman -Syuu" at which point a new "Cygwin" dll has been downloaded. However all Cygwin/msys2 runtimes share the same address space and thus you can't have multiple versions of the same runtime loaded at once. This is why subsequent calls to anything relying on the msys2 runtime will fail with a weird fork error. The solution is to just close all open msys2 window and re-open. Our own instructions page https://ghc.haskell.org/trac/ghc/wiki/Building/Preparation/Windows breaks this down in a few steps to avoid this issue. 1) first update the packages and 2) only after that update the msys-core files. But you will still need to restart the shell. > * should I worry about all those install errs No they're perfectly fine and expected. I would however re-run the pacman -Syu to make sure all packages were updated, now that the runtime has been updated already it shouldn't be updated again and you shouldn't see any fork errors. > * how can I debug what's happening with > that long delay If it's only startup and not executing of other commands or bash completion then my bet would be AV software. If bash completion is slow or commands like ls as well you may be hitting a long standing issue some computers have in which the domain controller is being hit for every invocation of commands, causing a slowdown https://github.com/Alexpux/MSYS2-packages/issues/138 , Solution 2 from https://gist.github.com/k-takata/9b8d143f0f3fef5abdab seems to fix it for most people. * Should I nuke the start menu shortcuts that the msys64 installer so carefully installed in favour of msys2_shell.cmd? Yes, these are now dead. you need to use msys2_shell.cmd but also pass it -mingw64 so it knows what shell to start. mintty is supposed to be the default, but in case that changes you can also pass it -mintty as well to be sure it doesn't change. I am working on a script to automate this setup, hopefully that would make it easier next time! Cheers, Tamar ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
RE: msys2 64 bit: help help!
Hi Simon, > 3. After this step, starting a shell failed altogether with > "c:/msys64/mingw64_shell.bat is > not recognised as an internal or external command". And sure enough, there is > no such file. > Presumably it existed in step 1. So perhaps step 2 deleted it? > [...] > 4. As you mention, I then tried msys2_shell.cmd. It worked -- with a > noticeable delay of 5 > seconds or so. I've also just got a new Win10 laptop and had the same issue with missing mingw64_shell.bat during msys2 install. I solved it by creating mingw64.bat with the following contents: msys2_shell.cmd -mingw64 -mintty I deleted all old shortcuts and use this script instead. Everything seems to work fine -- can build GHC. Cheers, Andrey ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
RE: msys2 64 bit: help help!
Tamar Thank you! Yes, it's in c:/msys64 Since I wrote more odd things have happened. 1. I just left the machine for 10-15 mins and lo! the shell windows opened up. It just took a lng time. At this point, starting a new shell no longer took a long time. It all seemed to be working. 2. I then ran pacman -Syuu as instructed on the installation page: https://sourceforge.net/p/msys2/wiki/MSYS2%20installation/ The log of what happened is below. There are numerous failures involving Cygwin, which I do not have installed, at least not so far as I know. I do not know if these failures matter. 3. After this step, starting a shell failed altogether with "c:/msys64/mingw64_shell.bat is not recognised as an internal or external command". And sure enough, there is no such file. Presumably it existed in step 1. So perhaps step 2 deleted it? 4. As you mention, I then tried msys2_shell.cmd. It worked -- with a noticeable delay of 5 seconds or so. So I'm less stuck than before but * should I worry about all those install errs * how can I debug what's happening with that long delay * Should I nuke the start menu shortcuts that the msys64 installer so carefully installed in favour of msys2_shell.cmd? Thanks Simon pacman -Syuu :: Synchronising package databases... mingw32 277.4 