Hi,
Can someone point me to some source with information about name demangling when compiling some D program and the famous linker error 42 appears?
Thank

On 2012-10-11 15:17, Lubos Pintes wrote:
> Hi,
> Can someone point me to some source with information about name
> demangling when compiling some D program and the famous linker error 42
> appears?
> Thank

This page contains the reference for the name mangling:

http://dlang.org/abi.html

-- 
/Jacob Carlborg

On Oct 11, 2012, at 6:17 AM, Lubos Pintes <lubos.pintes@gmail.com> wrote:

> Hi,
> Can someone point me to some source with information about name demangling when compiling some D program and the famous linker error 42 appears?

Filter the symbol names through core.demangle.

On 11-10-2012 22:56, Sean Kelly wrote:
> On Oct 11, 2012, at 6:17 AM, Lubos Pintes <lubos.pintes@gmail.com> wrote:
>
>> Hi,
>> Can someone point me to some source with information about name demangling when compiling some D program and the famous linker error 42 appears?
>
> Filter the symbol names through core.demangle.
>

We even have a tool for that: https://github.com/D-Programming-Language/tools/blob/master/ddemangle.d

-- 
Alex Rønne Petersen
alex@lycus.org
http://lycus.org

On Thursday, 11 October 2012 at 21:04:28 UTC, Alex Rønne Petersen wrote:
> On 11-10-2012 22:56, Sean Kelly wrote:
>> On Oct 11, 2012, at 6:17 AM, Lubos Pintes <lubos.pintes@gmail.com> wrote:
>>
>>> Hi,
>>> Can someone point me to some source with information about name demangling when compiling some D program and the famous linker error 42 appears?
>>
>> Filter the symbol names through core.demangle.
>>
>
> We even have a tool for that: https://github.com/D-Programming-Language/tools/blob/master/ddemangle.d

Does this work and if so how? I've seen a thread about the demangling tool not working on types, just functions. The comment at top of this file says
"Replaces *all* occurrences of mangled D symbols in the input..." so I assume it deals with types? The help says: 'If <inputfile> is a single dash '-', standard input is read.'

I've tried:
$ echo "this & _D56TypeInfo_xS8plusauri6bsheet13balance_sheet12BalanceSheet6__initZ unq & _D55TypeInfo_S8plusauri6bsheet13balance_sheet12BalanceSheet6__initZ" > ./goo.txt
$ ./ddmangle ./goo.txt

which just hangs, and this - which does not demangle:

$ echo "this & _D56TypeInfo_xS8plusauri6bsheet13balance_sheet12BalanceSheet6__initZ unq & _D55TypeInfo_S8plusauri6bsheet13balance_sheet12BalanceSheet6__initZ" | ./ddmangle
this & _D56TypeInfo_xS8plusauri6bsheet13balance_sheet12BalanceSheet6__initZ unq & _D55TypeInfo_S8plusauri6bsheet13balance_sheet12BalanceSheet6__initZ

What is the correct way to do this?

Thanks
Dan

On Thu, Nov 01, 2012 at 03:16:25PM +0100, Dan wrote:
> On Thursday, 11 October 2012 at 21:04:28 UTC, Alex Rønne Petersen wrote:
> >On 11-10-2012 22:56, Sean Kelly wrote:
> >>On Oct 11, 2012, at 6:17 AM, Lubos Pintes <lubos.pintes@gmail.com> wrote:
> >>
> >>>Hi,
> >>>Can someone point me to some source with information about
> >>>name demangling when compiling some D program and the famous
> >>>linker error 42 appears?
> >>
> >>Filter the symbol names through core.demangle.
> >>
> >
> >We even have a tool for that: https://github.com/D-Programming-Language/tools/blob/master/ddemangle.d
> 
> Does this work and if so how? I've seen a thread about the demangling tool not working on types, just functions. The comment at top of this file says "Replaces *all* occurrences of mangled D symbols in the input..." so I assume it deals with types? The help says: 'If <inputfile> is a single dash '-', standard input is read.'
[...]

This program only calls the library core.demangle. So if core.demangle doesn't demangle something, then it won't get demangled. The problem is, there's a comment in core.demangle that seems to say that some mangled symbols are skipped because they are not pertinent to the ABI, or something to that effect. Meaning that *not* all symbols are getting demangled.

I think this should be an enhancement request in the bugtracker.


