22 Oct
2003
22 Oct
'03
10:41 a.m.
Hi lukas
- Here I cannot be connected all the time, so it would be good to have
one way to have the css locally by for example having a method
generating it.
- as I cannot be connected all the time I downloaded the style.css file
and put just near the image,
then I edited the stylesheet and put @import
"file:///Volumes/Data/Users/ducasse/Workspace/FirstCircle/Coding/
style.css";
and it worked :) cooooool
- I spent more than half and hour to upgrade to get the new version and
at the end SmallWiki could not work because
Smacc was not updated (idiot store). I think that we should pay
attention to have the right number of the right components.
I do not know what is the solution but we should try to find one.
- for the templates it would be nice to have a title above the pane so
that we know which list represent the possible choices and which one is
the current one.
- we should remove the code from the first page and put it in a page
called how to access your children programmatically
because we can scare a lot of people that way.
- removed the validator by login and modify the templates and it works.
- I started small and wanted to add a button inspect
Smalltalk.SmallWiki defineClass: #Inspect
superclass: #{SmallWiki.Action}
indexedType: #none
private: false
instanceVariableNames: ''
classInstanceVariableNames: ''
imports: ''
category: 'SmallWiki Moka'
renderContent
html submitButtonWithAction: #inspect text: 'inspect'
class>>title
^'Inspect'
I can see it in the action list, I can remove predefined actions from
the list but when I add 'inspect' nothing changes.
When I typed http://localhost:8080/?action=Inspect I can see the button
rendered but there is no inspector that
show up in VW
two questions:
1- What should I do to get the actions showing up in the actions list
once I modify the template?
2- When I redefined renderContent as
renderContent
self inspect
The inspector pops up in VW but not with html submitButtonWithAction:
#inspect text: 'inspect'
I looked at the permission but you mentioned that the default for
permission is that everybody can see the action.
I think that we should start a FAQ :)
Stef
22 Oct
22 Oct
11:18 a.m.
Hi Stef,
- Here I cannot be connected all the time, so it would
be good to have
one way to have the css locally by for example having a method
generating it.
- as I cannot be connected all the time I downloaded the style.css
file and put just near the image, then I edited the stylesheet and put
@import
"file:///Volumes/Data/Users/ducasse/Workspace/FirstCircle/Coding/
style.css"; and it worked :) cooooool
You could even put the style.css and the image as resource into your
wiki and serve it from the wiki itself, so you don't need to give a
local file-url.
- I spent more than half and hour to upgrade to get
the new version
and at the end SmallWiki could not work because Smacc was not updated
(idiot store).
That are the timeframes I get at home too, when loading the whole code
into a fresh image. Even-tough Monticello is considered to be slow, I
would rather consider StORE to be terribly slow using remote
connections ... ;)
I think that we should pay attention to have the right
number of the
right components. I do not know what is the solution but we should try
to find one.
I updated the SmaCC on the SCG-StORE with the one of Cincom StORE, so
it should be the latest. What if you try to load SmallWiki from Cincom
StORE, I think it should work with the latest version from there.
- for the templates it would be nice to have a title
above the pane so
that we know which list represent the possible choices and which one
is the current one.
Ok, this is an easy thing to change ...
- we should remove the code from the first page and
put it in a page
called how to access your children programmatically because we can
scare a lot of people that way.
I am not sure: that code is put into every folder to simulate the look
of the old folders that were no subclasses of Page. Like this you have
an dynamically created list of children by default on all the folders.
But probably we should extract that code into an external method, so it
would only look like this for the end-user:
[ self renderChildrenOn: html ]
- removed the validator by login and modify the
templates and it works.
Pages containing code can only be saved (reparsed) when being logged in
as administrator. We should probably generate more specific
error-messages, to help the user in that case.
I can see it in the action list, I can remove
predefined actions from
the list but when I add 'inspect' nothing changes.
An action has to register to the structures it is able to handle, so to
make it show up on pages and folders, but not on resources, you have to
add the following code to the class-side of your action:
InfoAction class>>initilize
Page registerAction: self.
Folder registerAction: self.
When I typed http://localhost:8080/?action=Inspect I
can see the
button rendered but there is no inspector that show up in VW
HTML requires a form-tag around any form-element (buttons, combo-boxes,
list-boxes, text-input, text-areas, etc) inside a page. Your code
should look somehow like this and your browser will understand:
Inspect>>renderContent
html form: [
html submitButtonWithAction: #inspect text: 'inspect' ]
1- What should I do to get the actions showing up in
the actions list
once I modify the template?
See above :)
2- When I redefined renderContent as
renderContent
self inspect
The inspector pops up in VW but not with html submitButtonWithAction:
#inspect text: 'inspect'
See above :)
Probably it is easier to start with a link, as links do no require
additional HTML-tags around:
Inspect>>renderContent
html anchorWithAction: #inspect text: 'inspect'
Hope that helps!
