On 21.02.2011 11:53, Stephan wrote:
> On 18.02.2011 11:18, Walter Bright wrote:
>> Now with 64 bit Linux support! (Though expect problems with it, it's
>> brand new.)
>>
>>
>> http://www.digitalmars.com/d/1.0/changelog.html
>> http://ftp.digitalmars.com/dmd.1.067.zip
>>
>> http://www.digitalmars.com/d/2.0/changelog.html
>> http://ftp.digitalmars.com/dmd.2.052.zip
>
> Nice release in theorie but since i did not get the time to test the beta it seems some regressions made it into the release that rendered my codebase unbuildable...
>
> The problem is that i cannot reduce it properly. (win32, dmd 2.052)
>
> 1) Without changing anything suddenly i get linker errors. I am building an executable linking in a static lib that contains some methods using std.random.uniform. When i now try to build the executable the linker complains about std.random.uniform not being found. What the heck ? The whole rest of phobos is there and this is missing ??
>

Having hit similar things earlier, I'll tell you my recipe. I usually rebuild *everything* step by step  starting with and including all 3rd party D libraries with new dmd. And making sure nothing from old version sliped on the path. Yes, that's tiresome, especially when you haven't touch those libraries for quite a long time (assuming they are sort of stable).

> 2) This one is really weird. After removing all the std.random.uniform crap from above it compiles and links but crashes right in the beginning befor even entering my main method. This just happens in debug builds. Funny thing is that this is not due to some unittest of mine. And correct me if i am wrong but phobos is shipped without unittests either, right ?
>
> Any help is welcome.
> Stephan


-- 
Dmitry Olshansky

On 21.02.2011 12:34, Dmitry Olshansky wrote:
> On 21.02.2011 11:53, Stephan wrote:
>> On 18.02.2011 11:18, Walter Bright wrote:
>>> Now with 64 bit Linux support! (Though expect problems with it, it's
>>> brand new.)
>>>
>>>
>>> http://www.digitalmars.com/d/1.0/changelog.html
>>> http://ftp.digitalmars.com/dmd.1.067.zip
>>>
>>> http://www.digitalmars.com/d/2.0/changelog.html
>>> http://ftp.digitalmars.com/dmd.2.052.zip
>>
>> Nice release in theorie but since i did not get the time to test the
>> beta it seems some regressions made it into the release that rendered
>> my codebase unbuildable...
>>
>> The problem is that i cannot reduce it properly. (win32, dmd 2.052)
>>
>> 1) Without changing anything suddenly i get linker errors. I am
>> building an executable linking in a static lib that contains some
>> methods using std.random.uniform. When i now try to build the
>> executable the linker complains about std.random.uniform not being
>> found. What the heck ? The whole rest of phobos is there and this is
>> missing ??
>>
>
> Having hit similar things earlier, I'll tell you my recipe. I usually
> rebuild *everything* step by step starting with and including all 3rd
> party D libraries with new dmd. And making sure nothing from old version
> sliped on the path. Yes, that's tiresome, especially when you haven't
> touch those libraries for quite a long time (assuming they are sort of
> stable).
>

I doubt that. That was the first thing i tried coming from a professional C++ work flow ;)


>> 2) This one is really weird. After removing all the std.random.uniform
>> crap from above it compiles and links but crashes right in the
>> beginning befor even entering my main method. This just happens in
>> debug builds. Funny thing is that this is not due to some unittest of
>> mine. And correct me if i am wrong but phobos is shipped without
>> unittests either, right ?
>>
>> Any help is welcome.
>> Stephan
>
>

On 18/02/2011 10:18, Walter Bright wrote:
> Now with 64 bit Linux support! (Though expect problems with it, it's
> brand new.)
>
>
> http://www.digitalmars.com/d/1.0/changelog.html
> http://ftp.digitalmars.com/dmd.1.067.zip
>
> http://www.digitalmars.com/d/2.0/changelog.html
> http://ftp.digitalmars.com/dmd.2.052.zip


Some doc typos:
http://www.digitalmars.com/d/2.0/statement.html
"Error also contains a pointer to the points to the original exception"

-- 
Bruno Medeiros - Software Engineer

On 21/02/11 16:14, Michel Fortin wrote:
> On 2011-02-20 20:21:20 -0500, Graham St Jack <Graham.StJack@internode.on.net> said:
>
>> In particular, are there any plans to re-examine the tail-const issue in light of the compiler patch proposed by Michel Fortin in his post: "const(Object)ref is here!" back in December?
>
> Note that there's now a pull request for that:
> <https://github.com/D-Programming-Language/dmd/pull/3>
>
> And if someone wants to test it, just download and compile the const-object-ref branch of my dmd fork:
> <https://github.com/michelf/dmd/tree/const-object-ref>
>
> I'm currently waiting for feedback from Walter about this (and possibly others who dare to test it before it's in the mainline) before putting more work on it.
>
>
Excellent - hopefully Walter will like it. I will have a go at compiling your branch and trying it out, but my use cases aren't all that stressful.

-- 
Graham St Jack

dsimcha Wrote:

> == Quote from Don (nospam@nospam.com)'s article
> > Walter Bright wrote:
> > > Now with 64 bit Linux support! (Though expect problems with it, it's
> > > brand new.)
> > >
> > >
> > > http://www.digitalmars.com/d/1.0/changelog.html http://ftp.digitalmars.com/dmd.1.067.zip
> > >
> > > http://www.digitalmars.com/d/2.0/changelog.html http://ftp.digitalmars.com/dmd.2.052.zip
> > Eleven man-months to implement a 64-bit backend is pretty impressive, I
> > reckon. Contratulations, Walter!
> > BTW despite the emphasis on D2, this release has one of the highest
> > number of D1 bugfixes, ever.
> 
> Since when was it even 11?  I thought the first 64 commits weren't until June of last year.

The guy has been promising 64 bits since over a year ago. WTF is wrong with you? Not that impressive anymore.

phobophile wrote:
> dsimcha Wrote:
> 
>> == Quote from Don (nospam@nospam.com)'s article
>>> Walter Bright wrote:
>>>> Now with 64 bit Linux support! (Though expect problems with it, it's
>>>> brand new.)
>>>>
>>>>
>>>> http://www.digitalmars.com/d/1.0/changelog.html
>>>> http://ftp.digitalmars.com/dmd.1.067.zip
>>>>
>>>> http://www.digitalmars.com/d/2.0/changelog.html
>>>> http://ftp.digitalmars.com/dmd.2.052.zip
>>> Eleven man-months to implement a 64-bit backend is pretty impressive, I
>>> reckon. Contratulations, Walter!
>>> BTW despite the emphasis on D2, this release has one of the highest
>>> number of D1 bugfixes, ever.
>> Since when was it even 11?  I thought the first 64 commits weren't until June of
>> last year.
> 
> The guy has been promising 64 bits since over a year ago. WTF is wrong with you? Not that impressive anymore.

The first commit was on 21 June 2010. (All that first commit was, was defining the 64 bit register set -- it was really the very beginning of implementation). So it's 8 months today.

On 21.02.2011 09:53, Stephan wrote:
> On 18.02.2011 11:18, Walter Bright wrote:
>> Now with 64 bit Linux support! (Though expect problems with it, it's
>> brand new.)
>>
>>
>> http://www.digitalmars.com/d/1.0/changelog.html
>> http://ftp.digitalmars.com/dmd.1.067.zip
>>
>> http://www.digitalmars.com/d/2.0/changelog.html
>> http://ftp.digitalmars.com/dmd.2.052.zip
>
> Nice release in theorie but since i did not get the time to test the
> beta it seems some regressions made it into the release that rendered my
> codebase unbuildable...
>
> The problem is that i cannot reduce it properly. (win32, dmd 2.052)
>
> 1) Without changing anything suddenly i get linker errors. I am building
> an executable linking in a static lib that contains some methods using
> std.random.uniform. When i now try to build the executable the linker
> complains about std.random.uniform not being found. What the heck ? The
> whole rest of phobos is there and this is missing ??

Turns out the root of all evil are the version (ddoc) blocks in phobos. what a great idea to change method SIGNATURES using them.
as soon as one uses those methods and tries to build ddocs out of the own code (using -D) methods get referenced that are not actually in the phobos lib. great idea since the signatures differ !
manifests in this bug btw:
http://d.puremagic.com/issues/show_bug.cgi?id=5642

I wish i had tested the beta before release..


> 2) This one is really weird. After removing all the std.random.uniform
> crap from above it compiles and links but crashes right in the beginning
> befor even entering my main method. This just happens in debug builds.
> Funny thing is that this is not due to some unittest of mine. And
> correct me if i am wrong but phobos is shipped without unittests either,
> right ?

I still have no clue what this is about.

> Any help is welcome.
> Stephan

On Tuesday, February 22, 2011 16:17:39 Extrawurst wrote:
> On 21.02.2011 09:53, Stephan wrote:
> > On 18.02.2011 11:18, Walter Bright wrote:
> >> Now with 64 bit Linux support! (Though expect problems with it, it's
> >> brand new.)
> >> 
> >> 
> >> http://www.digitalmars.com/d/1.0/changelog.html http://ftp.digitalmars.com/dmd.1.067.zip
> >> 
> >> http://www.digitalmars.com/d/2.0/changelog.html http://ftp.digitalmars.com/dmd.2.052.zip
> > 
> > Nice release in theorie but since i did not get the time to test the beta it seems some regressions made it into the release that rendered my codebase unbuildable...
> > 
> > The problem is that i cannot reduce it properly. (win32, dmd 2.052)
> > 
> > 1) Without changing anything suddenly i get linker errors. I am building an executable linking in a static lib that contains some methods using std.random.uniform. When i now try to build the executable the linker complains about std.random.uniform not being found. What the heck ? The whole rest of phobos is there and this is missing ??
> 
> Turns out the root of all evil are the version (ddoc) blocks in phobos.
> what a great idea to change method SIGNATURES using them.
> as soon as one uses those methods and tries to build ddocs out of the
> own code (using -D) methods get referenced that are not actually in the
> phobos lib. great idea since the signatures differ !
> manifests in this bug btw:
> http://d.puremagic.com/issues/show_bug.cgi?id=5642
> 
> I wish i had tested the beta before release..

You pretty much _have_ to do that in a number of cases. For instance, if a function only exists on Linux or or it only exists on Windows, you have to have a different version block for the documentation. There's no way around that.

Now, given that Phobos and druntime are the most likely places to have to worry about that and that many programs don't need to do that sort of thing, Phobos and druntime are going to be changed so that they use a different version identifier for their documentation blocks, so you'll only get the problem if you yourself are forced to have separate documentation blocks. Andrei needs to change the makefiles before the version identifier in the code can be changed, and he didn't do that before the release (we only figured out that we needed it shortly before dmd went into beta, and I believe that Andrei was pretty busy while it was in beta, so he didn't make the change).

Regardless, Phobos does it by necessity. Separate blocks for documentation are required in the cases where it uses them.

- Jonathan M Davis