KiB 501K/s 00:01 [#] 100% mingw32.sig 96.0 B 0.00B/s 00:00 [#] 100% mingw64 277.0 KiB 481K/s 00:01 [#] 100% mingw64.sig 96.0 B 0.00B/s 00:00 [#] 100% msys 134.9 KiB 2014K/s 00:00 [#] 100% msys.sig 96.0 B 0.00B/s 00:00 [#] 100% :: Starting full system upgrade... :: Replace repman-git with msys/pactoys-git? [Y/n] y resolving dependencies... looking for conflicting packages... Packages (41) bash-completion-2.3-1 bsdcpio-3.2.0-1 bsdtar-3.2.0-1 coreutils-8.25-1 crypt-1.3-1 curl-7.49.1-1 file-5.25-1 filesystem-2016.05-3 gcc-libs-5.3.0-3 gettext-0.19.7-3 gmp-6.1.0-2 gnupg-1.4.20-1 grep-2.22-3 gzip-1.7-1 heimdal-libs-1.5.3-9 libarchive-3.2.0-1 libasprintf-0.19.7-3 libassuan-2.4.2-1 libcrypt-1.3-1 libcurl-7.49.1-1 libexpat-2.1.1-1 libgettextpo-0.19.7-3 libgpg-error-1.21-1 libgpgme-1.6.0-1 libintl-0.19.7-3 liblzma-5.2.2-1 libnettle-3.2-1 libopenssl-1.0.2.h-1 libssh2-1.7.0-1 mintty-1~2.2.3-1 mpfr-3.1.4-1 msys2-launcher-git-0.3.29.4028b6c-1 msys2-runtime-2.5.1-1 ncurses-6.0.20160220-1 openssl-1.0.2.h-1 pacman-5.0.1.6403.520736d-1 pactoys-git-r1.e58a7ac-1 rebase-4.4.2-1 repman-git-r23.87bf865-1 [removal] wget-1.17.1-3 xz-5.2.2-1 Total Download Size:24.85 MiB Total Installed Size: 124.28 MiB Net Upgrade Size:6.09 MiB :: Proceed with installation? [Y/n] y :: Retrieving packages ... msys2-runtime-2.5.1... 2.2 MiB 912K/s 00:03 [#] 100% bash-completion-2.3... 184.1 KiB 4.50M/s 00:00 [#] 100% gcc-libs-5.3.0-3-x86_64 783.6 KiB 1818K/s 00:00 [#] 100% libintl-0.19.7-3-x86_64 25.6 KiB 3.13M/s 00:00 [#] 100% libgettextpo-0.19.7... 110.7 KiB 3.60M/s 00:00 [#] 100% libasprintf-0.19.7-...11.8 KiB 1686K/s 00:00 [#] 100% gettext-0.19.7-3-x86_64 1524.1 KiB 1774K/s 00:01 [#] 100% liblzma-5.2.2-1-x86_6476.6 KiB 6.23M/s 00:00 [#] 100% gmp-6.1.0-2-x86_64 369.2 KiB 2.03M/s 00:00 [#] 100% libnettle-3.2-1-x86_64 101.1 KiB 5.20M/s 00:00 [#] 100% coreutils-8.25-1-x86_642.2 MiB 1794K/s 00:01 [#] 100% ncurses-6.0.2016022... 1147.2 KiB 1833K/s 00:01 [#] 100% bsdcpio-3.2.0-1-x86_64 747.1 KiB 2.08M/s 00:00 [#] 100% bsdtar-3.2.0-1-x86_64785.4 KiB 1879K/s 00:00 [#] 100% libcrypt-1.3-1-x86_64 13.5 KiB 1692K/s 00:00 [#] 100% crypt-1.3-1-x86_6413.9 KiB 1979K/s 00:00 [#] 100% libopenssl-1.0.2.h-... 803.0 KiB 1921K/s 00:00 [#] 100% heimdal-libs-1.5.3-... 592.7 KiB 1976K/s 00:00 [#] 100% openssl-1.0.2.h-1-x... 1354.4 KiB 1929K/s 00:01 [#] 100% gzip-1.7-1-x86_64 97.0 KiB 3.16M/s 00:00 [#] 100% libssh2-1.7.0-1-x86_64 170.7 KiB 2.69M/s 00:00 [#] 100% libexpat-2.1.1-1-x86_64 56.9 KiB 3.27M/s 00:00 [#] 100% libcurl-7.49.1-1-x86_64 191.1 KiB 1855K/s 00:00 [#] 100% curl-7.49.1-1-x86_64 608.4 KiB 627K/s 00:01
Re: Using the GHC API to write an interpreter
On 27 June 2016 at 13:31, Christopher Donewrote: > On 27 June 2016 at 10:01, Simon Marlow wrote: > > On 26 June 2016 at 11:28, Christopher Done wrote: > >> > >> I've been pondering how feasible it would be to: > >> > >> * Compile in stages a module with the byte code linker > >> * Keep hold of the Core source > >> * Interpret the Core AST within Haskell > > > > Interestingly, the first implementation of GHCi was a Core interpreter, > but > > it ran into a lot of problems. For starters it would have unsafeCoerce > > everywhere. Support for unboxed values is very very difficult. > > What year is that implementation from? I wouldn't mind taking a look > for it in the GHC repo history. > > I think around here is a good place to start looking: https://phabricator.haskell.org/rGHCbca9dd54c2b39638cb4638aaccf6015a104a1df5#021fe2a9 Cheers Simon > >> * is not tagless, so we preserve type info > > > > Not sure what you mean here. Your interpreter would be running on top of > > the same RTS with the same data representation, so it would have to use > the > > same tagging and representation conventions as the rest of GHC > > That's true, if a value comes from a compiled RTS function with a > polymorphic type then I don't know what its real type is to marshal it > properly. Drat. > > >> * allows top-level names to be redefined > > > > This you could do with the extisting byte-code interpreter, by instead of > > linking Names directly you link to some runtime