semanticscuttle-devel — General discussion about SemanticScuttle development

Re: [Semanticscuttle-devel] [ semanticscuttle-Bugs-3139493 ] add in / xml tags vs .php

[Christian W. <cw...@cw...>]

Hello Tomamplius,


> The <?xml ..... > is not understant by my version of php or
> apache in a php script (PHP 5.2.6-1+lenny9 with
> Suhosin-Patch 0.9.6.2 ). This version is the last php and
> apache version in debian 5.0. Is same with 0.97.2

You have "short open tags" activated in your php.ini file, which is
something you really should switch off. Change the setting in php.ini
and restart your web server; after that it will work.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

[Semanticscuttle-devel] SemanticScuttle 0.97.2 is out

[Christian W. <cw...@cw...>]

Hello fellow bookmarkers,


SemanticScuttle version 0.97.2 has just been released. This release is
a bugfix release that repairs some problems in SemanticScuttle 0.97.1:

- Fix bug #3178597: Broken link to context in gsearch admin index page
- Fix bug #3074816: French translation breaks edit javascript
   This also fixes #3094047 and #3178592
- Fix bug #3111254: Search in my_watchlist results in error
- Fix bug #3073215: Updating bookmark time does not work
- Fix bug #3065284: AjaxVote problem with Webkit browsers
- Make Unittests run with phpunit 3.5.x


-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] google custom searches limited to main page of sites?

[Christian W. <cw...@cw...>]

Hi Chris,


> I was wondering if the Google Custom Search (gsearch/) is intended to
> limit searches to the main page of the sites listed? I have
> SemanticScuttle installed at: http://clinki.es/gsearch/ and this
> appears to be the behavior... for example, search for 'David' and this
> should return many pages, but returns only one:
> http://www.google.com/cse?cref=http://clinki.es/gsearch/context.php&q=david&sa=Search&siteurl=clinki.es/gsearch/
> 
> If this is intended behavior-- and I could understand that being the
> case-- is there either a) an option to change this or b) a way to use
> SemanticScuttle to drive a GCSE in this way?

What do you mean by "main page"? All pages exported by
> http://clinki.es/api/export_gcs.php
should be found by google. This script is linked from
http://clinki.es/gsearch/


-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-= Geeking around in the name of science since 1982 =-

[Semanticscuttle-devel] google custom searches limited to main page of sites?

[Chris L. <ch...@ch...>]

I was wondering if the Google Custom Search (gsearch/) is intended to
limit searches to the main page of the sites listed? I have
SemanticScuttle installed at: http://clinki.es/gsearch/ and this
appears to be the behavior... for example, search for 'David' and this
should return many pages, but returns only one:
http://www.google.com/cse?cref=http://clinki.es/gsearch/context.php&q=david&sa=Search&siteurl=clinki.es/gsearch/

If this is intended behavior-- and I could understand that being the
case-- is there either a) an option to change this or b) a way to use
SemanticScuttle to drive a GCSE in this way?

Thanks!

c
--
Chris Lott

Re: [Semanticscuttle-devel] fixing the opensearch problem #3097187

[Christian W. <cw...@cw...>]

Hi Golanor,


> The whole purpose of using "rawurlencode" is that it could be used as
> an url.
> If I decode it then it will still show a '+' instead of "%20", as it
> should show.
> Are you sure the problem is not fixed?
Yes, otherwise I would not have said so :)

Please you describe in your own words which problem your patch fixes
and how to reproduce the problem.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-= Geeking around in the name of science since 1982 =-

Re: [Semanticscuttle-devel] fixing the opensearch problem #3097187

[<go...@gm...>]

Hi Chrisitian,
The whole purpose of using "rawurlencode" is that it could be used as an
url.
If I decode it then it will still show a '+' instead of "%20", as it should
show.
Are you sure the problem is not fixed?

