Re: [jamboost] patch boost-base.jam project wide settings

From: Thomas Witt (witt_at_[hidden])
Date: 2002-02-26 11:53:14

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 26 February 2002 17:15, David Abrahams wrote:
> Looking things over more closely...
>
> One thing I'm really unhappy about with this submission is the amount of
> code duplication.

This was almost intentionally. My Jam knowledge is still limited and this was 
the first serious (if you'd like to call it so) piece of code I wrote for it. 
So I tried to reduce the risk of breaking something. This leads to the 
overall design goal to limit the impact on existing code.

I think this is reasonable for a prototype. 

> The current sources are already hard-to-maintain; any new
> code duplication should really be considered unacceptable.

I agree the final version should avoid code duplication. The general idea was: 
Get it going , see whether the feature is accepted, ask for help, refactor 
it.

> Also, why is
> TEMPLATE a target type? It really doesn't seem to me that a template should
> be treated as a target.

Mainly due to historical reasons. I first wanted to hijack 
declare-local-target for template. But therefor I would have had to change 
the argument requirements of declare-local-target. I decided to play it safe 
with regard to existing code.

Thomas

>
> -Dave
>
>
> ----- Original Message -----
> From: "Thomas Witt" <witt_at_[hidden]>
> To: <jamboost_at_[hidden]>
> Sent: Tuesday, February 26, 2002 10:22 AM
> Subject: Re: [jamboost] patch boost-base.jam project wide settings
>
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Dave,
> >
> > On Tuesday 26 February 2002 15:34, David Abrahams wrote:
> > > Hi Thomas,
> > >
> > > your diffs don't appear to correspond to the current CVS state, so I
>
> can't
>
> > > apply your patch.
> >
> > It seems I mixed up the names. Swapping the names in the header should
>
> make
>
> > the problem go away. Anyway I made a new patch that should work, thoughI
>
> use
>
> > diff/patch seldomly so I may still miss a point. Sorry for the
>
> inconvenience.
>
> > > I am wondering a bit why the "template" solution is superior to using
> > > global variables which contain the dependencies.
> >
> > The main reason is, it never came to my mind using global variables for
>
> this.
>
> > One advantage of templates is that they can be located anywhere in the
>
> project
>
> > tree. The other advantage is they are not global variables :-). Seriously
>
> I
>
> > think template is more expressive compared to global variables.
> >
> > > Also, have you considered
> > > how this would interact with the proposed
> > > "project/subproject-as-target" feature?
> >
> > Nope, the intent was to have a temporary solution as long as v2 is not
> > available. My understanding was that development of v1 with regard to new
> > features stopped some time ago. I badly needed to manage global settings
> > and I did not think about global variables.
> >
> > The template "feature" was designed so that current behaviour does not
>
> change.
>
> > Anybody not using templates will not notice any change in behaviour or
> > runtime. So those who think they benefit from the fix can use it and fix
> > their Jamfiles when moving to v2. All others are not affected.
> >
> > > I certainly would like to have documentation before committing
> > > something of this magnitude.
> >
> > This was not intended as a "come'on apply this patch fast" message.
> > I will write the documentation if we/you decide it/something like this
>
> should
>
> > go in. And it's ok for me to commit after documentation work is
> > finished.
> >
> > There is only minimal change to the existing code with regard to
> > template. Nearly the whole functionality is realised in new code. The
> > only "old"
>
> rule
>
> > affected is declare-local-target. If template is dropped I would like to
>
> add
>
> > hooks to declare-local-target to implement it locally.
> >
> > This is not true for the with-command-file change. This is mostly
> > existing code that is modified, but still the changes should not affect