Name-lookup function. > You > > would probably want to revert all CAFs when the code changes too; this is > > currently not implemented for byte code. > > Right, I considered this but without the type information it's going > to blow up if I change the arity of a function or a data type or > whatever. > > >> * when a function is applied, it checks the type of its arguments > > > > Aha, but what if the arguments come from compiled code? GHC doesn't > carry > > type information around at runtime, except that it is possible > reconstruct > > types in a limited kind of way (this is what the GHC debugger does). > > Indeed, from compiled code e.g. id then id (undefined :: Foo) would > come back as something unidentifiable as being of type Foo. That's the > flaw in my plan. > > Looks like the current interpreter would have to be extended to > support this or a whole new one re-implementing all the primitives like > in GHCJS. > > Thanks! > ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: Using the GHC API to write an interpreter
On 27 June 2016 at 10:01, Simon Marlowwrote: > On 26 June 2016 at 11:28, Christopher Done wrote: >> >> I've been pondering how feasible it would be to: >> >> * Compile in stages a module with the byte code linker >> * Keep hold of the Core source >> * Interpret the Core AST within Haskell > > Interestingly, the first implementation of GHCi was a Core interpreter, but > it ran into a lot of problems. For starters it would have unsafeCoerce > everywhere. Support for unboxed values is very very difficult. What year is that implementation from? I wouldn't mind taking a look for it in the GHC repo history. >> * is not tagless, so we preserve type info > > Not sure what you mean here. Your interpreter would be running on top of > the same RTS with the same data representation, so it would have to use the > same tagging and representation conventions as the rest of GHC That's true, if a value comes from a compiled RTS function with a polymorphic type then I don't know what its real type is to marshal it properly. Drat. >> * allows top-level names to be redefined > > This you could do with the extisting byte-code interpreter, by instead of > linking Names directly you link to some runtime Name-lookup function. You > would probably want to revert all CAFs when the code changes too; this is > currently not implemented for byte code. Right, I considered this but without the type information it's going to blow up if I change the arity of a function or a data type or whatever. >> * when a function is applied, it checks the type of its arguments > > Aha, but what if the arguments come from compiled code? GHC doesn't carry > type information around at runtime, except that it is possible reconstruct > types in a limited kind of way (this is what the GHC debugger does). Indeed, from compiled code e.g. id then id (undefined :: Foo) would come back as something unidentifiable as being of type Foo. That's the flaw in my plan. Looks like the current interpreter would have to be extended to support this or a whole new one re-implementing all the primitives like in GHCJS. Thanks! ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: Using the GHC API to write an interpreter
On 27 June 2016 at 04:11, Edward Z. Yangwrote: > I don't understand what the bytecode format has to do here. Since > your suggestion is to just store Core you can just compile to object > code. True, I could compile to either as long as I can link it dynamically. > > Any input into this? How far away is GHC's current architecture from > > supporting such a concept? > > Well, if you are going to support update you need to make sure that the > tag information is more elaborate than what GHC currently supports > (a type would just be a Name, which is going to get reused when you > recompile.) Indeed -- like in GHCi when you redefine a named thing, I'd hope to implement an incrementing Name[n] versioning for names. But Core's AST is trivial so it'd be easy to make this kind of transformation. ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: msys2 64 bit: help help!
