Well yes and no, I am a firm believer in not recreating the wheel. I will still have to modify glade just a little but we as Open Source developers need to use what is at our disposal. It is to our distinct advantage to use glade as the interface builder in my opinion.

My most major goal right now is beating kylix out the door with a 1.0 release. They cannot use glade or I bet they would be doing it.

I am looking at the glade code right now to see what I have to do with it if anything to make this possible.

Believe me if I had the time to write it in Lazarus I sure would but as it stands there are only but about 4 or 5 acitve developers so I have to do what I have to do.

Take a look at the KDevelop product they actually integrated about 6 different projects into the IDE. I do not really care about what it is written in as long as it just plain works.

Anyhow I am going to take a quick stab at it and see how well my theory works. Mike I will send you a xml file from glad in a minute just so that you can visualize how it would work.

Cliff

+ From: Michael A. Hess

+ Right now we are setting here looking at the effort to built the
+ designer and not really knowing where to start or attack it.
+ With GLADE the interface is already available and available on
+ several platforms.
+ The other advantage is that even if we develop our own designer which
+ saves info in an XML file we need a way to go from the XML to either
+ Pascal code turning it into a resource that is linked into the
+ application.
+
+ Since GLADE already saves it's data into an XML file we can
+ work on this portion of the tool. We can work on converting
+ the XML data into something that can be used in an LCL bases
+ application. Once we have that worked out with the GLADE output
+ we can then start working on our own designer that saves XML
+ output. The work spent on GLADE will be of value and give us
+ some RAD capabilities sooner.

I don't know if I'm to happy with this approach. Is Glade GTK based or can you define your own widget set. If you can define your own widget set I cant see how they will be shown. So I guess you never are able to design something native LCL with it. Unless you take a large amount of time to convert things I think the look and feel of your designed form wont be the same. Further more how do you want to handle the on the fly creation of control and event entries in your source. If this is not supported, I see something created as braindead as M$VC.

Ok it is a nice tool to generate an xml template for your form. If we want to use it for a quick design of a form, as a separate tool, OK, but I think we shouldn't take effort into integrating it into Lazarus. IMO if we do that, we kill a great advantage Lazarus/Delphi has. Better put effort in creating forms form xml.

Marc

Hey why are we even trying to build our own
designer how about we use GLADE instead. We may need to hack it a little to make a clean integration but it should be absolutley possible to move things along very quickly.
�
Cliff

I started a code upload before leaving the house for work.

The address is http://www.codepunk.com/files/glade.zip it should be doen by now.

I do not wish to commit it to the main CVS until we decide to actually use it.

just unzip in the component directory it may or may not have a binary in it. I did a make
clean before uploading it.

To build drop into the glade directory and run ./configure --disable-gnome then just run
make to build the binary. The binary will be in the glade directory after the build.

Cliff

Cliff and I discussed this a little and I agree with him. While I eventually would like to see a design tool built into Lazarus we are short on two things. Developers and time. Developers in that only a few people are actually contributing code or effort to the project. Time, in that we want to beat Kylix to the punch.

Right now we are setting here looking at the effort to built the designer and not really knowing where to start or attack it. With GLADE the interface is already available and available on several platforms. The other advantage is that even if we develop our own designer which saves info in an XML file we need a way to go from the XML to either Pascal code turning it into a resource that is linked into the application.

Since GLADE already saves it's data into an XML file we can work on this portion of the tool. We can work on converting the XML data into something that can be used in an LCL bases application. Once we have that worked out with the GLADE output we can then start working on our own designer that saves XML output. The work spent on GLADE will be of value and give us some RAD capabilities sooner.

Since GLADE has it's own team of developers constantly improving their tool we have an additional ready made team that is in a sense adding to the Lazarus effort.

In any event I think it is an avenue that is worth exploring.

Mike I will send you a xml file from glad in a minute just so that you can visualize how it would work.

Okay do-kay. :-)

This was my thinking. I DO NOT want it to be the final end all solution to form design for Lazarus. However it could be a nice stop gap method and a good learning experience. From Cliffs effort it appears as though it is something that could be workable in a week or so. It is something that should be investigated and see what we can use from it.

I am going to make a small mod to glade to force a fixed widget on every window so the use of glade should be pretty transparent.

Cliff

I haven't looked at glade before this so I don't know enough to give my opinion (but I'll give it anyway:-) ) I would like to nothing else, perhaps the method of coding will help ME determine how to do the object inspector or designer. Maybe we can implement GLADE until we add our own. If it doesn't take a ton of time, perhaps it's worth it just for the learning value. Shane

>> 04/26/00 06:54AM >>>

+ From: Michael A. Hess

+ Right now we are setting here looking at the effort to built the
+ designer and not really knowing where to start or attack it.
+ With GLADE the interface is already available and available on
+ several platforms.
+ The other advantage is that even if we develop our own designer which
+ saves info in an XML file we need a way to go from the XML to either
+ Pascal code turning it into a resource that is linked into the
+ application.
+
+ Since GLADE already saves it's data into an XML file we can
+ work on this portion of the tool. We can work on converting
+ the XML data into something that can be used in an LCL bases
+ application. Once we have that worked out with the GLADE output
+ we can then start working on our own designer that saves XML
+ output. The work spent on GLADE will be of value and give us
+ some RAD capabilities sooner.

I don't know if I'm to happy with this approach. Is Glade GTK based or can you define your own widget set. If you can define your own widget set I cant see how they will be shown. So I guess you never are able to design something native LCL with it. Unless you take a large amount of time to convert things I think the look and feel of your designed form wont be the same. Further more how do you want to handle the on the fly creation of control and event entries in your source. If this is not supported, I see something created as braindead as M$VC.

Ok it is a nice tool to generate an xml template for your form. If we want to use it for a quick design of a form, as a separate tool, OK, but I think we shouldn't take effort into integrating it into Lazarus. IMO if we do that, we kill a great advantage Lazarus/Delphi has. Better put effort in creating forms form xml.

Marc

_________________________________________________________________ To unsubscribe: mail with
"unsubscribe" as the Subject
archives at http://www.miraclec.com/list_archives/lazarus