Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[rbistolfi]

Actually, something like this was requested a while back...

Sorry Joe, I didnt notice your request before...

Err... no, it wasn't me. I simply remembered the long discussion about it...

Oh, must be before I joined the forum. but do you like the idea? I think is a useful addition. And if people dont use it, is just a little box in a side... so no harm.

But the database is another story. It needs a little more work. I think can be useful for package managing, repo maintenance, etc. You could access very quick to package info, such as package location, packager info, build scripts, dates of updates and versions, ./configure used, deps, etc. If this is useful, the search tool can use this more rich info instead the one in packages.txt and take advantages from the sql language. If this is not needed, well I could just make a simple search tool based on the pkg.txt file.

[rbistolfi]

Well, not too much input on this... 
I will start to work on the search tool for the website and we can see if it is a useful tool. I think I can have a minimal version to test in a few days.

[Xeon]

I don't have the time to code it, I got 5 other php projects pending that have to be distributed.

Just one hint, don't make a PHP cron to fetch PACKAGES.TXT into a database.

[rbistolfi]

I don't have the time to code it, I got 5 other php projects pending that have to be distributed.

Its ok, I will do it. I think it t is simple enough for me, and I can optimize it with time. Good luck with your projects.

Just one hint, don't make a PHP cron to fetch PACKAGES.TXT into a database.

Oh, no, I just proposed, if a database is useful, to make a script to ran once time, and then work with the database from there. I think we dont have to put much stress to that file, I understand it is important for slapt-get. Anyway, seems like a database is not needed at all, so no problem there.

[blurymind]

definatelly the package builder,taken with vasm,vl-hot and many other things put vector aside of the other distros. These innovations and little bonus additions make a distro interesting to the mainstream.I noticed that distros today always tend to try to put as much as possible things to make themselfs unique.In many cases features that are not as useful as this here.When officially out,its gonna cause a buzz around vec,im sure of that...But the website still lacks any kind of information about the new and unique features of vector linux.For that reason,i am thinking of making a little presentation page,where to highlight these features. When vl6 is officially out,i think of writing (with nice screenshots) about vl-hot,vasm,vpackager,vl-installer,and everything else that only vec has at a small "features overview" page,like the one at linuxmint:

we should have a "whats new in latest version" page too:
http://linuxmint.com/celena.html

the presentation page can even have youtube presentation links,voiced by the author of the application...thats if he/she is willing to make one,i am gonna include a link to it. It can be presentation and a tut,and can be made with wink,instead of being captured and voiced (like the ones they have at puppy linux)... definatelly content to highlight features.

the website is crucial for the distro's popularity.Our current website doesnt look good enough and lacks some things.

I am sorry for the delay and i am still working on its template+a presentation of features page (maybe link to it from "discover the difference..." )

With Vec6,we are gonna hit it mainstream.

[rbistolfi]

Hi guys!
I was able to setup a search tool for testing purposes here.

It is a very simple tool. It has no fancy features like logical operators or things like that, but I think it works and it can find what you are looking for. I will describe it: the script takes the PACKAGES.TXT file and split the content in several parts, one for each package section. Then it can compare your keyword with each package section, and return the results. It can search in a specific portion of the text too, as in pkg description, pkg name, pkg required, and ./configure if available (vpackager adds this to the description, I think is useful). The tool is made for the user to find packages, of course, but I think it can be useful for packagers and advanced users, since you can find the command build for some package, the list of packages depending on other package, etc...

This is not a full setup, since it will search just in vl5.8std testing repo, I will try to add the others repos in the next days, that is easy...

Well, I am open for any comment, even if it is somethink like "you should find another way to spend your time" or something like that 

PS: dont worry, it is not using the real packages.txt file.

[newt]

Hi guys!
I was able to setup a search tool for testing purposes here.

I think you've done a FANTASTIC job at creating a search tool for VL's package lists; it even has visual appeal!!  This idea is truely an added bonus and would be useful for any user of VL.  I like the toggle options you've included and can see how having more would be useful.  For instance, toggle options to search the various repos (i.e. packages, patches, extras, testing, unstable) with the "safe" ones selected by default.  Another suggestion would be to add somekind of visual "break" between the packages results (maybe as easy as a horizontal rule or perhaps a more visually appealing image rule).  Also, I noticed that when no results are found the text that's returned is in spanish and the slogan says 'discover de difference'.  Last, is VectorLinux one or two words?  All this time I've thought it was one word but now I see distrowatch has us as 'vector' and I see two words on your search page.