Congrats Simon : ) I use precisely the same machine! While I haven't run into that particular problem, I do use an alternative console that might provide a workaround in case you keep running into trouble: http://cmder.net/ My setup log is here, which explains how to use Cmder with MSYS2: https://gist.github.com/lukexi/e634067f1d7e3a629988#cmder All the best Luke On Mon, Jun 27, 2016 at 1:21 AM, David Macekwrote: > On 27. 6. 2016 10:01, Simon Peyton Jones via ghc-devs wrote: > > Friends, esp Tamar, > > > > I am in happy possession of a new Surface Book, running Windows 10, > which is delightful – except that I can’t make the msys64 installation > work, which is crucial for GHC. Can any of you help? > > > > ·I install it from here: > https://sourceforge.net/p/msys2/wiki/MSYS2%20installation/, which seems > to be the canonical place. The actual 64-bit exe seems to be > msys2-x86_64-20160205.exe. > > > > ·Installation goes fine > > > > ·But when I launch the “MinGW-w54 Win64 shell” from the Start > menu, the screen flashes as if it is briefly putting up a window, but then > the window disappears. > > > > ·Each time I do this the task manager shows that there is > another running ‘mintty.exe’. But it has no visible window > > > > > > > > Does anyone have any idea what I can do or how I can investigate? I > can’t do any GHC development without this! > > Hi. > > Please try `C:\msys64\usr\bin\mintty.exe -h always -` (incl. the trailing > dash). You can also try `C:\msys64\usr\bin\bash.exe -li` in case the > culprit is really mintty. > > > -- > David Macek > > > ___ > ghc-devs mailing list > ghc-devs@haskell.org > http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs > > ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: msys2 64 bit: help help!
On 27. 6. 2016 10:01, Simon Peyton Jones via ghc-devs wrote: > Friends, esp Tamar, > > I am in happy possession of a new Surface Book, running Windows 10, which is > delightful – except that I can’t make the msys64 installation work, which is > crucial for GHC. Can any of you help? > > ·I install it from here: > https://sourceforge.net/p/msys2/wiki/MSYS2%20installation/, which seems to be > the canonical place. The actual 64-bit exe seems to be > msys2-x86_64-20160205.exe. > > ·Installation goes fine > > ·But when I launch the “MinGW-w54 Win64 shell” from the Start menu, > the screen flashes as if it is briefly putting up a window, but then the > window disappears. > > ·Each time I do this the task manager shows that there is another > running ‘mintty.exe’. But it has no visible window > > > > Does anyone have any idea what I can do or how I can investigate? I can’t do > any GHC development without this! Hi. Please try `C:\msys64\usr\bin\mintty.exe -h always -` (incl. the trailing dash). You can also try `C:\msys64\usr\bin\bash.exe -li` in case the culprit is really mintty. -- David Macek smime.p7s Description: S/MIME Cryptographic Signature ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: Using the GHC API to write an interpreter
On 26 June 2016 at 11:28, Christopher Donewrote: > I've been pondering how feasible it would be to: > > * Compile in stages a module with the byte code linker > * Keep hold of the Core source > * Interpret the Core AST within Haskell > Interestingly, the first implementation of GHCi was a Core interpreter, but it ran into a lot of problems. For starters it would have unsafeCoerce everywhere. Support for unboxed values is very very difficult. > * When encountering built-in/primitives (or things from other libraries), > we compile that Core term and link it as an HValue and then run it with the > arguments expected. So () would be such a HValue, as would be "show" which > in interpretable unoptimized Core would take an extra argument for the Show > instance. When passing in values to such "foreign" functions it would wrap > them up in an interpretive way. > > This is the hypothetical idea, it seems like it would yield a really > trivial way to write a new and interesting interpreter for GHC Haskell > without having to re-implement any prim ops, ready to work on regular > Haskell code. > > In my case, I would use this to write an interpreter which: > > * is not tagless, so we preserve type info > Not sure what you mean here. Your interpreter would be running on top of the same RTS with the same data representation, so it would have to use the same tagging and representation conventions as the rest of GHC > * allows top-level names to be redefined > This you could do with the extisting byte-code interpreter, by instead of linking Names directly you link to some runtime Name-lookup function. You would probably want to revert all CAFs when the code changes too; this is currently not implemented for byte code. > * when a function is applied, it checks the type of its arguments > Aha, but what if the arguments come from compiled code? GHC doesn't carry type information around at runtime, except that it is possible reconstruct types in a limited kind of way (this is what the GHC debugger does). Cheers Simon > Both of these are pretty much necessary for being able to do in-place > update of a running program while developing (a la Emacs or Smalltalk), and > type tags let us throw a regular Haskell exception of type error, a la > deferred type errors. It means in your running program, if you make a > mistake or forget to update one part, it doesn't bring the whole program > down with an RTS error or a segfault, maybe a handler in a thread (like a > server or a video game) throws an exception and the developer just updates > their code and tries again. > > I'd love support for something like this, but I'd rather not have to > re-create the world just to add this capability. Because it's really just > conceptually regular interpreted GHC Haskell plus type tags and updating, > it seems like it should be a small diff. > > Any input into this? How far away is GHC's current architecture from > supporting such a concept? > > Ciao! > > ___ > ghc-devs mailing list > ghc-devs@haskell.org > http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs > > ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
msys2 64 bit: help help!