> > current behaviour. This change is not directly related to template. I
> > think it is needed with or without template.
> >
> > Thomas
> >
> > > Thanks for your work,
> > > Dave
> > >
> > >
> > > ----- Original Message -----
> > > From: "Thomas Witt" <witt_at_[hidden]>
> > > To: <jamboost_at_[hidden]>
> > > Sent: Tuesday, February 26, 2002 6:08 AM
> > > Subject: [jamboost] patch boost-base.jam project wide settings
> > >
> > > > -----BEGIN PGP SIGNED MESSAGE-----
> > > > Hash: SHA1
> > > >
> > > >
> > > > Hi,
> > > >
> > > > The attached patch introduces a new rule template that can be used to
> > > > conveniently define project wide requirements. Furthermore
> > > > a change to the with-command-file rule is made to allow for
> > > > additional prefix string that contains non-sources. I need this to
> > >
> > > circumvent
> > >
> > > > cmdline restrictions with long libpathes on windows.
> > > >
> > > > Here is an example of template usage.
> > > >
> > > > #
> > > > # Defining requirements template
> > > > #
> > > > template TplBase
> > > >
> > > >
> > > > # requirements
> > > > <include>../myinclude
> > > > <define>TEMPLATE
> > > > # build not covered so far.
> > > > ;
> > > >
> > > > #
> > > > # Defining another requirements template
> > > > # that inherits requirements
> > > > #
> > > > template TplDerived
> > > >
> > > > # only <template> dependencies go here
> > > > <template>TplBase
> > > >
> > > > <include>../anotherinclude
> > > > <define>DERIVED_TEMPLATE
> > > > ;
> > > >
> > > > #
> > > > # Exe that uses requirements from TplDerived
> > > > #
> > > > exe MyExe
> > > >
> > > > main.cpp
> > > > <template>TplDerived
> > > >
> > > >
> > > > debug
> > > > ;
> > > >
> > > > I am using template in two projects and it reduces the maintenance
>
> work
>
> > > > by at least an order of magnitude. Currently template covers only
> > > > requirements but I think it can be extended to cover build settings
>
> to.
>
> > > The
> > >
> > > > need for requirements was just more pressing :-). If this goes in I
>
> will
>
> > > > happily update the documentation.
> > > >
> > > > Thomas
> > > >
> > > > - --
> > > > Dipl.-Ing. Thomas Witt
> > > > Institut fuer Verkehrswesen, Eisenbahnbau und -betrieb, Universitaet
> > >
> > > Hannover
> > >
> > > > voice: [phone] - 4273, fax: [phone]
> > > > http://www.ive.uni-hannover.de
> > > > -----BEGIN PGP SIGNATURE-----
> > > > Version: GnuPG v1.0.6 (GNU/Linux)
> > > > Comment: For info see http://www.gnupg.org
> > > >
> > > > iD8DBQE8e2yg0ds/gS3XsBoRAlOtAKCGvE3yuhgpMiGQZf4voYt/YQr2CwCfWEoa
> > > > fv8flr9IuLiXrrp6vlELgX0=
> > > > =4NKn
> > > > -----END PGP SIGNATURE-----
> > > >
> > > >
> > > > To unsubscribe from this group, send an email to:
> > > > jamboost-unsubscribe_at_[hidden]
> > > >
> > > >
> > > >
> > > > Your use of Yahoo! Groups is subject to
>
> http://docs.yahoo.com/info/terms/
>
> > > To unsubscribe from this group, send an email to:
> > > jamboost-unsubscribe_at_[hidden]
> > >
> > >
> > >
> > > Your use of Yahoo! Groups is subject to
>
> http://docs.yahoo.com/info/terms/
>
> > - --
> > Dipl.-Ing. Thomas Witt
> > Institut fuer Verkehrswesen, Eisenbahnbau und -betrieb, Universitaet
>
> Hannover
>
> > voice: [phone] - 4273, fax: [phone]
> > http://www.ive.uni-hannover.de
> > -----BEGIN PGP SIGNATURE-----
> > Version: GnuPG v1.0.6 (GNU/Linux)
> > Comment: For info see http://www.gnupg.org
> >
> > iD8DBQE8e6hO0ds/gS3XsBoRAhiiAJ9YzHrlNfOD1FKPCiBhW+1qVYKJhwCdEfGe
> > snJv4THFdkEvCggVPhoFb0A=
> > =m4ai
> > -----END PGP SIGNATURE-----
> >
> >
> > To unsubscribe from this group, send an email to:
> > jamboost-unsubscribe_at_[hidden]
> >
> >
> >
> > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
> To unsubscribe from this group, send an email to:
> jamboost-unsubscribe_at_[hidden]
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
- -- 
Dipl.-Ing. Thomas Witt
Institut fuer Verkehrswesen, Eisenbahnbau und -betrieb, Universitaet Hannover
voice: [phone] - 4273, fax: [phone]
http://www.ive.uni-hannover.de
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8e71+0ds/gS3XsBoRAqMQAKCE7Abvxrh9wZcehosLZ/dJ6FPQNgCfW1GL
1bcjhrT9nDBvlrggeFML/2g=
=JAw/
-----END PGP SIGNATURE-----