Any way, you've done a great job in such little time.
Cheers!!

[exeterdad]

Great work my friend 

I like how you made it link to the package and the folder it resides in. Very slick.
I like the "almost" invisible Vector Linux text in the background of the Vector Linux graphic.

[blurymind]

this search tool is awesome- both graphically appealing and efficient

would be great if it had a drop-down list of all the repos avaiable for search

[rbistolfi]

Thank you very much guys!

I like the toggle options you've included and can see how having more would be useful.  For instance, toggle options to search the various repos (i.e. packages, patches, extras, testing, unstable) with the "safe" ones selected by default.

Yes, I was thinking about just perform a search on all the repos availables, and present the results with a title like "X matches found in testing" or something like that. I will find the way to add some stuff not filling the hole page with checkboxes and selects.

Another suggestion would be to add somekind of visual "break" between the packages results (maybe as easy as a horizontal rule or perhaps a more visually appealing image rule).

Some css will do it, in a very light color perhaps.

Also, I noticed that when no results are found the text that's returned is in spanish

It is a tool made by a spanish speaker for spanish speakers, no just kidding  , I forgot to do the trans 

and the slogan says 'discover de difference'.

 

  After 2 AM I just speak in spanglish 

Last, is VectorLinux one or two words?  All this time I've thought it was one word but now I see distrowatch has us as 'vector' and I see two words on your search page.

I have no idea, I found myself writing it sometimes as one and sometime as two words.

this search tool is awesome- both graphically appealing and efficient

would be great if it had a drop-down list of all the repos avaiable for search

Thanks Blury, said by guy who can draw! I will keep this one  . I didnt want to present it  as a white paper and added some color. I will keep working in the layout. Anyway, the final look should be the same as the website one.

Great work my friend 

I like how you made it link to the package and the folder it resides in. Very slick.
I like the "almost" invisible Vector Linux text in the background of the Vector Linux graphic.

Thanks Exeter! The link was a hard job for me, I worked several days to "cut" the right words from the text to build the link, but damn, it feels good when finally works!

[rbistolfi]

Something to fix: if you check more than one box and the query gives results in all the checked categories, it will repeat the results (try, for example, "gettext" with a few boxes checked.

[newt]

Good catch (rodrigo, correct?)!!

One idea for the duplicate listing issue would be to construct a multidimensional array of the results and then apply a modified array_unique function that can handle multidimensional arrays to scrub out the dupes.  But there are probably many other ways to deal with the same issue - just an idea.

Also, re-applying the search text/options when results are displayed would be great.  Just another idea.

Thanks again - it really does look/work great!

[rbistolfi]

Good catch (rodrigo, correct?)

Thats correct

One idea for the duplicate listing issue would be to construct a multidimensional array of the results and then apply a modified array_unique function that can handle multidimensional arrays to scrub out the dupes.  But there are probably many other ways to deal with the same issue - just an idea.

Thanks I will look into it, something like $results[pkg_name][n...] could work. I think array_unique() can do it, since the entrys are exactly the same.

Also, re-applying the search text/options when results are displayed would be great.  Just another idea.

That is in the todo list already, should be easy since there is some open source functions around to do it, we dont need to reinvent the wheel here (is not open source great?

Thank you very much for the input.

[newt]

Also, re-applying the search text/options when results are displayed would be great.  Just another idea.

That is in the todo list already, should be easy since there is some open source functions around to do it, we dont need to reinvent the wheel here (is not open source great?

Thank you very much for the input.

The way I've re-applied post/get data back to a form was just to add an if statement to the value attribute, or checked attribute.
For instance:

Code: [Select]

Assuming you don't extract($_GET)....

for text fields:
<input type="text" name="keyword" value="<?php if ( isset($_GET['keyword']) ) { echo $_GET['keyword']; } ?>" />

for checkbox fields:
<input type="checkbox" name="data[]" value="PKG_NAME" <?php if ( isset($_GET['data']) && in_array('PKG_NAME', $_GET['data']) ) { echo "checked"; } ?> />

Hope this helps!

[rbistolfi]

It does thank you. I was thinking in something like that. I see you can do it with javascript too, but I prefer the server side if possible. The client can always have javascript disabled.