Lev at mozdev

]

Bugs
None
Pages where repag doesn't work (these can be considered bugs, but the heuristics for determining the right link will never be perfect)
http://www.wincustomize.com (link runs Javascript)
Interoperability (priority depends on hardness, probably it will be one fix for several extensions)

BetterSearch
CustomizeGoogle
BumbleSearch
RIP

Improvements

Indication that loading is finished
Any number
Make links inside the page point to anchors
Make it persistent. E.g., enable going back in history to work properly, bookmarks.
Enable adding one more page (without pointing to the link? By toolbar?)
Delete repeated headers & footers.
Add only those pages where a certain string can be found.
Make a slideshow control (really distant future!)

26 comments|Leave a comment

Repagination 2006.4.4 [Apr. 4th, 2006|11:30 pm]

[

Tags

]

The only difference is a fix of the following bug: repaginating the same search twice returned the same results again instead of the next results.

The query for the link to use included [position() = last()], which was supposed to search for the last appearance of a link with the same text / image. Instead, it searched for one that was the last child of its parent. (?) So what I did was take the query results and iterate till I got to the last one.

Hmmm... Maybe I should delete [position() = last()] from the query, too?

TODO repagination

[Apr. 1st, 2006|10:38 pm]

[

Tags

]

Anonymous:
- Make it persistent. E.g., enable going back in history to work properly, bookmarks.
Bajrang reported:
- Doesn't work with sites like http://www.wincustomize.com. (On that site, the link runs Javascript. Solution: load current page in iframe, activate the link. Reuse the iframe.)
- Interoperation with extensions like "BetterSearch" and/or "CustomizeGoogle".
- "Repag again" shows the same pages again (oops...)
klode:
- Make links inside the page point to anchors
Jon C.:
- Enable adding a page
Me:
- Delete repeated headers & footers.
Also interoperation with BumbleSearch (linuxoid) and RIP (Brian).

TODO

[Mar. 17th, 2006|12:33 am]

[

Tags

]

1. Inform the user what's new on first loading. Copy from ImageSize.
2. Toni complains that if a page doesn't fit on the screen, it should scroll automatically. Ignore?
3. Make a speed control?
4. Delete headers and footers
5. Make the last "next" link load more pages?
6. Inform Csimbi
7. Answer Toni

2 comments|Leave a comment

Version 2006.3.16

[Mar. 17th, 2006|12:05 am]

[

Tags

]

Fixed Csimbi's bugs (http://levmozdev.livejournal.com/1410.html).

For the first bug, did as planned: copied each child of body separately.

For the second, used the following algorithm:

  if link text contains a single number
    for the first page loaded, try to increment the number and search for that link (e.g., search for a "3" link)
      if not found, search for a link with the original text
    for each successive page, behave consistently (i.e., either always increment or never)

Reason: There might be a self link with the page number.

Also fixed bug found in code: Originally: If the next page to be loaded is the same as the previous, stop. Fixed: If the next page to be loaded is the same as the current one, stop.

Test pages

[Mar. 13th, 2006|11:31 pm]

[

Tags

]

http://www.google.com/search?q=testing&start=0&ie=utf-8&oe=utf-8&client=firefox-a&rls=org.mozilla:en-US:official - simple case
http://community.webshots.com/photo/526149579/1526151793083129783aCWBPC - slideshow
http://thepiratebay.org/brwsearch.php?b=1&c=401 - csimbi's bug where font size is changed
http://bt.eastgame.net/ - to repaginate by numbers
http://sevastnikon.livejournal.com/ - where the link contains a number but is not incremented
http://www.pvponline.com/ - where the link is an image

TODO find test cases where
- The number is only part of the link, e.g. [1], [2], etc.
- The link contains only an input (special treatment in code)
- In the last page the next link points to itself.

3 comments|Leave a comment

TODO

[Mar. 13th, 2006|11:30 pm]

[

Tags

]

Jeffery suggests to make the throbber work when pages are loaded. Actually, I've been missing an indication myself. I'd rather it was the progress bar, though. Only I've no clue how to do any of the above.

More bugs

[Mar. 13th, 2006|11:22 pm]

[

Tags

]

Brian writes that after repaginating at http://housd.net/ and at http://www.pvponline.com/, the "next" link downloads the next page instead of loading it. Worksforme.
UPD:
After the release of version 2006.3.16, he writes the bugs are fixed.

Repagination bugs

[Mar. 11th, 2006|10:51 pm]

[

Tags

]

Csimbi has reported the following bugs:

* At http://thepiratebay.org/brwsearch.php?b=1&c=401 the appended pages have smaller fonts. I've no clue why this happens. Must be something with CSS. I'll try to append each element of each appended page separately instead of appending the whole <body> as is done now. I hope this works.

* Doesn't work with http://bt.eastgame.net/. It turns out that on that page, the ">>" link does not link to the next page, but to the last one. So there is no "next" link, and technically it's not a bug. But there is an easy solution to this problem: If the link's text is a number, increment the number each time. So I'm gonna do this. Then one will be able to repaginate that site using the "2" link.

The question arises: What about links like "[2]"? I can check if the link contains a number. But then, what about LiveJournal's "Next 20" links? There are several solutions. What I'll do is this: check for links which are numbers only, then wait to see if anyone complains about sites with [2]-like links.

Upd: The first bug fixed as planned, the second differently. Fixes are here.

Repagination technicalities

1 comment|Leave a comment

[Mar. 11th, 2006|10:26 pm]

[

Tags

]

So how did I enable repagination to work with scripts disabled?
( Read more... )

Repagination

[Mar. 11th, 2006|10:10 pm]

[

Tags

]

repagination is an improvement on antipagination. It enables to do two things:

1. Right-click on any link, choose "Re-pagination", "All". Repagination will append the page linked to at the end of the current page. Then at that page it will attempt to find the "same" link, and append the linked page again, etc. So this is useful for blogs, search results etc. with "next" or "previous" links. It is also possible to limit the number of pages appended, which I usually do myself.

2. Right-click on any link, choose "Re-pagination", "Slideshow", then the number of seconds. Repagination will load the page linked to (replacing the current page, as usual), wait for the specified time, then attempt to find the "same" link in that page and load the page linked to, etc. This is useful for photos. There is no way to limit the number of pages in the slideshow in advance. Of course, the slideshow will stop if repagination doesn't find the link (i.e., at the last photo). It can also be stopped by right-clicking anywhere on the page and choosing "Re-Pagination", "Stop".

Changes in the first version (2006.2.18) relative to the original antipagination: Made it work even with scripts disabled.