Friends, esp Tamar, I am in happy possession of a new Surface Book, running Windows 10, which is delightful - except that I can't make the msys64 installation work, which is crucial for GHC. Can any of you help? *I install it from here: https://sourceforge.net/p/msys2/wiki/MSYS2%20installation/, which seems to be the canonical place. The actual 64-bit exe seems to be msys2-x86_64-20160205.exe. *Installation goes fine *But when I launch the "MinGW-w54 Win64 shell" from the Start menu, the screen flashes as if it is briefly putting up a window, but then the window disappears. *Each time I do this the task manager shows that there is another running 'mintty.exe'. But it has no visible window Does anyone have any idea what I can do or how I can investigate? I can't do any GHC development without this! Thanks Simon ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: [Diffusion] [Build Failed] rGHCbb84ee44e30e: Improve pretty-printing of Avail
I believe I've just fixed this. I think it was non-deterministic in some way that I don't fully understand, but hopefully I've made it deterministic now. https://phabricator.haskell.org/rGHC7843c71c7e48cdba115bef422184e855ede23a67 On 24 June 2016 at 15:12, Simon Peyton Jones via ghc-devs < ghc-devs@haskell.org> wrote: > These failures are odd. One is this change in T8761 which I'm sure is not > my fault. I saw that simonmar did some hFlush thing -- could that be it? > > Simon > > > +++ ./T8761.run/T8761.comp.stderr.normalised2016-06-24 > 15:00:22.975083533 +0100 > @@ -1,3 +1,7 @@ > +pattern Q1 x1_0 x2_1 x3_2 <- ((x1_0, x2_1), [x3_2], _, _) > +pattern x1_0 Q2 x2_1 = ((x1_0, x2_1)) > +pattern Q3 {qx3, qy3, qz3} <- ((qx3, qy3), [qz3]) where > + Q3 qx3 qy3 qz3 = ((qx3, qy3), [qz3]) > T8761.hs:(16,1)-(39,13): Splicing declarations > do { [qx1, qy1, qz1] <- mapM >(/ i -> newName $ "x" ++ show i) [1, 2, 3]; > @@ -119,10 +123,6 @@ > pattern Pup x <- MkUnivProv x > pattern Puep :: forall a. forall b. Show b => a -> b -> (ExProv, a) > pattern Puep x y <- (MkExProv y, x) > -pattern Q1 x1_0 x2_1 x3_2 <- ((x1_0, x2_1), [x3_2], _, _) > -pattern x1_0 Q2 x2_1 = ((x1_0, x2_1)) > -pattern Q3 {qx3, qy3, qz3} <- ((qx3, qy3), [qz3]) where > - Q3 qx3 qy3 qz3 = ((qx3, qy3), [qz3]) > pattern T8761.P :: GHC.Types.Bool > pattern T8761.Pe :: () => forall (a0_0 :: *) . a0_0 -> T8761.Ex > pattern T8761.Pu :: forall (a0_0 :: *) . a0_0 -> a0_0 > > | -Original Message- > | From: nore...@phabricator.haskell.org > | [mailto:nore...@phabricator.haskell.org] > | Sent: 24 June 2016 14:50 > | To: Simon Peyton Jones> | Subject: [Diffusion] [Build Failed] rGHCbb84ee44e30e: Improve pretty- > | printing of Avail > | > | Harbormaster failed to build B10226: rGHCbb84ee44e30e: Improve pretty- > | printing of Avail! > | > | BRANCHES > |master, wip/rae > | > | USERS > |simonpj (Author) > | > | COMMIT > |https://phabricator.haskell.org/rGHCbb84ee44e30e > | > | EMAIL PREFERENCES > |https://phabricator.haskell.org/settings/panel/emailpreferences/ > | > | To: simonpj, Harbormaster > ___ > ghc-devs mailing list > ghc-devs@haskell.org > http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs > ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