On Thu, Feb 10, 2011 at 10:54 AM, Christian Weiske <cw...@cw...>wrote:

> Hi Golanor,
>
>
> > > The fix is now commited (under search-fix).
> You might want to set you author name and email address in git
> because currently it is:
> > author        Or <or@or-laptop.(none)>
>
> $ git config user.name Golanor
> $ git config user.email go...@gm...
>
>
> --
> Regards/Mit freundlichen Grüßen
> Christian Weiske
>
> -= Geeking around in the name of science since 1982 =-
>
>
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb
> _______________________________________________
> Semanticscuttle-devel mailing list
> Sem...@li...
> https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel
>

Re: [Semanticscuttle-devel] fixing the opensearch problem #3097187

[Christian W. <cw...@cw...>]

Hi Golanor,


> > The fix is now commited (under search-fix).
You might want to set you author name and email address in git
because currently it is:
> author	Or <or@or-laptop.(none)>

$ git config user.name Golanor
$ git config user.email go...@gm...


-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-= Geeking around in the name of science since 1982 =-

Re: [Semanticscuttle-devel] fixing the opensearch problem #3097187

[Christian W. <cw...@cw...>]

Hi Golanor,


> That was the error.
Good that it works now.

> The fix is now commited (under search-fix).
> Tell me what do you think.
Did you read that mail to you:
https://sourceforge.net/mailarchive/message.php?msg_id=27001945
?

Your patch does not solve the problem.

Please you describe in your own words which problem your patch fixes
and how to reproduce the problem.



-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-= Geeking around in the name of science since 1982 =-

Re: [Semanticscuttle-devel] Semanticscuttle-devel Digest, Vol 6, Issue 1

[<go...@gm...>]

Hello,
That was the error.
The fix is now commited (under search-fix).
Tell me what do you think.

On Thu, Feb 10, 2011 at 10:33 AM, Christian Weiske <cw...@cw...>wrote:

> Hi Golanor,
>
>
>
> > Ok, now when I try to push, this is the error:
> > $ git push origin search-fix
> > fatal: The remote end hung up unexpectedly
>
> Seems to be a similar problem as
> https://sourceforge.net/apps/trac/sourceforge/ticket/12546
>
> Run
> $ git config --list
> and post the output. The URL should begin with ssh:// and not git://.
> If it does, change it to
> ssh://
> go...@se.../gitroot/semanticscuttle/sc
>
> --
> Regards/Mit freundlichen Grüßen
> Christian Weiske
>
> -= Geeking around in the name of science since 1982 =-
>
>
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb
> _______________________________________________
> Semanticscuttle-devel mailing list
> Sem...@li...
> https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel
>

Re: [Semanticscuttle-devel] Semanticscuttle-devel Digest, Vol 6, Issue 1

[Christian W. <cw...@cw...>]

Hi Golanor,



> Ok, now when I try to push, this is the error:
> $ git push origin search-fix
> fatal: The remote end hung up unexpectedly

Seems to be a similar problem as
https://sourceforge.net/apps/trac/sourceforge/ticket/12546

Run
$ git config --list
and post the output. The URL should begin with ssh:// and not git://.
If it does, change it to
ssh://go...@se.../gitroot/semanticscuttle/sc

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-= Geeking around in the name of science since 1982 =-

Re: [Semanticscuttle-devel] Semanticscuttle-devel Digest, Vol 6, Issue 1

[<go...@gm...>]

Ok, now when I try to push, this is the error:
$ git push origin search-fix
fatal: The remote end hung up unexpectedly


On Thu, Feb 10, 2011 at 9:38 AM, Christian Weiske <cw...@cw...>wrote:

> Hi Golarnor,
>
>
>
> > The git clone command doesnt work
> > $ git clone ssh://
> > go...@se.../gitroot/semanticscuttle/sc
> > when i type this i get this error:
> > fatal: destination path 'sc' already exists and is not an empty
> > directory. what am I doing wrong?
>
> "sc" already exists. It seems you already have a directory named "sc"
> on your hard disk.
>
> Btw, please post to the list - maybe other people can respond faster
> than I do.
>
> --
> Regards/Mit freundlichen Grüßen
> Christian Weiske
>
> -=≡ Geeking around in the name of science since 1982 ≡=-
>
>
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb
> _______________________________________________
> Semanticscuttle-devel mailing list
> Sem...@li...
> https://lists.sourceforge.net/lists/listinfo/semanticscuttle-devel
>
>

Re: [Semanticscuttle-devel] Semanticscuttle-devel Digest, Vol 6, Issue 1

[Christian W. <cw...@cw...>]

Hi Golarnor,



> The git clone command doesnt work
> $ git clone ssh://
> go...@se.../gitroot/semanticscuttle/sc
> when i type this i get this error:
> fatal: destination path 'sc' already exists and is not an empty
> directory. what am I doing wrong?

