Difference between revisions of "Talk:DPL Example 027"
m (→Order=descending) |
|||
| (8 intermediate revisions by 2 users not shown) | |||
| Line 15: | Line 15: | ||
:::On other queries using %TOTALPAGES% in the header, there is no header displayed. See [http://www.mixesdb.com/db/index.php/MixesDB:Tests/Test_3 here]. --[[User:Subfader|Subfader]] 09:16, 6 July 2009 (UTC) | :::On other queries using %TOTALPAGES% in the header, there is no header displayed. See [http://www.mixesdb.com/db/index.php/MixesDB:Tests/Test_3 here]. --[[User:Subfader|Subfader]] 09:16, 6 July 2009 (UTC) | ||
| + | |||
| + | |||
| + | ::::Does debug= 3 show SELECT SQL_CALC_FOUND_ROWS DISTINCT as the first part of the SQL statement? [[User:Gero|Gero]] 14:37, 6 July 2009 (UTC) | ||
| + | |||
| + | :::::Yep. I enabled debug now. --[[User:Subfader|Subfader]] 16:05, 6 July 2009 (UTC) | ||
| + | |||
| + | ==Navi end & back== | ||
| + | Something's not quite working yet in the template. When you navigate to the end it reads <code>'''back''' next (23 listed)</code>, when you go back one step it's <code>'''back''' '''next''' (50 listed / 98 left)</code> while it should be 22 left again. --[[User:Subfader|Subfader]] 16:54, 6 July 2009 (UTC) | ||
| + | |||
| + | === Ideas === | ||
| + | We could always start at multiples of the count limit (1, 51, 101, etc). | ||
| + | This would lead to more stability. | ||
| + | |||
| + | This will only work if the last entry from the previous screen is NOT repeated in the next screen. | ||
| + | One could offer an option to decide whether the list should start at a value GREATER oder GREATER_EQUAL than the | ||
| + | value which is passed. DPL could also silently read ahead one entry and use that one for scroll navigation. I have a slight tendency towards the first alternative. | ||
| + | |||
| + | [[User:Gero|Gero]] 21:25, 6 July 2009 (UTC) | ||
| + | |||
| + | :Whatever is more logic and lighter. I'd go for the second. Prepare the next start page in the back. I'd see no use in listing the last of the previous set on top again. For the average user this looks like a mistake. --[[User:Subfader|Subfader]] 20:53, 7 July 2009 (UTC) | ||
| + | |||
| + | == Order=descending == | ||
| + | |||
| + | As soon as you use order=descending the system breaks. [[User:Subfader/Olli1/O2/O3|Example]] --[[User:Subfader|Subfader]] 16:24, 13 July 2009 (UTC) | ||
Latest revision as of 17:24, 13 July 2009
Contents
Offset + 1 on next page
The next page lists the last of the previous on top. This looks wrong. No need to reply if this is part of the super-light strategy, but it makes count = 1 impossible. ;) --Subfader 20:57, 5 July 2009 (UTC)
urlencode
From at least MW 1.14 it is {{urlencode: not {{#urlencode: --Subfader 21:08, 5 July 2009 (UTC)
TOTALPAGES = 1
%TOTALPAGES% always returns 1 in the header or footer for me. I guess you fixed it already sinec it's working here and I use the latest 1.8.6. --Subfader 06:15, 6 July 2009 (UTC)
- The dpldemo website runs exactly the version that is in the DOWNLOAD area.
- %TOTALPAGES% is only calculated if needed (use debug=3 to see the SQL statement) as it may cost additional CPU load. DPL checks if %TOTALPAGES% is used within header or footer. Maybe there is aproblem with that code...
- Where exactly do you get the "1" for %TOTALPAGES% ?
- I get it in the footer and header. E.g. (25 listed / 1) [using "if total > pages" meanwhile], but laso when I use %TOTALPAGES% before the links template. Therefor the links don't work as desired. --Subfader 07:01, 6 July 2009 (UTC)
- Does debug= 3 show SELECT SQL_CALC_FOUND_ROWS DISTINCT as the first part of the SQL statement? Gero 14:37, 6 July 2009 (UTC)
- Yep. I enabled debug now. --Subfader 16:05, 6 July 2009 (UTC)
Something's not quite working yet in the template. When you navigate to the end it reads back next (23 listed), when you go back one step it's back next (50 listed / 98 left) while it should be 22 left again. --Subfader 16:54, 6 July 2009 (UTC)
Ideas
We could always start at multiples of the count limit (1, 51, 101, etc). This would lead to more stability.
This will only work if the last entry from the previous screen is NOT repeated in the next screen. One could offer an option to decide whether the list should start at a value GREATER oder GREATER_EQUAL than the value which is passed. DPL could also silently read ahead one entry and use that one for scroll navigation. I have a slight tendency towards the first alternative.
Gero 21:25, 6 July 2009 (UTC)
- Whatever is more logic and lighter. I'd go for the second. Prepare the next start page in the back. I'd see no use in listing the last of the previous set on top again. For the average user this looks like a mistake. --Subfader 20:53, 7 July 2009 (UTC)
Order=descending
As soon as you use order=descending the system breaks. Example --Subfader 16:24, 13 July 2009 (UTC)