I have to go to university now,
Lukas
--
Lukas Renggli
http://renggli.freezope.org
26 Oct
26 Oct
9:09 a.m.
Hi lukas
You could even put the style.css and the image as
resource into your
wiki and serve it from the wiki itself, so you don't need to give a
local file-url.
Then what should I write in the import statement of the template?
Lukas I saved the image and reopen it and it seems to me that the
import I specified was lost.
Is it possible or my mistake?
- I do not have any reference or doc at hand but I found the following
code difficult
because you mix accept and visit in the same method
Visitor>>acceptPage: aPage
self acceptStructure: aPage.
self visit: aPage document
I was programming using the visitor that roel defined for bibtext
processing and he uses
BibVisitor>>visitFile:
BibVisitor>>visitField:
Then the structure provide the acceptMethod:
BibVisitor>> visitEntry: anEntry
anEntry body acceptVisitor: self
I checked and in SmallWiki you have accept on the Structure and accept*
on all the
visitor and this is confusing. Do am I wrong? Visitor been a big double
dispatch I would
prefer to have:
Visitor visiting and Structure accepting (I also checked in the GOF
book)
So may be it would be good to deprecate some methods and offer new
ones. This would improve the
readibility and avoid to get lost....
- I saw that the parent of a resource is not the page having a
reference to it but the folder in which the resource is located. I can
understand why: multiple pages in the same folder can refer to the same
resource. Now is there a simple way to know from which page a resource
is accessed?
I saw that in the searchVisitor you use the currentPage to keep this
information and I did the same but I would like to know if there is a
way to know that information directly.
- how can I find the name and not the title of an uploaded image?
- My first dummy version of a wiki analyzer works :) (look at the
SmallWiki - Moka package). You can get all the pages that contain
uploads. I took me some hours and now I'm sure that I could redo it in
less than 20 min. Now the real questions is how to provide and
categorize results. I will continue to extend it. My idea is that after
this functionality should be built via the search engine.
I would like to be able to plug different results renderers for example
I will try to have thumbnails.
Stef
7:23 p.m.
Hi Stef,
Basically the url to your resource, that means if you upload the
style-sheet to your root folder as a resource and give it the title
'Style Sheet', it is accessible with:
@import "http://localhost:8080/StyleSheet";
For a production server you have to replace the host-name with your
real domain-name, else people from outside won't see your design.
You could even
put the style.css and the image as resource into your
wiki and serve it from the wiki itself, so you don't need to give a
local file-url.
Then what should I write in the import statement of the template? Lukas I saved the image and reopen it and it seems to
me that the
import I specified was lost. Is it possible or my mistake?
This should not happen, as the template and the style-sheet are saved
as a property within the structure where you defined it.
One thing that could cause that problem is if you use this line to
start your server again:
server := SmallWiki.SwazooServer startOn: 8080.
As I write in the comment is only intended to be executed the first
time and should be removed afterwards, because it initializes a
completely new wiki. Just use
server start.
to restart the wiki again, after you opened an image. Of course,
shutting-down and starting-up the server could be made automatically,
but it is not easy in VW as it is in Squeak, I think.
- I do not have any reference or doc at hand but I
found the following
code difficult because you mix accept and visit in the same method
Visitor>>acceptPage: aPage
self acceptStructure: aPage.
self visit: aPage document
I was programming using the visitor that roel defined for bibtext
processing and he uses
BibVisitor>>visitFile:
BibVisitor>>visitField:
Then the structure provide the acceptMethod:
BibVisitor>> visitEntry: anEntry
anEntry body acceptVisitor: self
I checked and in SmallWiki you have accept on the Structure and
accept* on all the visitor and this is confusing. Do am I wrong?
Visitor been a big double dispatch I would prefer to have:
Visitor visiting and Structure accepting (I also checked in the GOF
book)
So may be it would be good to deprecate some methods and offer new
ones. This would improve the readibility and avoid to get lost....
You are basically suggesting the naming-conventions of the design
patterns book. When I implemented those visitors I didn't had that book
yet, so I copied the conventions of the refactoring-browser. I don't
know what is better and what the idea of John Brant was when writing
his visitors.
Of course it is easy to change the message-names, but I don't know if
that really helps much. I can life with both conventions, in 'ROE' the
design-pattern and in 'Formulator' a modified refactoring-browser
approach is taken.
- I saw that the parent of a resource is not the page
having a
reference to it but the folder in which the resource is located. I can
understand why: multiple pages in the same folder can refer to the
same resource. Now is there a simple way to know from which page a
resource is accessed?
Pages do not have children and they might even reference resources from
different folders than their parent.
If you want to know all the structures referencing a specific other
structure (this doesn't have to be a resource necessarily) just use
something similar to the code below. One could even add a message to
Structure itself to get all the references more easily:
Structure>>references
"Return all the structures referencing the receiver."
^(VisitorReferences collect: self root to: self)
collection
- how can I find the name and not the title of an
uploaded image?
The name? You might call #id to get the id of the image or call #url to
get the full path of it. But the filename where it is was uploaded from
is not stored anywhere, I think this is not really needed and
platform-dependant anyway.
Cheers (& see you tomorrow in the Java lecture :P),
Lukas
--
Lukas Renggli
http://renggli.freezope.org
27 Oct
27 Oct
10:33 a.m.
Basically the url to your resource, that means if you
upload the
style-sheet to your root folder as a resource and give it the title
'Style Sheet', it is accessible with:
@import "http://localhost:8080/StyleSheet";
For a production server you have to replace the host-name with your
real domain-name, else people from outside won't see your design.
We should start a FAQ on SmallWiki :)
Lukas I saved the image and reopen it and it
seems to me that the
import I specified was lost. Is it possible or my mistake?
This should not happen, as the template and the style-sheet are saved
as a property within the structure where you defined it.
One thing that could cause that problem is if you use this line to
start your server again:
server := SmallWiki.SwazooServer startOn: 8080.
As I write in the comment is only intended to be executed the first
time and should be removed afterwards, because it initializes a
completely new wiki. Just use
server start.
to restart the wiki again, after you opened an image. Of course,
shutting-down and starting-up the server could be made automatically,
but it is not easy in VW as it is in Squeak, I think.
I do not know but I can think that this is more robust in VW but we
should simply ask.
You are basically suggesting the naming-conventions of
the design
patterns book. When I implemented those visitors I didn't had that
book yet, so I copied the conventions of the refactoring-browser. I
don't know what is better and what the idea of John Brant was when
writing his visitors.
Of course it is easy to change the message-names, but I don't know if
that really helps much. I can life with both conventions, in 'ROE' the
design-pattern and in 'Formulator' a modified refactoring-browser
approach is taken.
Sure I can live with that too, but imagine all the newbie that will
follow my lecture and the lecture of roel, so this is only a cosmetic
change but it is really worth because it helps the people to know where
they are. I realized that because I was coding at least 5 visitors for
the bibtex stuff of roel and I was confused not to find the same.
- I saw that
the parent of a resource is not the page having a
reference to it but the folder in which the resource is located. I
can understand why: multiple pages in the same folder can refer to
the same resource. Now is there a simple way to know from which page
a resource is accessed?
Pages do not have children and they might even reference resources
from different folders than their parent.
If you want to know all the structures referencing a specific other
structure (this doesn't have to be a resource necessarily) just use
something similar to the code below. One could even add a message to
Structure itself to get all the references more easily:
Structure>>references
"Return all the structures referencing the receiver."
^(VisitorReferences collect: self root to: self)
collection
We should add that.
> - how can I find the name and not the title of an uploaded image?
The name? You might call #id to get the id of the image or call #url
to get the full path of it. But the filename where it is was uploaded
from is not stored anywhere, I think this is not really needed and
platform-dependant anyway.
It is important because imagine I want to upload a collection of
picture and then after some reason
I want to transform them (thumbnail or what ever) and I want to save
them. I do not want to have image1, or just the title. Here we are
losing an important semantical information. Note that I did not talked
about the path but the file name like stairInteractif.tif vs stair.tif
(those are just images of my books and their names are really
important).
Cheers (& see you tomorrow in the Java lecture :P),
I cannot attend it, but I will follow all the other ones.
Lukas
--
Lukas Renggli
http://renggli.freezope.org
1:58 p.m.
I was just going to say:
that information should be in the workspace. I want my students to be
able to load SmallWiki, read a bit on the workspace, start serving
their thing and start playing. So everything related to that should be
there. Most of it is, but some of these new things are not. Other stuff
that is good to know should be on a FAQ page somewhere, yes.
On Monday, Oct 27, 2003, at 10:33 Europe/Zurich, ducasse wrote:
OK
Ok
Roel Wuyts
DeComp
roel.wuyts(a)ulb.ac.be Université Libre de
Bruxelles
http://homepages.ulb.ac.be/~rowuyts/
Belgique
Board Member of the European Smalltalk User Group: www.esug.org
Basically the
url to your resource, that means if you upload the
style-sheet to your root folder as a resource and give it the title
'Style Sheet', it is accessible with:
@import "http://localhost:8080/StyleSheet";
For a production server you have to replace the host-name with your
real domain-name, else people from outside won't see your design.
We should start a FAQ on SmallWiki :)
Lukas I saved the image and reopen it and it
seems to me that the
import I specified was lost. Is it possible or my mistake?
This should not happen, as the template and the style-sheet are saved
as a property within the structure where you defined it.
One thing that could cause that problem is if you use this line to
start your server again:
server := SmallWiki.SwazooServer startOn: 8080.
As I write in the comment is only intended to be executed the first
time and should be removed afterwards, because it initializes a
completely new wiki. Just use
server start.
to restart the wiki again, after you opened an image. Of course,
shutting-down and starting-up the server could be made automatically,
but it is not easy in VW as it is in Squeak, I think.
I do not know but I can think that this is more robust in VW but we
should simply ask.
You are basically suggesting the
naming-conventions of the design
patterns book. When I implemented those visitors I didn't had that
book yet, so I copied the conventions of the refactoring-browser. I
don't know what is better and what the idea of John Brant was when
writing his visitors.
Of course it is easy to change the message-names, but I don't know if
that really helps much. I can life with both conventions, in 'ROE'
the design-pattern and in 'Formulator' a modified refactoring-browser
approach is taken.
Sure I can live with that too, but imagine all the newbie that will
follow my lecture and the lecture of roel, so this is only a cosmetic
change but it is really worth because it helps the people to know
where they are. I realized that because I was coding at least 5
visitors for the bibtex stuff of roel and I was confused not to find
the same.
- I saw
that the parent of a resource is not the page having a
reference to it but the folder in which the resource is located. I
can understand why: multiple pages in the same folder can refer to
the same resource. Now is there a simple way to know from which page
a resource is accessed?
Pages do not have children and they might even reference resources
from different folders than their parent.
If you want to know all the structures referencing a specific other
structure (this doesn't have to be a resource necessarily) just use
something similar to the code below. One could even add a message to
Structure itself to get all the references more easily:
Structure>>references
"Return all the structures referencing the receiver."
^(VisitorReferences collect: self root to: self)
collection
We should add that.
> - how can I find the name and not the title of an uploaded image?
The name? You might call #id to get the id of the image or call #url
to get the full path of it. But the filename where it is was uploaded
from is not stored anywhere, I think this is not really needed and
platform-dependant anyway.
It is important because imagine I want to upload a collection of
picture and then after some reason
I want to transform them (thumbnail or what ever) and I want to save
them. I do not want to have image1, or just the title. Here we are
losing an important semantical information. Note that I did not talked
about the path but the file name like stairInteractif.tif vs stair.tif
(those are just images of my books and their names are really
important).
Cheers (& see you tomorrow in the Java lecture :P),
I cannot attend it, but I will follow all the other ones.
Lukas
--
Lukas Renggli
http://renggli.freezope.org
3:50 p.m.
Of course it
is easy to change the message-names, but I don't know if
that really helps much. I can life with both conventions, in 'ROE'
the design-pattern and in 'Formulator' a modified refactoring-browser
approach is taken.
Sure I can live with that too, but imagine all the newbie that will
follow my lecture and the lecture of roel, so this is only a cosmetic
change but it is really worth because it helps the people to know
where they are. I realized that because I was coding at least 5
visitors for the bibtex stuff of roel and I was confused not to find
the same.
[...]
If you want to know all the structures
referencing a specific other
structure (this doesn't have to be a resource necessarily) just use
something similar to the code below. One could even add a message to
Structure itself to get all the references more easily:
Structure>>references
"Return all the structures referencing the receiver."
^(VisitorReferences collect: self root to: self)
collection
We should add that.
9:04 p.m.
Sure lectures have top priority!
no problem, just create a public to do list, and when we have a meeting
we just do the dummy ones. I think that we should deprecate the old
methods and not remove them completely.
Stef
On Lundi, oct 27, 2003, at 15:50 Europe/Zurich, Lukas Renggli wrote:
The same here, I will change the code accordingly an publish it to SCG
and Cincom StORE. Unfortunately my time is quite limited at the moment
(there is a whole bunch of lectures starting at university right now),
so it will take some more time.
Cheers,
Lukas
--
Lukas Renggli
http://renggli.freezope.org
Of course it is easy to change the message-names, but
I don't know
if that really helps much. I can life with both conventions, in
'ROE' the design-pattern and in 'Formulator' a modified
refactoring-browser approach is taken.
Sure I can live with that too, but imagine all the newbie that will
follow my lecture and the lecture of roel, so this is only a cosmetic
change but it is really worth because it helps the people to know
where they are. I realized that because I was coding at least 5
visitors for the bibtex stuff of roel and I was confused not to find
the same.
[...]
If you want to know all the structures
referencing a specific other
structure (this doesn't have to be a resource necessarily) just use
something similar to the code below. One could even add a message to
Structure itself to get all the references more easily:
Structure>>references
"Return all the structures referencing the receiver."
^(VisitorReferences collect: self root to: self)
collection
We should add that.
7756
days inactive
7761
days old
7 comments
3 participants
participants (3)
-
ducasse -
Lukas Renggli -
Roel Wuyts