"sc" already exists. It seems you already have a directory named "sc"
on your hard disk.

Btw, please post to the list - maybe other people can respond faster
than I do.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] Feature request #3164348: Make default privacy configurable

[Christian W. <cw...@cw...>]

Hi Brett,


> I will also need to learn more about the unit testing tool.

We use PHPUnit from www.phpunit.de
Running the tests is pretty easy:
$ cd tests
$ phpunit .


-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] Feature request #3164348: Make default privacy configurable

[Brett D. <bs...@fr...>]

Thanks for the clarifications. I will do a more thorough recursive
grep.

I will also need to learn more about the unit testing tool.

Rgrds,
Brett

On Wed, 9 Feb 2011 07:20:33 +0100, Christian Weiske
<cw...@cw...> wrote:
> Hi again,
> 
> 
>> The idea is correct, but you missed some points where bookmarks are
>> added/updated:
>> - www/bookmarks.php (what you probably meant)
>> - www/edit.php
>> - www/api/posts_add.php
> I also missed some points, i.e. the import scripts. Just grep/ack-grep
> the files.

Re: [Semanticscuttle-devel] Feature request #3164348: Make default privacy configurable

[Christian W. <cw...@cw...>]

Hi again,


> The idea is correct, but you missed some points where bookmarks are
> added/updated:
> - www/bookmarks.php (what you probably meant)
> - www/edit.php
> - www/api/posts_add.php
I also missed some points, i.e. the import scripts. Just grep/ack-grep
the files.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] Feature request #3164348: Make default privacy configurable

[Christian W. <cw...@cw...>]

Hi Brett,


> This is in regards to feature request #3164348
> [http://sourceforge.net/tracker/?func=detail&aid=3164348&group_id=211356&atid=1017433],
> "Make default privacy configurable."
> 
> Privacy can be made configurable when the bookmark is added by the
> modifications described in the feature request, but the assignment to
> 'bStatus' is also made 16 lines before in the preceding if block. In
> the event that POST_TAGS is not set, an array is created with a
> default bStatus value of 0. I think this should also be set to the
> global default:
The idea is correct, but you missed some points where bookmarks are
added/updated:
- www/bookmarks.php (what you probably meant)
- www/edit.php
- www/api/posts_add.php

All of them need to be modified, and you will need to add unit tests to
verify it works.


> 'tags' => (POST_TAGS ? explode(',', stripslashes(POST_TAGS)) :
> array()),
>                 'bStatus' => $GLOBALS['privacy_default'], //New
> assignment.
> 		//'bStatus' => 0, //Old assignment.
>            );
We need to structure the settings a bit, so I'd call the variable
> $GLOBALS['defaults']['privacy']
You should add it to the config.defaults.php, too.

I'd say go on and implement it in a new feature branch, i.e.
configurable-privacy.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

[Semanticscuttle-devel] Feature request #3164348: Make default privacy configurable

[Brett D. <bs...@fr...>]

Hi all,

This is in regards to feature request #3164348
[http://sourceforge.net/tracker/?func=detail&aid=3164348&group_id=211356&atid=1017433],
"Make default privacy configurable."

Privacy can be made configurable when the bookmark is added by the
modifications described in the feature request, but the assignment to
'bStatus' is also made 16 lines before in the preceding if block. In the
event that POST_TAGS is not set, an array is created with a default
bStatus value of 0. I think this should also be set to the global
default:

'tags' => (POST_TAGS ? explode(',', stripslashes(POST_TAGS)) :
array()),
                'bStatus' => $GLOBALS['privacy_default'], //New
assignment.
		//'bStatus' => 0, //Old assignment.
           );

Let me know if you have other suggestions/thoughts.

Thanks and Regards,
Brett


On Fri, 28 Jan 2011 06:24:20 +0100, Christian Weiske
<cw...@cw...> wrote:
> Hi Brett,
> 
> 
>> I am new to open-source work and am currently familiarizing myself
>> with git and the product code. I think this will be primarily a
>> learning experience for me for a few weeks at least.
> 
> That's totally fine. There is no need to rush, and we have plenty of
> small tasks you can do to get into the code. If you need help, just
> send a mail to the mailing list or join me/us in #semanticscuttle on
> freenode. If you have jabber/xmpp, you might add cw...@cw... to
> your contact list.
> 
> Btw, the mailing list is on 
> https://sourceforge.net/mailarchive/forum.php?forum_name=semanticscuttle-devel
> 
> You should at first get familiar with SemanticScuttle itself - download
> it, install it and get it to work on your own machine.
> 
> Since you are very new, you might implement a small feature request:
>> "Make default privacy configurable"
> http://sourceforge.net/tracker/?func=detail&aid=3164348&group_id=211356&atid=1017433
> 
> The steps to take are:
> 1. Find out what the things are needed to do to implement the feature
> 2. Look at the code and see what needs to be done where to do it
> 3. Write that up and send it to the mailing list. We can discuss it then
> 4. Create your own feature branch (i.e. "configurable-privacy") from
>    master and implement it in there
> 5. Push the branch to sourceforge so I can review it.
> 6. If everything is fine, I will merge it into the master branch
> 
> If you find out that that's not your thing, you could help at
> other places, i.e. the documentation.

Re: [Semanticscuttle-devel] Updates to privatekey branch (aka Private Rss Feed)

[Christian W. <cw...@cw...>]

Hello Mark,



> Ok, pushed some additional updates.  I'm authenticating using the
> user name and the private key.  Sounds quite reasonable.  I didn't
> want to monkey with the login() function so I copied it and  simply
> altered it to authenticate the user.  I haven't completed the unit
> testing for this yet, but will do so this week as time permits.  If
> you look at www/rss.php, you'll notice I log the user out.
> 
> At your leisure, please let me know if this is closer to what you were
> wanting.

Yes, the login code looks better now.

I'm still pondering what to do with the javascript regeneration.
Downside of your JS code is that it works only with JavaScript while
most of SC works fine without JS. Also, you do not have any error
handling on the javascript side yet if the key regeneration fails. This
is what I hate/fear most when doing js things; it's most time not easier
to do than the "normal" way if you do it correctly.

And yes, I meant that the private RSS feed on the main page is missing.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-= Geeking around in the name of science since 1982 =-

Re: [Semanticscuttle-devel] SemanticScuttle and Scuttle

[Christian W. <cw...@cw...>]

Hello Chris,


> I was just wondering: it looks like development of SemanticScuttle is 
> ongoing-- is it tracking with development of Scuttle in addition to
> doing its own things? I'm looking to implement social bookmarking for
> my design team, but don't want to get stranded... it looks like
> Scuttle has had a few releases in the last months, but it's hard to
> tell if they were significant.

No, I do not follow Scuttle's development. Since Scuttle was not in
development for years, the code bases differ significantly now and even
if we'd like, it's not easy to merge changes back into SemanticScuttle.

We @SemanticScuttle are several developers now, so I think we are the
best bet now :)

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] Updates to privatekey branch (aka Private Rss Feed)

[Christian W. <cw...@cw...>]

Hi Mark,


> Ok, I think I'm done.  Unit tested (added an additional test on the
> private key generation and exists).  Updated as per your comments
> below.  I think it's pretty good.  I'm all for constructive
> criticism, so feel free to let me know what you think.
I went through all commits and noted what I think could be better. Some
of the "bad" changes have been reverted in later commits (i.e.
htaccess), but I still note them here.

ce9124bfd576394b75c5e7a8f3a97dff7a89e747 mostly done 
- scripts/create-testbookmarks.php 
- javascript for key re-generation? why?
- "onClick" -> onclick
- utf8_strtolower?
- "function updateUser($uId, $password, $name, $privateKey, $email,
  $homepage, $uContent) {"
  breaks BC - backwards compatibility. If someone used the
  semanticscuttle API from his script, it would behave .. strangely now
  or stop working. I know that SC is not stable yet so BC breaks are
  allowed, but nevertheless it's not nice.
- "function PrivateKeyExists" - methods begin with a lowercase letter.
  Classes with an uppercase one
- same method: $output is a bad variable name since it's a return
  value. $exists would be better
- src/SemanticScuttle/db/mysqli.php: echo $sql_report?
- Changes to www/.htaccess? Why?
- why change getBookmarks() to add privatekey?
  why not "log in" the user temporarily?

97717684c429ee4e4827f4e4a120a019cbd3f78f added new file
- commitdescription is .. ambiguous
- function getNewPrivateKey() should be in the user model, it belongs
  there. maybe someone else wants to re-generate the key not with ajax

a32c9a1578c218728e0c40219d993b0a709033e7 added secondary rss feed to
bookmark page to test
- scripts/create-testbookmarks.php changed again
- www/bookmarks.php: "(private)" not translatable. Use sprintf and a
  configurable string from the translation

43ad8e77250cbc8e67f7f22127314fbe7c76b23d Cleaned up User.php and moved
Private Key function to it
- src/SemanticScuttle/Service/User.php: constructors do not have a
  @return value since it's known what they return
- "@param integer $nb Max number of usrs"
- "@return array Array of users" - array of user objects? array of user
  data arrays from database? I have to look at the code to see what it
  really returns
- "function & getUsers($nb = 0)" -> add visibility modificator
  (public/protected/private) -- why "&"?

f6873268e58a4cd3a1e32d1aa03ce7f5795752bc Updated tests to reflect
private key changes
- "public function testTestUniquePrivateId()":
  "$this->assertTrue($this->us->privateKeyExists($randKey));" is the
  last test, but you should verify if the user you created is the one
  that got the key.


- main page misses private key support
- you would need to test getBookmarks() with private key to make sure
  that other user's private bookmarks do not get returned when I pass
  my private key
- wrong privatekey should return an auth error http status - currently
  the user does not know if it worked or not
- my main problem is that you changed getBookmarks() to add the
  privatekey parameter to it. This is unfortunate since it adds a new
  "hole" in the security of the application, deep inside the API. The
  current security measures (user authentication, fetching bookmarks
  for the authenticated user) are well tested and know to work. Now
  that you add a private key parameter, we need a new suite of tests
  covering all the cases like
  "privatekey, own private bookmarks, foreign private bookmarks"
    -> user needs only to get his own private bookmarks, not the
    private bookmarks of other users.
  I personally would prefer that when accessing the feed with a private
  key, the user is logged in but not stored in the session. With the
  current user being available, getBookmarks() would automatically work
  correctly. Not storing it in the session would make sure the user
  cannot continue to do other things with the private key.
- You often committed changes to the privatekey feature and cleanup work
  in one single commit. while cleaning up is great and needs to be
  done, please commit them in separate commits. This makes review and
  merging easier. Ideally, cleanup would be made in a separate branch,
  but even I am too lazy to do that :) Separate commits however are a
  must.

Generally, great first work on SemanticScuttle. Don't let my comments
scare you off, I really appreciate your work.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] regarding "help wanted" post / a first task

[Christian W. <cw...@cw...>]

Hi Golanor,


> Looks to me like it was simply a case of not encoding the $terms
> string before adding it to the url, so i simply added:
>     $terms = rawurlencode($terms);
> 
> after the "else" of the search terms (line 88, www/search.php).

Do you mean "urldecode"?

php > echo rawurlencode('foo+bar');
foo%2Bbar
php > echo urldecode('foo+bar');
foo bar

And yes, I'd fix that on the same place.

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] Semanticscuttle-devel Digest, Vol 6, Issue 1

[<go...@gm...>]

the file is search.php

Re: [Semanticscuttle-devel] [semanticscuttle - Help] RE: CleanURLS problem on Godaddy linux grid

[Christian W. <cw...@cw...>]

Hi Mark,


> I chose to e-mail you directly since this is more of an internal
> discussion rather than public.
I think it's related to semanticscuttle development, so the dev list is
best suited here. Maybe other people read it, too, and we get insights
about the problem from them. Even if not, discussing it on the mailing
list archives our ongoing discussion and may help people in the future.

> I too have a GoDaddy account and attempted to install SemanticScuttle
> and using the base install, the application simply doesn't work.
> What I've found is that the $_SERVER['PATH_INFO'] variable is
> different than what the app is expecting.  I've attempted to alter
> the .htaccess to correct it, but I've had no luck.  However, I've
> been able to resolve the issue by parsing the PATH_INFO differently.
> My initial fix seems to be borderline hack and was looking for an
> alternate method to parse the PATH_INFO URL to get the appropriate
> information.  I think resolving this will fix most of the "hosted"
> issues that people seem to be experiencing.
I understand that it's a PATH_INFO issue. The problem I currently have
is that I cannot reproduce it @home. This is the reason I ask for
apache and php versions, and preferably even the settings the
virtualhost in apache has. Maybe it's not even an apache but lighttpd
or so? Information helps.

We might fix it based on debugging information from some var_dump
statements on a live semanticscuttle installation on godaddy, but
that's not the same as a reproducible testcase on the dev machine - I
can i.e. run unit tests against that.

About parsing PATH_INFO: That's historic crap, and I don't like it,
too. Better would be to have a central URL parsing method that takes
care of such things.


-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] regarding "help wanted" post / a first task

[Christian W. <cw...@cw...>]

Hi Golanor,


> > https://sourceforge.net/tracker/?func=detail&atid=1017430&aid=3097187&group_id=211356
> Looks to me like it was simply a case of not encoding the $terms
> string before adding it to the url, so i simply added:
>     $terms = rawurlencode($terms);
> after the "else" of the search terms (line 88).
Which file do you mean?

> i saw in the comments that someone suggested doing
> $terms = str_replace('+', ' ', $terms);
> but that looks to me like a hack to the specific problem, not a
> solution.
That's indeed a hack.


> about git - I have no idea how to use it, so can you please explain
> to me how to create a new branch and commit this change to it?
$ git clone
ssh://go...@se.../gitroot/semanticscuttle/sc
$ git branch my-fix
$ git checkout my-fix
.. edit files
$ git add files you changed
$ git commit -m 'I fixed this and that'
$ git push origin my-fix

-- 
Regards/Mit freundlichen Grüßen
Christian Weiske

-=≡ Geeking around in the name of science since 1982 ≡=-

Re: [Semanticscuttle-devel] regarding "help wanted" post / a first task

[<go...@gm...>]

First of all - Hello everyone!

Looks to me like it was simply a case of not encoding the $terms string
before adding it to the url, so i simply added:
    $terms = rawurlencode($terms);

after the "else" of the search terms (line 88).

i saw in the comments that someone suggested doing
$terms = str_replace('+', ' ', $terms);

but that looks to me like a hack to the specific problem, not a solution.

anyway, I checked it on my installation, and it seems to be working fine.

about git - I have no idea how to use it, so can you please explain to me
how to create a new branch and commit this change to it?


On Tue, Feb 1, 2011 at 8:52 AM, Christian Weiske <cw...@cw...> wrote:

> Hi Golanor,
>
>
>
> > ok, I subscribed to the mailing list and got the code from git.
>
> The first - easy - task for you to solve could be bug #3097187 [1]
>
> The steps to take are:
> 1. Find out what the things are needed to do to implement the feature
> 2. Look at the code and see what needs to be done where to do it (this
>   might be a bit overkill for that bug, but it is important when
>   implementing new features)
> 3. Write that up and send it to the mailing list. We can discuss it
>   then
> 4. Create your own feature branch (i.e. "fix-twoword-opensearch")
>   from master and implement it in there
> 5. Push the branch to sourceforge so I can review it.
> 6. If everything is fine, I will merge it into the master branch
>
> Please consider the following things:
> 1. SemanticScuttle tries to follow the PEAR Coding Standards. Use the
>   PHP_CodeSniffer tool from PEAR to make sure you don't leave more
>   style errors than there were when you began working on the file. Not
>   all files comply to the styles yet, but we're on the right track
> 2. We have unit tests. Run them. Make sure that they still work when
>   you finished your task. Better: Write a test for your own bug first
>   - it should fail. Then fix the bug, and your test should work now.
> 3. If you find other bugs, open a bug report.
> 4. Do not fix several things in one branch. This really makes it hard
>   to merge changes into the master branch. Creating a new branch in
>   git is cheap, so use that.
> 5. Update the Changelog
> 6. I repeat myself, but: Only commit in your own branches.
>
>
> [1]
>
> https://sourceforge.net/tracker/?func=detail&atid=1017430&aid=3097187&group_id=211356
>
> --
> Regards/Mit freundlichen Grüßen
> Christian Weiske
>
> -=≡ Geeking around in the name of science since 1982 ≡=-
>