T

-- 
Let's call it an accidental feature. -- Larry Wall

On Thu, 01 Nov 2012 15:40:23 -0000, H. S. Teoh <hsteoh@quickfur.ath.cx> wrote:

> On Thu, Nov 01, 2012 at 03:16:25PM +0100, Dan wrote:
>> On Thursday, 11 October 2012 at 21:04:28 UTC, Alex Rønne Petersen
>> wrote:
>> >On 11-10-2012 22:56, Sean Kelly wrote:
>> >>On Oct 11, 2012, at 6:17 AM, Lubos Pintes
>> >><lubos.pintes@gmail.com> wrote:
>> >>
>> >>>Hi,
>> >>>Can someone point me to some source with information about
>> >>>name demangling when compiling some D program and the famous
>> >>>linker error 42 appears?
>> >>
>> >>Filter the symbol names through core.demangle.
>> >>
>> >
>> >We even have a tool for that:
>> >https://github.com/D-Programming-Language/tools/blob/master/ddemangle.d
>>
>> Does this work and if so how? I've seen a thread about the demangling
>> tool not working on types, just functions. The comment at top of this
>> file says "Replaces *all* occurrences of mangled D symbols in the
>> input..." so I assume it deals with types? The help says: 'If
>> <inputfile> is a single dash '-', standard input is read.'
> [...]
>
> This program only calls the library core.demangle. So if core.demangle
> doesn't demangle something, then it won't get demangled. The problem is,
> there's a comment in core.demangle that seems to say that some mangled
> symbols are skipped because they are not pertinent to the ABI, or
> something to that effect. Meaning that *not* all symbols are getting
> demangled.
>
> I think this should be an enhancement request in the bugtracker.

There is also a bug in demangle.d, line 55:

foreach (line; stdin.byLine())

should read:

foreach (line; f.byLine())

R

-- 
Using Opera's revolutionary email client: http://www.opera.com/mail/

On Thu, 01 Nov 2012 15:40:23 -0000, H. S. Teoh <hsteoh@quickfur.ath.cx> wrote:

> On Thu, Nov 01, 2012 at 03:16:25PM +0100, Dan wrote:
>> On Thursday, 11 October 2012 at 21:04:28 UTC, Alex Rønne Petersen
>> wrote:
>> >On 11-10-2012 22:56, Sean Kelly wrote:
>> >>On Oct 11, 2012, at 6:17 AM, Lubos Pintes
>> >><lubos.pintes@gmail.com> wrote:
>> >>
>> >>>Hi,
>> >>>Can someone point me to some source with information about
>> >>>name demangling when compiling some D program and the famous
>> >>>linker error 42 appears?
>> >>
>> >>Filter the symbol names through core.demangle.
>> >>
>> >
>> >We even have a tool for that:
>> >https://github.com/D-Programming-Language/tools/blob/master/ddemangle.d
>>
>> Does this work and if so how? I've seen a thread about the demangling
>> tool not working on types, just functions. The comment at top of this
>> file says "Replaces *all* occurrences of mangled D symbols in the
>> input..." so I assume it deals with types? The help says: 'If
>> <inputfile> is a single dash '-', standard input is read.'
> [...]
>
> This program only calls the library core.demangle. So if core.demangle
> doesn't demangle something, then it won't get demangled. The problem is,
> there's a comment in core.demangle that seems to say that some mangled
> symbols are skipped because they are not pertinent to the ABI, or
> something to that effect. Meaning that *not* all symbols are getting
> demangled.
>
> I think this should be an enhancement request in the bugtracker.

Also, running this:

dmd test.d | demangle.exe -

on windows, results in this:

invalid UTF-8 sequence

Usage: demangle.exe [options] <inputfile>
Demangles all occurrences of mangled D symbols in the input and writes to
standard output.
If <inputfile> is a single dash '-', standard input is read.
Options:
    --help, -h    Show this help

Adding writefln to demangle.d to debug the issue shows that args[1] is "-" and that getopt is throwing the exception "invalid UTF-8 sequence".

Omitting the - works, and demangle reads stdin.

R

-- 
Using Opera's revolutionary email client: http://www.opera.com/mail/

On Thu, Nov 01, 2012 at 04:25:56PM -0000, Regan Heath wrote:
> On Thu, 01 Nov 2012 15:40:23 -0000, H. S. Teoh <hsteoh@quickfur.ath.cx> wrote:
> 
> >On Thu, Nov 01, 2012 at 03:16:25PM +0100, Dan wrote:
> >>On Thursday, 11 October 2012 at 21:04:28 UTC, Alex Rønne Petersen wrote:
> >>>On 11-10-2012 22:56, Sean Kelly wrote:
[...]
> >>>We even have a tool for that: https://github.com/D-Programming-Language/tools/blob/master/ddemangle.d
[...]
> Also, running this:
> 
> dmd test.d | demangle.exe -
> 
> on windows, results in this:
> 
> invalid UTF-8 sequence
[...]
> Omitting the - works, and demangle reads stdin.
[...]

https://github.com/D-Programming-Language/tools/pull/35


T

-- 
English has the lovely word "defenestrate", meaning "to execute by throwing someone out a window", or more recently "to remove Windows from a computer and replace it with something useful". :-) -- John Cowan

On Thu, 01 Nov 2012 16:25:56 -0000, Regan Heath <regan@netmail.co.nz> wrote:
> Adding writefln to demangle.d to debug the issue shows that args[1] is "-" and that getopt is throwing the exception "invalid UTF-8 sequence".

I was wrong here, this line is the issue:
  stderr.writeln(e.msg);

For some reason, when run from command prompt and passed "-" as an arg it errors.

If I run it in debug mode in VisualD it works and outputs:
  Unrecognized option -

R

-- 
Using Opera's revolutionary email client: http://www.opera.com/mail/