binarycloud-dev — development list for binarycloud

Re: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

On Mon, Oct 08, 2001 at 11:49:13AM -0700, bin...@li... wrote:
> Hey dave,
> 
> if you want sablot to work fine, (and you're on a linux box)
Nope, BSD baby. Though may last set of production machines were all
Slack boxes and run great, I prefer BSD.

> 
> http://sourceforge.net/project/showfiles.php?group_id=16402&release_id=52725
> 
> doenload that tarball and compile it from scratch.
> 
> that package is a little thing I made up that will properly compile and
> install apache (comes with a bunch of configure scripts for each package.
> 
> I have sab .5 working fine with php406/apache 1320
> 
> :)
> 
> _alex

Looking forward to getting this done up and running, I know I talked to
you a loooong time ago but didn't have the time to investigate BC. I
looked hard at Tomcat but I find PHP does pretty much everything faster
and with less effort/constraints. The one advantage I saw with Tomcat was
a structured environment. Thats what I think BC will give me.

DAve

-- 
My other computer is your Windows machine... 

Re: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

On Mon, Oct 08, 2001 at 08:26:55PM +0200, bin...@li... wrote:
> Hi Dave,
> 
> >> >I'm a new user and maybe out of line here but...
> >> You are not! :-)
> >Well........ New to BC!
> 
> But not out of line (if i've understood the term right in sense of "my
> opinion does not matter", i'm not a native speaker).
I was only remarking that I may be speaking on a subject already covered
in depth. I Didn't want to bring up an old topic. You understood correctly
so don't appologize for being bi-lingual, which I am not ;^)
> 
> 
> >Got that already in PHP ;^) If I could ever get Sablotron
> >to work I'm not sure I would need/want *.jsp.
> That was ironic a ironic remark. But what sablot concerns, what does not
> work with your setup ?
That would be off-topic, but I have issues with the Sablot build and
iconv.h, It's a FreeBSD issue, but frustrating non the less.
> 
> 
> >> That's a point. On the other hand you need a working
> >> make and all related tools. The advantage here is that
> >> we can provide jre and a env setting script with bc.
> >> And if everything is packaged and configured well, you don't
> >> need to install something optional.
> 
> >Yes, but that was already installed and in use for the
> >multitude of other tools/apps I have/will install. Agreed you
> >could package everything together, but that moots the point of
> >moving to ANT so developers can move
> >on with other BC things. I don't want them to get tied up
> >developing an ANT package installer under the assumption
> >it will save time over developing a PHP based installer.
> 
> Of course, complete PHP based make would be really killer. I personally
> preffer that much over all other options.
> 
> >>
> >> >Not real sure I would want to learn Java in order to use PHP ;^)
> >> No, no. You don't need to know anything about coding in java or something
> >> like that, as well as you don't have to know make/autoconf stuff
> >to make bc
> >> or other projects now. They only think you have to do is edit
> >MakeConf.xml.
> 
> >Well, maybe. I've had to dive into make files many times,
> >and I've helped newbies around some problems with installing
> >software before as well. You can't say you will *never* have to
> >use/edit/write Java to have an ANT install package.
> 
> Really? As i said before, I don't know ant very much. Do you have experience
> with it? Can you give a example when ant needs to be extended with java
> code?
No I don't, just with Tomcat, which didn't install out of the box as they
claimed. Not a big deal, we all expect it with new software. Just noting
that one more addition to the build process will not likely make life
easier. More often it makes things more complicated.
> 
> 
> >> Of course, in this case, jre and and would be provided by bc (i.e.
> >> binarycloud/make/jre/win32|linx|sparc).
> >And BSD and Mac OSX and Darwin and AIX and .......?
> 
> Hmm.
> 
> 
> >Of course I am not as informed as to why PHP is during
> >the install as I've just started looking at the cvs tree
> >I have. But I'm real curious, had to drag out the PHP
> >sources again to build a cgi version, just to install BC.
> 
> Don't understand what you mean. Are you reffering compilation of php
> binaries or something bc specific?
Yes, the need to have a PHP binary in order to install BC. Having always
used PHP as a module I actually forgot how to make a standalone PHP!