Re: [commit: ghc] master: Accept new (lower) allocations for T7257 (15641b0)
Yes, it was the sizeExpr fix. It validated locally and on Travis, but I'm guessing it was right on the boundary. On 22 June 2016 at 22:37, Bartosz Nitkawrote: > Appears to be: > a47b62cb3685 Second attempt to fix sizeExpr > > https://perf.haskell.org/ghc/#revision/9d62d09a6c399c98491b7a63a7a1366c89fcf5db > > 2016-06-22 22:24 GMT+01:00 Simon Peyton Jones via ghc-devs < > ghc-devs@haskell.org>: > >> Does anyone know what made T7257 better? >> >> Simon >> >> | -Original Message- >> | From: ghc-commits [mailto:ghc-commits-boun...@haskell.org] On Behalf Of >> | g...@git.haskell.org >> | Sent: 22 June 2016 22:08 >> | To: ghc-comm...@haskell.org >> | Subject: [commit: ghc] master: Accept new (lower) allocations for T7257 >> | (15641b0) >> | >> | Repository : ssh://g...@git.haskell.org/ghc >> | >> | On branch : master >> | Link : >> | >> https://na01.safelinks.protection.outlook.com/?url=http%3a%2f%2fghc.hask >> | ell.org >> %2ftrac%2fghc%2fchangeset%2f15641b07f1d3ccb5f35b4f31539ecceb5fb38 >> | c17%2fghc=01%7c01%7csimonpj%40064d.mgd.microsoft.com >> %7c32fafc88b758 >> | 43148b2a08d39ae1ea3f%7c72f988bf86f141af91ab2d7cd011db47%7c1=eLrMob >> | XyWHA691jI3t4pf0CbGybfrVqptmkpTDdtDI8%3d >> | >> | >--- >> | >> | commit 15641b07f1d3ccb5f35b4f31539ecceb5fb38c17 >> | Author: Simon Marlow >> | Date: Wed Jun 22 20:54:59 2016 +0100 >> | >> | Accept new (lower) allocations for T7257 >> | >> | >> | >--- >> | >> | 15641b07f1d3ccb5f35b4f31539ecceb5fb38c17 >> | testsuite/tests/perf/should_run/all.T | 3 ++- >> | 1 file changed, 2 insertions(+), 1 deletion(-) >> | >> | diff --git a/testsuite/tests/perf/should_run/all.T >> | b/testsuite/tests/perf/should_run/all.T >> | index caf4eff..3cb6f8e 100644 >> | --- a/testsuite/tests/perf/should_run/all.T >> | +++ b/testsuite/tests/perf/should_run/all.T >> | @@ -276,9 +276,10 @@ test('T7257', >> |[(wordsize(32), 989850664, 10), >> | # expected value: 1246287228 (i386/Linux) >> | # 2016-04-06: 989850664 (i386/Linux) no idea >> | what happened >> | - (wordsize(64), 1654893248, 5)]), >> | + (wordsize(64), 1414893248, 5)]), >> | # 2012-09-21: 1774893760 (amd64/Linux) >> | # 2015-11-03: 1654893248 (amd64/Linux) >> | +# 2016-06-22: 1414893248 (amd64/Linux, sizeExpr >> | fix) >> |stats_num_field('peak_megabytes_allocated', >> |[(wordsize(32), 217, 5), >> | # 2012-10-08: 217 (x86/Linux) >> | >> | ___ >> | ghc-commits mailing list >> | ghc-comm...@haskell.org >> | >> https://na01.safelinks.protection.outlook.com/?url=http%3a%2f%2fmail.has >> | kell.org%2fcgi-bin%2fmailman%2flistinfo%2fghc- >> | commits=01%7c01%7csimonpj%40064d.mgd.microsoft.com >> %7c32fafc88b75843 >> | 148b2a08d39ae1ea3f%7c72f988bf86f141af91ab2d7cd011db47%7c1=RftvNpBH >> | H7bw53EdhgnsdV%2f6M2LNgFVB1t7gCwOTT7A%3d >> ___ >> ghc-devs mailing list >> ghc-devs@haskell.org >> http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs >> > > > ___ > ghc-devs mailing list > ghc-devs@haskell.org > http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs > > ___ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs