RetroBBS - comp.lang.xharbour - Re: A problem with HB

A problem with HB_FUNC

<uebrlr$29rml$1@dont-email.me>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17827&group=comp.lang.xharbour#17827

copy link Newsgroups: comp.lang.xharbour

by: Enrico Maria Giordan - Tue, 19 Sep 2023 10:03 UTC

This sample works fine with Harbour but not with xHarbour. Ron, can you
try to fix it, please?

FUNCTION MAIN()

RETURN NIL

#pragma BEGINDUMP

#ifdef __MYTEST__

HB_FUNC( SCAVENGEAL )
{ }

#endif

#pragma ENDDUMP

With xHarbour:

error LNK2001: unresolved external symbol _HB_FUN_SCAVENGEAL

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Re: A problem with HB_FUNC

<uebro9$29rml$2@dont-email.me>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17828&group=comp.lang.xharbour#17828

copy link Newsgroups: comp.lang.xharbour

Path: i2pn2.org!i2pn.org!eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: e.m.giordano@emagsoftware.it (Enrico Maria Giordano)
Newsgroups: comp.lang.xharbour
Subject: Re: A problem with HB_FUNC
Date: Tue, 19 Sep 2023 12:04:57 +0200
Organization: A noiseless patient Spider
Lines: 16
Message-ID: <uebro9$29rml$2@dont-email.me>
References: <uebrlr$29rml$1@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Tue, 19 Sep 2023 10:04:57 -0000 (UTC)
Injection-Info: dont-email.me; posting-host="94201aaebae2902b8e403ec493ded080";
logging-data="2420437"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/B3o5VBc6sHo3zhz6mNsCR"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:Qd9Xpv6IOjd7MPgcwPQesv0Gru0=
Content-Language: it
In-Reply-To: <uebrlr$29rml$1@dont-email.me>

by: Enrico Maria Giordan - Tue, 19 Sep 2023 10:04 UTC

Il 19/09/2023 12:03, Enrico Maria Giordano ha scritto:

> HB_FUNC( SCAVENGEAL )

This function name is irrelevant. It errors out with any function name.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Re: A problem with HB_FUNC

<4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17829&group=comp.lang.xharbour#17829

copy link Newsgroups: comp.lang.xharbour

X-Received: by 2002:a05:620a:878d:b0:76f:c32:705f with SMTP id py13-20020a05620a878d00b0076f0c32705fmr16043qkn.2.1695164563265;
Tue, 19 Sep 2023 16:02:43 -0700 (PDT)
X-Received: by 2002:a05:6808:1a13:b0:3ac:ab4f:ee6 with SMTP id
bk19-20020a0568081a1300b003acab4f0ee6mr402318oib.1.1695164562796; Tue, 19 Sep
2023 16:02:42 -0700 (PDT)
Path: i2pn2.org!i2pn.org!usenet.blueworldhosting.com!diablo1.usenet.blueworldhosting.com!3.us.feeder.erje.net!feeder.erje.net!border-1.nntp.ord.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.xharbour
Date: Tue, 19 Sep 2023 16:02:42 -0700 (PDT)
In-Reply-To: <uebrlr$29rml$1@dont-email.me>
Injection-Info: google-groups.googlegroups.com; posting-host=2806:10be:8:467e:95d8:54f:210d:9f1a;
posting-account=8KTLLQoAAABrqmjk_im031yR_2Rimshb
NNTP-Posting-Host: 2806:10be:8:467e:95d8:54f:210d:9f1a
References: <uebrlr$29rml$1@dont-email.me>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>
Subject: Re: A problem with HB_FUNC
From: ronpinkas@gmail.com (Ron Pinkas)
Injection-Date: Tue, 19 Sep 2023 23:02:43 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Lines: 50

by: Ron Pinkas - Tue, 19 Sep 2023 23:02 UTC

On Tuesday, September 19, 2023 at 5:03:43 AM UTC-5, Enrico Maria Giordano wrote:
> This sample works fine with Harbour but not with xHarbour. Ron, can you
> try to fix it, please?
>
>
> FUNCTION MAIN()
>
> RETURN NIL
>
>
> #pragma BEGINDUMP
>
> #ifdef __MYTEST__
>
> HB_FUNC( SCAVENGEAL )
> {
> }
>
> #endif
>
> #pragma ENDDUMP
>
>
> With xHarbour:
>
> error LNK2001: unresolved external symbol _HB_FUN_SCAVENGEAL
>
> --
> Enrico Maria Giordano
>
> http://www.emagsoftware.it
> http://www.emagsoftware.it/emgmusic
> http://www.emagsoftware.it/spectrum
> http://www.emagsoftware.it/tbosg

Enrico,

The issue is because the function is not included due to the #ifdef. Technically the dump scanner can attempt to parse #ifdef to avoid generation of the entry into the symbol table but I am not sure this is desired. Can you please provide a real world sample where such case can exist intentionally? I mean if the function is in the BEGINDUMP than it will likely also have a prg code calling it so I don't understand what can be a real world case of such issue. Even if there was such case, I suppose it is much simpler to have a PRG level #ifdef around the #pragma BEGINDUMP. - or maybe I just don't understand the problem.

Ron

Re: A problem with HB_FUNC

<ueee69$2sedi$1@dont-email.me>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17830&group=comp.lang.xharbour#17830

copy link Newsgroups: comp.lang.xharbour

Path: i2pn2.org!i2pn.org!eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: e.m.giordano@emagsoftware.it (Enrico Maria Giordano)
Newsgroups: comp.lang.xharbour
Subject: Re: A problem with HB_FUNC
Date: Wed, 20 Sep 2023 11:31:55 +0200
Organization: A noiseless patient Spider
Lines: 38
Message-ID: <ueee69$2sedi$1@dont-email.me>
References: <uebrlr$29rml$1@dont-email.me>
<4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Wed, 20 Sep 2023 09:31:53 -0000 (UTC)
Injection-Info: dont-email.me; posting-host="2fd0b81a6161d53147ab7541d39ff69d";
logging-data="3029426"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+60+miyVbEeG84GlziO8NT"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:jFY3S0LmsFFbNPc0Xfm2RrTz21Y=
Content-Language: it
In-Reply-To: <4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>

by: Enrico Maria Giordan - Wed, 20 Sep 2023 09:31 UTC

Il 20/09/2023 01:02, Ron Pinkas ha scritto:

> The issue is because the function is not included due to the #ifdef. Technically the dump scanner can attempt to parse #ifdef to avoid generation of the entry into the symbol table but I am not sure this is desired. Can you please provide a real world sample where such case can exist intentionally? I mean if the function is in the BEGINDUMP than it will likely also have a prg code calling it so I don't understand what can be a real world case of such issue. Even if there was such case, I suppose it is much simpler to have a PRG level #ifdef around the #pragma BEGINDUMP. - or maybe I just don't understand the problem.

Imagine you want to define a function only for 64-bit mode compilation.
Well, you can't:

FUNCTION MAIN()

RETURN NIL

#pragma BEGINDUMP

#ifdef _WIN64

HB_FUNC( MYFUNC64 )
{ }

#endif

#pragma ENDDUMP

In 32-bit mode:

error LNK2001: unresolved external symbol _HB_FUN_MYFUNC64

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Re: A problem with HB_FUNC

<3518150f-f447-446f-9b99-a4aab7292ecan@googlegroups.com>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17831&group=comp.lang.xharbour#17831

copy link Newsgroups: comp.lang.xharbour

X-Received: by 2002:a05:622a:1a96:b0:412:2646:9995 with SMTP id s22-20020a05622a1a9600b0041226469995mr36356qtc.10.1695219216864;
Wed, 20 Sep 2023 07:13:36 -0700 (PDT)
X-Received: by 2002:a05:6808:3288:b0:3ad:fc2e:fbc6 with SMTP id
cg8-20020a056808328800b003adfc2efbc6mr31440oib.10.1695219216447; Wed, 20 Sep
2023 07:13:36 -0700 (PDT)
Path: i2pn2.org!i2pn.org!usenet.blueworldhosting.com!diablo1.usenet.blueworldhosting.com!peer01.iad!feed-me.highwinds-media.com!news.highwinds-media.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.xharbour
Date: Wed, 20 Sep 2023 07:13:36 -0700 (PDT)
In-Reply-To: <ueee69$2sedi$1@dont-email.me>
Injection-Info: google-groups.googlegroups.com; posting-host=2806:10be:8:467e:95d8:54f:210d:9f1a;
posting-account=8KTLLQoAAABrqmjk_im031yR_2Rimshb
NNTP-Posting-Host: 2806:10be:8:467e:95d8:54f:210d:9f1a
References: <uebrlr$29rml$1@dont-email.me> <4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>
<ueee69$2sedi$1@dont-email.me>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <3518150f-f447-446f-9b99-a4aab7292ecan@googlegroups.com>
Subject: Re: A problem with HB_FUNC
From: ronpinkas@gmail.com (Ron Pinkas)
Injection-Date: Wed, 20 Sep 2023 14:13:36 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Received-Bytes: 2777

by: Ron Pinkas - Wed, 20 Sep 2023 14:13 UTC

On Wednesday, September 20, 2023 at 4:31:56 AM UTC-5, Enrico Maria Giordano wrote:
> Il 20/09/2023 01:02, Ron Pinkas ha scritto:
>
> > The issue is because the function is not included due to the #ifdef. Technically the dump scanner can attempt to parse #ifdef to avoid generation of the entry into the symbol table but I am not sure this is desired. Can you please provide a real world sample where such case can exist intentionally? I mean if the function is in the BEGINDUMP than it will likely also have a prg code calling it so I don't understand what can be a real world case of such issue. Even if there was such case, I suppose it is much simpler to have a PRG level #ifdef around the #pragma BEGINDUMP. - or maybe I just don't understand the problem.
> Imagine you want to define a function only for 64-bit mode compilation.
> Well, you can't:
> FUNCTION MAIN()
>
> RETURN NIL
>
>
> #pragma BEGINDUMP
> #ifdef _WIN64
>
> HB_FUNC( MYFUNC64 )
> {
> }
>
> #endif
>
> #pragma ENDDUMP
>
> In 32-bit mode:
>
> error LNK2001: unresolved external symbol _HB_FUN_MYFUNC64
> --
> Enrico Maria Giordano
>
> http://www.emagsoftware.it
> http://www.emagsoftware.it/emgmusic
> http://www.emagsoftware.it/spectrum
> http://www.emagsoftware.it/tbosg

Enrico,

Of course you can:

#ifdef WIN64

#pragma BEGINDUMP

HB_FUNC( SCAVENGEAL )
{ }

#endif

#pragma ENDDUMP

#endif

Re: A problem with HB_FUNC

<ueevf7$2vdvr$1@dont-email.me>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17832&group=comp.lang.xharbour#17832

copy link Newsgroups: comp.lang.xharbour

Path: i2pn2.org!i2pn.org!eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: e.m.giordano@emagsoftware.it (Enrico Maria Giordano)
Newsgroups: comp.lang.xharbour
Subject: Re: A problem with HB_FUNC
Date: Wed, 20 Sep 2023 16:26:49 +0200
Organization: A noiseless patient Spider
Lines: 72
Message-ID: <ueevf7$2vdvr$1@dont-email.me>
References: <uebrlr$29rml$1@dont-email.me>
<4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>
<ueee69$2sedi$1@dont-email.me>
<3518150f-f447-446f-9b99-a4aab7292ecan@googlegroups.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Wed, 20 Sep 2023 14:26:47 -0000 (UTC)
Injection-Info: dont-email.me; posting-host="2fd0b81a6161d53147ab7541d39ff69d";
logging-data="3127291"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/knq05L79ph9Seof/aozBd"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:F3qEomVZ7U6SwyQQWQNwEs5zpWU=
Content-Language: it
In-Reply-To: <3518150f-f447-446f-9b99-a4aab7292ecan@googlegroups.com>

by: Enrico Maria Giordan - Wed, 20 Sep 2023 14:26 UTC

Il 20/09/2023 16:13, Ron Pinkas ha scritto:

>> Imagine you want to define a function only for 64-bit mode compilation.
>> Well, you can't:
>>
>> FUNCTION MAIN()
>>
>> RETURN NIL
>>
>>
>> #pragma BEGINDUMP
>> #ifdef _WIN64
>>
>> HB_FUNC( MYFUNC64 )
>> {
>> }
>>
>> #endif
>>
>> #pragma ENDDUMP
>>
>> In 32-bit mode:
>>
>> error LNK2001: unresolved external symbol _HB_FUN_MYFUNC64
>
> Of course you can:
>
> #ifdef WIN64
>
> #pragma BEGINDUMP
>
> HB_FUNC( SCAVENGEAL )
> {
> }
>
> #endif
>
> #pragma ENDDUMP
>
> #endif

Yes, I know, but with Harbour it works in this form too:

FUNCTION MAIN()

RETURN NIL

#pragma BEGINDUMP

#ifdef _WIN64

HB_FUNC( MYFUNC64 )
{ }

#endif

#pragma ENDDUMP

Can we fix it?

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Re: A problem with HB_FUNC

<f7930fd8-4f71-47d8-a866-ccbeb9aac5c4n@googlegroups.com>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17833&group=comp.lang.xharbour#17833

copy link Newsgroups: comp.lang.xharbour

X-Received: by 2002:a05:620a:8f01:b0:773:f7e8:efff with SMTP id rh1-20020a05620a8f0100b00773f7e8efffmr16975qkn.9.1695222461586;
Wed, 20 Sep 2023 08:07:41 -0700 (PDT)
X-Received: by 2002:a05:6870:3e4:b0:1d5:c5a1:f18f with SMTP id
h36-20020a05687003e400b001d5c5a1f18fmr900498oaf.11.1695222461222; Wed, 20 Sep
2023 08:07:41 -0700 (PDT)
Path: i2pn2.org!i2pn.org!weretis.net!feeder6.news.weretis.net!border-2.nntp.ord.giganews.com!border-1.nntp.ord.giganews.com!nntp.giganews.com!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.xharbour
Date: Wed, 20 Sep 2023 08:07:40 -0700 (PDT)
In-Reply-To: <ueevf7$2vdvr$1@dont-email.me>
Injection-Info: google-groups.googlegroups.com; posting-host=2806:10be:8:467e:95d8:54f:210d:9f1a;
posting-account=8KTLLQoAAABrqmjk_im031yR_2Rimshb
NNTP-Posting-Host: 2806:10be:8:467e:95d8:54f:210d:9f1a
References: <uebrlr$29rml$1@dont-email.me> <4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>
<ueee69$2sedi$1@dont-email.me> <3518150f-f447-446f-9b99-a4aab7292ecan@googlegroups.com>
<ueevf7$2vdvr$1@dont-email.me>
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <f7930fd8-4f71-47d8-a866-ccbeb9aac5c4n@googlegroups.com>
Subject: Re: A problem with HB_FUNC
From: ronpinkas@gmail.com (Ron Pinkas)
Injection-Date: Wed, 20 Sep 2023 15:07:41 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Lines: 105

by: Ron Pinkas - Wed, 20 Sep 2023 15:07 UTC

On Wednesday, September 20, 2023 at 9:26:50 AM UTC-5, Enrico Maria Giordano wrote:
> Il 20/09/2023 16:13, Ron Pinkas ha scritto:
>
> >> Imagine you want to define a function only for 64-bit mode compilation..
> >> Well, you can't:
> >>
> >> FUNCTION MAIN()
> >>
> >> RETURN NIL
> >>
> >>
> >> #pragma BEGINDUMP
> >> #ifdef _WIN64
> >>
> >> HB_FUNC( MYFUNC64 )
> >> {
> >> }
> >>
> >> #endif
> >>
> >> #pragma ENDDUMP
> >>
> >> In 32-bit mode:
> >>
> >> error LNK2001: unresolved external symbol _HB_FUN_MYFUNC64
> >
> > Of course you can:
> >
> > #ifdef WIN64
> >
> > #pragma BEGINDUMP
> >
> > HB_FUNC( SCAVENGEAL )
> > {
> > }
> >
> > #endif
> >
> > #pragma ENDDUMP
> >
> > #endif
> Yes, I know, but with Harbour it works in this form too:
> FUNCTION MAIN()
>
> RETURN NIL
>
>
> #pragma BEGINDUMP
>
> #ifdef _WIN64
>
> HB_FUNC( MYFUNC64 )
> {
> }
>
> #endif
>
> #pragma ENDDUMP
> Can we fix it?
> --
> Enrico Maria Giordano
>
> http://www.emagsoftware.it
> http://www.emagsoftware.it/emgmusic
> http://www.emagsoftware.it/spectrum
> http://www.emagsoftware.it/tbosg

Enrico, I believe the "fix" is wrong - As the documentation in inline_c.txt clearly states:

-------
#PRAGMA BEGINDUMP
* ...
* [#PRAGMA ENDDUMP]|EOF
*
* The transition from Harbour to C* occurs when encountering the FIRST NewLine token, at the end of the #pragme BEGINDUMP directive line. The transition from C* back to Harbour occurs when encountering the FIRST NewLine token, at the *end* of the FIRST FOUND line containing a #pragma ENDDUMP directive, or at the EOF.
*
* There may be any number of lines in such C* block. Any legal C* syntax including #directives, and any number of nested balanced {} is allowed.
*
* Misc:
*
*
* C* Syntax errors and warnings, are reported with the correct PRG source file name, and line number.
*
* No parsing of the C* code is being performed, and thus MACROS of unbalanced '{}' are allowed.
*
* This implementation is actually language independent, and will allow inclusion of any code compatible with the generated output

-----

It's up to the Programmer to make sure that the dumped section is valid in the context of the PRG compilation unit. Also, since such function clearly must be called from the PRG code at SOME TIME to be of value, then the PRG code *MUST* have such #ifdef anyway!

Ron

Re: A problem with HB_FUNC

<uef2o7$302qv$1@dont-email.me>

copy mid

https://www.rocksolidbbs.com/devel/article-flat.php?id=17834&group=comp.lang.xharbour#17834

copy link Newsgroups: comp.lang.xharbour

Path: i2pn2.org!i2pn.org!eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: e.m.giordano@emagsoftware.it (Enrico Maria Giordano)
Newsgroups: comp.lang.xharbour
Subject: Re: A problem with HB_FUNC
Date: Wed, 20 Sep 2023 17:22:49 +0200
Organization: A noiseless patient Spider
Lines: 16
Message-ID: <uef2o7$302qv$1@dont-email.me>
References: <uebrlr$29rml$1@dont-email.me>
<4921407f-b649-44c5-ad0a-a780c5fe90f5n@googlegroups.com>
<ueee69$2sedi$1@dont-email.me>
<3518150f-f447-446f-9b99-a4aab7292ecan@googlegroups.com>
<ueevf7$2vdvr$1@dont-email.me>
<f7930fd8-4f71-47d8-a866-ccbeb9aac5c4n@googlegroups.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Wed, 20 Sep 2023 15:22:47 -0000 (UTC)
Injection-Info: dont-email.me; posting-host="2fd0b81a6161d53147ab7541d39ff69d";
logging-data="3148639"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/2oRqaPNV5OMjl7C4odQJH"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:fFUqI3lFopzuPe+EW7h+aXdi5Hs=
In-Reply-To: <f7930fd8-4f71-47d8-a866-ccbeb9aac5c4n@googlegroups.com>
Content-Language: it

by: Enrico Maria Giordan - Wed, 20 Sep 2023 15:22 UTC

Il 20/09/2023 17:07, Ron Pinkas ha scritto:

> Enrico, I believe the "fix" is wrong

Ok, no problem.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Elliptic paraboloids for sale.

devel / comp.lang.xharbour / Re: A problem with HB_FUNC

Subject	Author
A problem with HB_FUNC	Enrico Maria Giordano
Re: A problem with HB_FUNC	Enrico Maria Giordano
Re: A problem with HB_FUNC	Ron Pinkas
Re: A problem with HB_FUNC	Enrico Maria Giordano
Re: A problem with HB_FUNC	Ron Pinkas
Re: A problem with HB_FUNC	Enrico Maria Giordano
Re: A problem with HB_FUNC	Ron Pinkas
Re: A problem with HB_FUNC	Enrico Maria Giordano