I understand now the issue, Windows users have no access to GNU tools for
the build process. In that sense I understand the need now for a build
system not requiring Makefiles.

DAve
-- 
My other computer is your Windows machine... 

Re: FW: [binarycloud-dev] r2 and FreeBSD

[<bin...@li...>]

On Mon, Oct 08, 2001 at 11:28:08AM -0700, bin...@li... wrote:
> can anyone with FreeBSD and a little experience with make on FreeBSD point
> me in the right direction?
> 
> It looks like gmake is reading the makefile wrong?
> 
> I'm tempted to use a pure php make process now... steal some ideas from ant
> and go about our business :)
> 
> _a
> 
I must confess a great ignorance of make other than editing paths, but I
can take a look at it as see what pops.

DAve

-- 
My other computer is your Windows machine... 

[binarycloud-dev] please go here:

[<bin...@li...>]

http://sourceforge.net/tracker/?func=detail&atid=200001&aid=468604&group_id=
1

If you are a memver of sourceforge, and respond to the support guy in the
message. I want to overwhelm him a bit so he sees that turning off features
in software expressly designed to give users the ability to use those
features is a stupid idea.

thanks all,

and I'm registering on tigris right now. egh.

_a

Re: [binarycloud-dev] FW: [ alexandria-Support Requests-468604 ] Reply-to option is gone?

[<bin...@li...>]

I am now considering doing just that.

I wanted to stay with sourceforge because it was convenient to do so (pain
in the ass moving cvs every 6 months :)

but if they're going to pull antics like this I need to look at
alternatives. we have work to do.



> Hi Alex,
> 
> did you consider hosting bc on tigris.org as Manuel proposed some time ago.
> They look very serious and competent.
> 
> Andi
> 
> 
> 
>> -----Original Message-----
>> AGH!!!!
>> 
>> Sorceforge is now in the misguided business of mandating policy
>> where use of
>> its software is concerned...
>> 
>> Well, it may be time again to return to privately hosted cvs and mailing
>> lists... just when I thought it would work well with sourceforge.
>> Fantastic.
>> 
>> https://sourceforge.net/docman/display_doc.php?docid=6693&group_id=1
>> 
>> what a pain in the ass.
>> 
>> _a
>> 
>> 
>> 
>> _______________________________________________
>> binarycloud-dev mailing list
>> bin...@li...
>> https://lists.sourceforge.net/lists/listinfo/binarycloud-dev
>> http://www.binarycloud.com
>> http://www.sf.net/projects/binarycloud/
>> 
> 
> 
> _______________________________________________
> binarycloud-dev mailing list
> bin...@li...
> https://lists.sourceforge.net/lists/listinfo/binarycloud-dev
> http://www.binarycloud.com
> http://www.sf.net/projects/binarycloud/
> 

Re: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

> Of course, complete PHP based make would be really killer. I personally
> preffer that much over all other options.

I'm thinking more and more that this is the best option, actually.

_alex

Re: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

>>> I've just finished the install dance for Java to test Tomcat.
>> Huh.. gratulations. No you've got a damned fast server side language ;->
> Got that already in PHP ;^) If I could ever get Sablotron to work I'm not
> sure I would need/want *.jsp.

Hey dave,

if you want sablot to work fine, (and you're on a linux box)

http://sourceforge.net/project/showfiles.php?group_id=16402&release_id=52725

doenload that tarball and compile it from scratch.

that package is a little thing I made up that will properly compile and
install apache (comes with a bunch of configure scripts for each package.

I have sab .5 working fine with php406/apache 1320

:)

> Of course I am not as informed as to why PHP is during the install as I've
> just started looking at the cvs tree I have. But I'm real curious, had to
> drag out the PHP sources again to build a cgi version, just to install BC.

yeah, we could convert all the makefiles to perl! :)

hehe....  kidding..

