|
View:
New views
3 Messages
—
Rating Filter:
Alert me
|
|
|
[GHC] #2429: error building full dll name to be loaded in ghci#2429: error building full dll name to be loaded in ghci
------------------------+--------------------------------------------------- Reporter: jvl | Owner: Type: bug | Status: new Priority: normal | Component: Driver Version: 6.8.2 | Severity: normal Keywords: dll | Testcase: Architecture: Unknown | Os: Windows ------------------------+--------------------------------------------------- problem encountered ghc 6.8.2 on windows XP. This problem was identified whilst trying to load the wxHaskell dll into ghci. The wxHaskell dll is named wxc-msw2.6.4-0.10.3.dll and will not load into ghci, the following error is encountered {{{ >ghci -fglasgow-exts -lwxc-msw2.6.4-0.10.3 GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Loading object (dynamic) wxc-msw2.6.4-0.10.3 ... failed. Dynamic linker error message was: addDLL: unknown error Whilst trying to load: (dynamic) wxc-msw2.6.4-0.10.3 Directories to search are: : user specified .o/.so/.DLL could not be loaded. }}} yet if we rename the dll {{{ >ghci -fglasgow-exts -lwxc-msw2.6.4-0.10 GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Loading object (dynamic) wxc-msw2.6.4-0.10 ... done final link ... done Prelude> }}} i.e. the final .3.dll causes the problem, exceeding some limit? There appears to be an error in the parsing and construction of the full dll name. i.e. ghci postfix's a .dll adds the path etc, to the base name, but in the above, it can't handle the extra ".3" in the name. ghc in compile mode does not have this error. The above can be verified by constructibe a dummy dll using the following c file, then renaming it appropriately, file wxc.cpp: {{{ #define ROBEXP __declspec(dllexport) void ROBEXP foo_addref(int* f){ } void ROBEXP foo_release(int* f){ } }}} as follows {{{ >cl /LD /MD wxc.cpp >ghci -fglasgow-exts -lwxc GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Loading object (dynamic) wxc ... done final link ... done }}} rename and test {{{ >rename wxc.dll wxc-msw2.6.4-0.10.3.dll >ghci -fglasgow-exts -lwxc-msw2.6.4-0.10.3 GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Loading object (dynamic) wxc-msw2.6.4-0.10.3 ... failed. Dynamic linker error message was: addDLL: unknown error Whilst trying to load: (dynamic) wxc-msw2.6.4-0.10.3 Directories to search are: : user specified .o/.so/.DLL could not be loaded. }}} again the same load error A number of other points ... * dll load error messages * dll load name stuffing == error messages == The same error message "addDLL: unknown error" will apear regardless of whether the error is a genuine "unknown" error or not. The load error reporting should at least distinguish between not being able to locate the dll and other errors, for example, lets give the load a spurious name {{{ >ghci -fglasgow-exts -lwxc-msw2.6.4-X GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Loading object (dynamic) wxc-msw2.6.4-X ... failed. Dynamic linker error message was: addDLL: unknown error Whilst trying to load: (dynamic) wxc-msw2.6.4-X Directories to search are: : user specified .o/.so/.DLL could not be loaded. }}} OK, now lets create a bad dll {{{ >cp wxc.cpp wxc-msw2.6.4-X.dll }}} {{{ >ghci -fglasgow-exts -lwxc-msw2.6.4-X GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Loading object (dynamic) wxc-msw2.6.4-X ... failed. Dynamic linker error message was: addDLL: unknown error Whilst trying to load: (dynamic) wxc-msw2.6.4-X Directories to search are: : user specified .o/.so/.DLL could not be loaded. }}} so we can't even distinguish between a missing dll and a bad one ! == name stuffing == Also, if possible wouldn't it be better for ghci to display the same behaviour as gcc in treatment of dll names i.e. it will postfix .dll, but if that fails it will try the supplied name without any extension, so something like the following will work. {{{ ghci -lsomelib.dll }}} a somewhat related issue #1883 can be treated in the same way, i.e. generate a set of possible target dll full names, by prefixing, postfixing, or in pseudo haskell in the list monad {{{ names base = do prefix <- all_possible_prefix_including_null postfix <- all_possible_postfix_including_null return (prefix++base++postfix) }}} then test each one for load, (gcc does something very similar when searching for .a libraries) -- Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2429> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler _______________________________________________ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@... http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs |
|
|
Re: [GHC] #2429: error building full dll name to be loaded in ghci#2429: error building full dll name to be loaded in ghci
---------------------+------------------------------------------------------ Reporter: jvl | Owner: Type: bug | Status: new Priority: normal | Milestone: 6.10.1 Component: Driver | Version: 6.8.2 Severity: normal | Resolution: Keywords: dll | Difficulty: Unknown Testcase: | Architecture: Unknown Os: Windows | ---------------------+------------------------------------------------------ Changes (by igloo): * difficulty: => Unknown * milestone: => 6.10.1 Comment: Thanks for the report -- Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2429#comment:1> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler _______________________________________________ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@... http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs |
|
|
Re: [GHC] #2429: error building full dll name to be loaded in ghci#2429: error building full dll name to be loaded in ghci
---------------------+------------------------------------------------------ Reporter: jvl | Owner: Type: bug | Status: closed Priority: normal | Milestone: 6.10.1 Component: Driver | Version: 6.8.2 Severity: normal | Resolution: fixed Keywords: dll | Difficulty: Unknown Testcase: | Architecture: Unknown Os: Windows | ---------------------+------------------------------------------------------ Changes (by simonmar): * status: new => closed * resolution: => fixed Comment: I couldn't reproduce the reported bug (with HEAD), but I did improve the error messages. {{{ Wed Sep 3 11:49:51 GMT Daylight Time 2008 Simon Marlow <marlowsd@...> * Windows: print an error message in addDLL Also, look for libXXX.dll in addition to XXX.dll (see #1883, this isn't really a proper fix, but it'll help in some cases). And I tidied up the error message for a DLL load failure, though it's still a bit of a mess because addDLL is supposed to return a (static) string with the error message, but this isn't possible on Windows. Shall I push this patch? (1/2) [ynWvpxdaqjk], or ? for help: Wed Sep 3 11:50:18 GMT Daylight Time 2008 Simon Marlow <marlowsd@...> * sysErrorBelch: don't put an extra \n on Windows }}} -- Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2429#comment:2> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler _______________________________________________ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@... http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs |
| Free Forum Powered by Nabble | Forum Help |