PHP is in the install because we're processing files as part of the make -
it's not just a packaging utility. if it was, you wouldn't need php on the
command line and the make system we have would be fine with a few
modifications.

as it stands we're doing things with make that will be _very_ cool when
they're done, which requires php on the command line.

... And I do not intend to use ANTs as the make system. But I will look at
it, certainly.

_alex

RE: [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

Hi Alex,

>> - standardized and established
>> - really portable (unix, linux, windows) without having
>> extra utilities installed (only requirement is jre)
>
>The JRE requirement bugs me a little (yet more pain in the
>ass setup... but I'm still interested)

Hmm. I'd no problems with it. Just untared it into /opt and set up
JAVA_HOME, and included it in path.

>> - very powerful and flexible xml configuration of
>> the build process
>
>ooh.

I attached a example build file. It's from EIRC java chat client applet.
It's a small project and the build.xml gives a brief overview of the very
basic ant feautures.

>> - don't mess around with make & various command line tools
>
>If we build a configure system I don't see that being a problem.

Not really, at least under unix like systems. But much people develop and
set up complete dev system on windows, so I think they don't want to mess
around with cygwin. Especially in the enterprise sector java is established
and it would be convinient for those people to handle this, they want to
have a _fast_ setup process, not messing around downloading additional
tools.


>> - recursive build support (call ant on every package level,
>We'll do that (easily) with gnu make.

Jepp, I just wanted to point out that ANT can do this, too :-)


>> Of courses it is based on java and so the startup, and
>> execution is not as fast as make and friends.

>And that is a big thing for me, at least.
>I'm taking great pains to design a make system
>that will be as screaming fast as possible.

Yep, thats a good point. But it's only the make process. It's a matter of
how much we could benefit from feautures at the cost of speed. I can't say
not if it's worth it.

>And what we're doing is mostly implemented in php anyway..
>so the role of gnu make will really be just to call our
>utilities in a convenient way.

And that's a thorn in my eye. All this tools required only for basically
calling others.

>Some of the XML config stuff sounds very nice.
>I'd like to see it function before I bit though :)

Ok, I'll experiment with it some time, basically replacing the Make we
currently have. And we'll see how it works.

>I have no hatred for java, _but_ we would be mixing languages and
>that means lots of extra software to install.
That's a point.

>Could you send a little quick-setup example to the list
>so we can check it out and see what we think?

ok, see other mail.


andi

Re: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

> I felt that I had to bring this up 'caus I think if we can reduce
> development time on the make system and simultaniously archive more
> flexibility and compatibility - even if this means introducing new
> technology - it's worth a *thought*. I think it's not a good idea to reject
> something that _might_ bring a real advantage to bc without discussing or
> maybe experimenting with it.

And here is the central point:

We should _always_ look at everything we can to see if there are good ideas
to be had. We're all smart people but we can't think of everything: so if
ANTs has some cool gizmo we forgot, we can look at it, steal it, give them
credit, and be the better for it.

However, as andi's _commits_ to cvs will show you: we're all in the business
of building a functional production system as soon as we possibly can.

best,

_alex

RE: [binarycloud-dev] r2 and FreeBSD

[<bin...@li...>]

Hi Alex,

>I'm tempted to use a pure php make process now... steal some 
>ideas from ant and go about our business :)

That would be killer.

andi

RE: [binarycloud-dev] FW: [ alexandria-Support Requests-468604 ] Reply-to option is gone?

[<bin...@li...>]

Hi Alex,

did you consider hosting bc on tigris.org as Manuel proposed some time ago.
They look very serious and competent.

Andi



>-----Original Message-----
>AGH!!!!
>
>Sorceforge is now in the misguided business of mandating policy
>where use of
>its software is concerned...
>
>Well, it may be time again to return to privately hosted cvs and mailing
>lists... just when I thought it would work well with sourceforge.
>Fantastic.
>
>https://sourceforge.net/docman/display_doc.php?docid=6693&group_id=1
>
>what a pain in the ass.
>
>_a
>
>
>
>_______________________________________________
>binarycloud-dev mailing list
>bin...@li...
>https://lists.sourceforge.net/lists/listinfo/binarycloud-dev
>http://www.binarycloud.com
>http://www.sf.net/projects/binarycloud/
>

FW: [binarycloud-dev] r2 and FreeBSD

[<bin...@li...>]

can anyone with FreeBSD and a little experience with make on FreeBSD point
me in the right direction?

It looks like gmake is reading the makefile wrong?

I'm tempted to use a pure php make process now... steal some ideas from ant
and go about our business :)

_a


------ Forwarded Message
> From: Dave Goodrich <da...@pi...>
> Date: Mon, 8 Oct 2001 08:48:55 -0500
> To: Alex Black <en...@tu...>
> Subject: Re: [binarycloud-dev] r2 and FreeBSD
> 
> On Fri, Oct 05, 2001 at 12:30:56PM -0700, Alex Black wrote:
>> can you send the errors?
>> 
>> it's almost certainly a problem with make.
>> 
>> thanks,
>> 
>> _alex
>> 
> Inserted below, thanks,
> 
> DAve
> 
>>> Hello all.
>>> 
>>> I've been trying to get to playing with BC and now I have the time. CVS co
>>> this morning and tried to run make, it fails. Ran gmake and got much
>>> farther, but still many errors. Is ther something I should know?
>>> 
>>> FreeBSD 4.1
>>> PHP 4.0.6
>>> BCHOME is set to /usr/local/binarycloud/binarycloud
>>> (where /usr/local/binarycloud is the top of the CVS tree)
>>> 
> 
> Errors!!
> 
> [dave@redbird:/usr/local/binarycloud/binarycloud]$ gmake
> 
> Building da site
> 
> in base
> 
> in core
> 
> /usr/local/binarycloud/binarycloud
> /base
> /utils
> /installcode.sh
> :No such file or directory
> 
> *** Error code 1
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/base/core.
> 
> *** Error code 1
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/base.
> 
> in ext
> 
> /usr/local/binarycloud/binarycloud
> /base
> /utils
> /installcode.sh
> : not found
> 
> *** Error code 127
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/ext.
> 
> in user
> 
> in user/htdocs
> 
> /usr/local/binarycloud/binarycloud
> /base
> /utils
> /installcode.sh
> : not found
> 
> *** Error code 127
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/user/htdocs.
> 
> *** Error code 1
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/user.
> 
> Building en site
> 
> in base
> 
> in core
> 
> /usr/local/binarycloud/binarycloud
> /base
> /utils
> /installcode.sh
> :No such file or directory
> 
> *** Error code 1
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/base/core.
> 
> *** Error code 1
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/base.
> 
> in ext
> 
> /usr/local/binarycloud/binarycloud
> /base
> /utils
> /installcode.sh
> : not found
> 
> *** Error code 127
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/ext.
> 
> in user
> 
> in user/htdocs
> 
> /usr/local/binarycloud/binarycloud
> /base
> /utils
> /installcode.sh
> : not found
> 
> *** Error code 127
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/user/htdocs.
> 
> *** Error code 1
> 
> 
> 
> Stop in /usr/local/binarycloud/binarycloud/user.
> 
> gmake: *** [langs] Error 1
> 
> 
> My other computer is your Windows machine...
> 

------ End of Forwarded Message

RE: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

Hi Dave,

>> >I'm a new user and maybe out of line here but...
>> You are not! :-)
>Well........ New to BC!

But not out of line (if i've understood the term right in sense of "my
opinion does not matter", i'm not a native speaker).


>Got that already in PHP ;^) If I could ever get Sablotron
>to work I'm not sure I would need/want *.jsp.
That was ironic a ironic remark. But what sablot concerns, what does not
work with your setup ?


>> That's a point. On the other hand you need a working
>> make and all related tools. The advantage here is that
>> we can provide jre and a env setting script with bc.
>> And if everything is packaged and configured well, you don't
>> need to install something optional.

>Yes, but that was already installed and in use for the
>multitude of other tools/apps I have/will install. Agreed you
>could package everything together, but that moots the point of
>moving to ANT so developers can move
>on with other BC things. I don't want them to get tied up
>developing an ANT package installer under the assumption
>it will save time over developing a PHP based installer.

Of course, complete PHP based make would be really killer. I personally
preffer that much over all other options.

>>
>> >Not real sure I would want to learn Java in order to use PHP ;^)
>> No, no. You don't need to know anything about coding in java or something
>> like that, as well as you don't have to know make/autoconf stuff
>to make bc
>> or other projects now. They only think you have to do is edit
>MakeConf.xml.

>Well, maybe. I've had to dive into make files many times,
>and I've helped newbies around some problems with installing
>software before as well. You can't say you will *never* have to
>use/edit/write Java to have an ANT install package.

Really? As i said before, I don't know ant very much. Do you have experience
with it? Can you give a example when ant needs to be extended with java
code?


>> Of course, in this case, jre and and would be provided by bc (i.e.
>> binarycloud/make/jre/win32|linx|sparc).
>And BSD and Mac OSX and Darwin and AIX and .......?

Hmm.


>Of course I am not as informed as to why PHP is during
>the install as I've just started looking at the cvs tree
>I have. But I'm real curious, had to drag out the PHP
>sources again to build a cgi version, just to install BC.

Don't understand what you mean. Are you reffering compilation of php
binaries or something bc specific?

Andi

Re: [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

> Hi All,
> 
> preior I want to state that I'm not trying to force or somehow to replace
> the make-system with ant. This is just a evil thought of mine occured some
> time ago and maybe ant is a option. Maybe.
> 
> Has anybody deeper insight and experience with ANT (a very versatile java
> based tool)? I just had a brief look over it and I personally think ant does
> implement very much of the advanced make we're going to create. Thus we
> would create double work by implementing ant flexibility with make.

hmm... go on..


> I'm not experienced with ant, nor did I try to setup a base build config to
> see how it performs (but I will :-)). Ant is primary designed for Java
> applications, but it can be used for any other builds. From my knowledege of
> ant, I've gained so far, the befits are:
> 
> - standardized and established
> - really portable (unix, linux, windows) without having
> extra utilities installed (only requirement is jre)

The JRE requirement bugs me a little (yet more pain in the ass setup... but
I'm still interested)

> - very powerful and flexible xml configuration of
> the build process

ooh.

> - very extendable with other tasks (so we could easyliy
> implement php2xml, bcc etc in a short time

ooh.

> - covers file permissions, copying files, building distributions
> and snapshots, and much more

ooh.

> - don't mess around with make & various command line tools

If we build a configure system I don't see that being a problem.

> - recursive build support (call ant on every package level,
> and only build files in that package or in that package
> and everything below) with seperation of source and output.

We'll do that (easily) with gnu make.

> - apache license

That's cool.

> The only thing it requires is jre and ant itself and as it runs withing
> native windows (no cygwin required, afaik) we don't need all the make tools
> and don't take care of all windows specific stuff (/ = \ etc). Of courses it
> is based on java and so the startup, and execution is not as fast as make
> and friends.

And that is a big thing for me, at least.
I'm taking great pains to design a make system that will be as screaming
fast as possible.

And what we're doing is mostly implemented in php anyway.. so the role of
gnu make will really be just to call our utilities in a convenient way.

> The big advantage is protability, flexiblity and most stuff we plan is
> already utilized with ant. So I think we should seriously consider a
> possible usage of ant.

Some of the XML config stuff sounds very nice.
I'd like to see it function before I bit though :)

> I know some of you guys don't love Java, so I do. And introducing another
> technology in bc may be not wanted. But not everything about java is bad,
> and ant might be a candidate of "take the good things leave the bad".

I have no hatred for java, _but_ we would be mixing languages and that means
lots of extra software to install.

> I'm going play around a bit with it.

Could you send a little quick-setup example to the list so we can check it
out and see what we think?

If it's cool enough... :)

> As I said I'm not very experieced with ant, so what do you think?

Let's play with ant a bit and see.

_a

[binarycloud-dev] FW: [ alexandria-Support Requests-468604 ] Reply-to option is gone?

[<bin...@li...>]

AGH!!!!

Sorceforge is now in the misguided business of mandating policy where use of
its software is concerned...

Well, it may be time again to return to privately hosted cvs and mailing
lists... just when I thought it would work well with sourceforge. Fantastic.

https://sourceforge.net/docman/display_doc.php?docid=6693&group_id=1

what a pain in the ass.

_a

Re: [binarycloud-dev] r2 and FreeBSD

[<bin...@li...>]

In case anyone was wondering I did send those directly to Alex. If the
grooup wants I can post them.

DAve

On Fri, Oct 05, 2001 at 12:30:56PM -0700, Alex Black wrote:
> can you send the errors?
> 
> it's almost certainly a problem with make.
> 
> thanks,
> 
> _alex
> 
> > Hello all.
> > 
> > I've been trying to get to playing with BC and now I have the time. CVS co
> > this morning and tried to run make, it fails. Ran gmake and got much
> > farther, but still many errors. Is ther something I should know?
> > 
> > FreeBSD 4.1
> > PHP 4.0.6
> > BCHOME is set to /usr/local/binarycloud/binarycloud
> > (where /usr/local/binarycloud is the top of the CVS tree)
> > 
> > Thanks,
> > 
> > DAve
> 
> 
> _______________________________________________
> binarycloud-dev mailing list
> bin...@li...
> https://lists.sourceforge.net/lists/listinfo/binarycloud-dev

-- 
My other computer is your Windows machine... 

Re: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

On Mon, Oct 08, 2001 at 06:53:33PM +0200, bin...@li... wrote:
> Hi Dave,
> 
> >I'm a new user and maybe out of line here but...
>
> You are not! :-)
Well........ New to BC!
> 
> >I've just finished the install dance for Java to test Tomcat.
> Huh.. gratulations. No you've got a damned fast server side language ;->
Got that already in PHP ;^) If I could ever get Sablotron to work I'm not
sure I would need/want *.jsp.
> 
> >Install, upgrade, join mail list to fix problems not covered
> >in docs. Took a few days.
> 
> hehe.. hmm sounds familiar to me :(
> 
> 
> >I for one would not want to require Java to complete an
> >install of a product that did not require Java in it's
> >operation.
> 
> That's a point. On the other hand you need a working make and all related
> tools. The advantage here is that we can provide jre and a env setting
> script with bc. And if everything is packaged and configured well, you don't
> need to install something optional.
>
Yes, but that was already installed and in use for the multitude of other
tools/apps I have/will install. Agreed you could package everything
together, but that moots the point of moving to ANT so developers can move
on with other BC things. I don't want them to get tied up developing an
ANT package installer under the assumption it will save time over
developing a PHP based installer.
> 
> >Not real sure I would want to learn Java in order to use PHP ;^)
> No, no. You don't need to know anything about coding in java or something
> like that, as well as you don't have to know make/autoconf stuff to make bc
> or other projects now. They only think you have to do is edit MakeConf.xml.
Well, maybe. I've had to dive into make files many times, and I've helped
newbies around some problems with installing software before as well. You
can't say you will *never* have to use/edit/write Java to have an ANT
install package.
> 
> >Thats one more thing to install, upgrade, track, on each
> >production box I may use BC on.
> 
> Of course, in this case, jre and and would be provided by bc (i.e.
> binarycloud/make/jre/win32|linx|sparc).
And BSD and Mac OSX and Darwin and AIX and .......?

Of course I am not as informed as to why PHP is during the install as I've
just started looking at the cvs tree I have. But I'm real curious, had to
drag out the PHP sources again to build a cgi version, just to install BC.

DAve
-- 
My other computer is your Windows machine... 

RE: Re (PRODUCTION!): [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

Hi Dave,

>I'm a new user and maybe out of line here but...

You are not! :-)


>I've just finished the install dance for Java to test Tomcat.
Huh.. gratulations. No you've got a damned fast server side language ;->

>Install, upgrade, join mail list to fix problems not covered
>in docs. Took a few days.

hehe.. hmm sounds familiar to me :(


>I for one would not want to require Java to complete an
>install of a product that did not require Java in it's
>operation.

That's a point. On the other hand you need a working make and all related
tools. The advantage here is that we can provide jre and a env setting
script with bc. And if everything is packaged and configured well, you don't
need to install something optional.


>Not real sure I would want to learn Java in order to use PHP ;^)
No, no. You don't need to know anything about coding in java or something
like that, as well as you don't have to know make/autoconf stuff to make bc
or other projects now. They only think you have to do is edit MakeConf.xml.

>Thats one more thing to install, upgrade, track, on each
>production box I may use BC on.

Of course, in this case, jre and and would be provided by bc (i.e.
binarycloud/make/jre/win32|linx|sparc).


andi

Re: [binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

Booooooooooo!

=]



> I know some of you guys don't love Java, so I do. And introducing another
> technology in bc may be not wanted. But not everything about java is bad,
> and ant might be a candidate of "take the good things leave the bad".

[binarycloud-dev] ANT vs. MAKE

[<bin...@li...>]

Hi All,

preior I want to state that I'm not trying to force or somehow to replace
the make-system with ant. This is just a evil thought of mine occured some
time ago and maybe ant is a option. Maybe.

Has anybody deeper insight and experience with ANT (a very versatile java
based tool)? I just had a brief look over it and I personally think ant does
implement very much of the advanced make we're going to create. Thus we
would create double work by implementing ant flexibility with make.

I'm not experienced with ant, nor did I try to setup a base build config to
see how it performs (but I will :-)). Ant is primary designed for Java
applications, but it can be used for any other builds. From my knowledege of
ant, I've gained so far, the befits are:

- standardized and established
- really portable (unix, linux, windows) without having
  extra utilities installed (only requirement is jre)
- very powerful and flexible xml configuration of
  the build process
- very extendable with other tasks (so we could easyliy
  implement php2xml, bcc etc in a short time
- covers file permissions, copying files, building distributions
  and snapshots, and much more
- don't mess around with make & various command line tools
- recursive build support (call ant on every package level,
  and only build files in that package or in that package
  and everything below) with seperation of source and output.
- apache license

The only thing it requires is jre and ant itself and as it runs withing
native windows (no cygwin required, afaik) we don't need all the make tools
and don't take care of all windows specific stuff (/ = \ etc). Of courses it
is based on java and so the startup, and execution is not as fast as make
and friends.

The big advantage is protability, flexiblity and most stuff we plan is
already utilized with ant. So I think we should seriously consider a
possible usage of ant.

I know some of you guys don't love Java, so I do. And introducing another
technology in bc may be not wanted. But not everything about java is bad,
and ant might be a candidate of "take the good things leave the bad".

I'm going play around a bit with it.

As I said I'm not very experieced with ant, so what do you think?

andi

Re: [binarycloud-dev] RE: [PEAR] PHPDoc segfaults..?

[<bin...@li...>]

> hi all,
> 
> did a checkout of php4 cvs, found php4/pear/HTML/ITX.php...
> copied that alone to my include dir, didn't work (same problem)
> then I copied the _entire_ contents of HTML from the checkout, and that did
> work...
> 
> so, problem solved :)

errr..

not :(

I ran the test on a larger set of files, the Segfault problem still
persists...

I'll spend a bit more time trying to see if anything changes that and do a
backtrace later in the week.

best,

_alex

Re: [binarycloud-dev] entity history

[<bin...@li...>]

hey andi,

glad you think so.

I've added that field and its description to the entity example in
docs/specs/xml/

:)

-a


> Hi Alex,
> 
> this is a *very* useful feauture we should implement.
> 
> Andi
> 
> 
> 
> 
>> I was looking at a sourceforge support request (mine, actually, the one
>> about reply-to on this list)
>> 
>> and a relatively simple means of getting 'entity change histories' to work
>> within the system occurred to me:
>> 
>> a top-level tag in the entity definition:
>> 
>>    <entity:history>true</entity:history>
>> 
>> what would that do?
>> 
>>    -When you generate metabase xml schemas, _two_ schemas would
>> be created,
>> one with the name of the entity: "furbees" and another: "furbees_history"
>>    -furbees_history would contain the same fields as furbees, with the
>> addition of two important fields:
>>        -history_id
>>        -history_timestamp
>> 
>> With those two tables, it would be _relatively_ easy to implement a
>> GetByHistory() method in EntityManager, which would trickle down to the
>> correct Query to furbees_history through QueryManager..
>> 
>> that's easy to deal with for applications like bugtrackers, support
>> requests, etc...
>> 
>> so you could ask for the 'current copy' of the entity, or one of the copies
>> by id or timestamp, or all but the current, or all.. etc.
>> 
>> something where you have a set of fields but need to maintain the same key
>> id.
> 
> 
> _______________________________________________
> binarycloud-dev mailing list
> bin...@li...
> https://lists.sourceforge.net/lists/listinfo/binarycloud-dev
> http://www.binarycloud.com
> http://www.sf.net/projects/binarycloud/
> 

RE: [binarycloud-dev] entity history

[<bin...@li...>]

Hi Alex,

this is a *very* useful feauture we should implement.

Andi




>I was looking at a sourceforge support request (mine, actually, the one
>about reply-to on this list)
>
>and a relatively simple means of getting 'entity change histories' to work
>within the system occurred to me:
>
>a top-level tag in the entity definition:
>
>    <entity:history>true</entity:history>
>
>what would that do?
>
>    -When you generate metabase xml schemas, _two_ schemas would
>be created,
>one with the name of the entity: "furbees" and another: "furbees_history"
>    -furbees_history would contain the same fields as furbees, with the
>addition of two important fields:
>        -history_id
>        -history_timestamp
>
>With those two tables, it would be _relatively_ easy to implement a
>GetByHistory() method in EntityManager, which would trickle down to the
>correct Query to furbees_history through QueryManager..
>
>that's easy to deal with for applications like bugtrackers, support
>requests, etc...
>
>so you could ask for the 'current copy' of the entity, or one of the copies
>by id or timestamp, or all but the current, or all.. etc.
>
>something where you have a set of fields but need to maintain the same key
>id.

RE: [binarycloud-dev] File permissions

[<bin...@li...>]

Hi Alex,

that looks fine to me.

Andi


>find ${BC_PATH}/build/site_name/en/ -name '*.php' -type f -exec chmod
>u+rw,g+rw,a+r {} \;
>and
>find ${BC_PATH}/build/site_name/en/htdocs -type f -exec chmod -R
>u+rw,g+rw,a+r {} \;

><fsperms>
>    <file>
>        <match>*.php</match>
>        <perms>
>            <user>rw</user>
>            <group>rw</group>
>            <all>r</all>
>        </perms>
>    </file>
>    <location>
>        <!-- this is the path from lang_code root in the build dir -->
>        <path>htdocs/</path>
>        <resursive>true</recursive>
>        <perms>
>            <user>rw</user>
>            <group>rw</group>
>            <all>r</all>
>        </perms>
>    </location>
></fsperms>

Re: [PEAR] Re: [binarycloud-dev] RE: [PEAR] PHPDoc segfaults..?

[<bin...@li...>]

On Sun, 7 Oct 2001 12:02:25 -0700, alex black wrote:

>what's the expected ETA on a 100% separate CVS for pear?

Right now we can't do that because the infrastructure on pear.php.net
isn't finished yet. Today I did some work for the account requesting
thingie (which is working properly now) and Tomas is working hard on
the package/installer system.

When this task is finished, we can (hopefully) finally move the
rest of the code to /pear.

My next effort will be to set up a proper documentation system (with
Docbook) and helping Tomas in finishing the installer.

Stay tuned,
	- Martin