RCFV.1.1.2304.4757 (Updated 2022-08-03)

  • Version RCFV.1.1.2304.4757
  • Download 2383
  • File Size 6.02 MB
  • Create Date 2022-08-06

Updated 2022-08-03

===============================================================================
Change log for version 1.1
===============================================================================
( 1) Added multipage support for urls with numeric suffixes which precede an extension, e.g., http://www.rickfischer.org/arc001.html
( 2) Added more predefined fractions translated from UNICODE
( 3) Fixed unrecognized left tag bracket in certain urls
( 4) Improved the recognition of images in image tags when the attribute is not obvious, e.g., data-original="https://images.idgesg.net/images/..."
( 5) Added support for partially "encoded" urls in url pages, e.g., https:u002Fu002Fwww.stitcher.com. This will allow the batch download of Stitcher mp3 files.
( 6) Ensured that changes to a default folder in the Settings will be reflected immediately
( 7) Fixed problem that prevented standard link images from being wrapped horizontally in a browser window
( 8) Fixed Esc key behavior such that it just stops scrolling and does not force a spurious refresh
( 9) Fixed a problem accessing https://www.theskepticsguide.org/podcasts
(10) If the File List is visible but the main window is not then clicking on any line in the list (to go to that file number) will make the main window visible
(11) Some sites, such as www.newsweek.com, hide relevant information and require the "Display hidden text" option
(12) Hopefully improved recognition of web pages which contain non-empty tables. This will result in fewer false warning messages about word-wrap being used with table data.
(13) Fixed next page recognition when it is "hidden" in the ALT attribute text of an image
(14) Allows next page recognition when the assumed next page is significantly different from the initial url but the latter contains a redirect
(15) Repaired several formatting errors when converting web text back to HTML via 'Copy and edit'
(16) Changed redundant {Alt}+{p} key to {Alt}+{f} for PDF creation in the browser
(17) Using {y} key to toggle filtering caused focus to be lost in viewer
(18) Fixed potential overflow when evaluating images when converting to HTML
(19) The "Display hidden text" option has been extended to include text included within <FORM...> tags
(20) Improved recognition and download of new releases when right-clicking the Help button and selecting option 2
(21) Fixed incorrect handling of some unreadable urls such as https://twitter.com/i/redirect?url=https%3A%2F%2Ftwitter.com....
(22) Added support for image urls without an explicit image extension but which contain the literal string "IMAGE", e.g., https://www.wsws.org/asset/523f984c-7ad9-403a-bc71-f80e0a82358f?rendition=image1280

plus other miscellaneous, minor problem fixes

 


The following notification is for existing users of RCFV versions prior to version 1.1. New users need not be concerned.

Due to an unfortunate choice of some "default" regular expression meta characters used in the provided fvf files we have discovered that they are too effective at being selected as the "best" matching filter file.

Below is an explanation of the problem if you are interested:

Some of the provided fvf files contained the following regular expression segment (within the quotes), "`S.+`S". We should have known better than to use the "match everything" wildcard, ".+", particularly in its greedy mode, but we goofed. Consider the following two pattern files with their equivalent regular expression value and the two urls which follow.

(www`E.)`Qfamousnewspaper.com`S.+`S.fvf
==> (www.)?famousnewspaper.com/.+/
(www`E.)`Qfamousnewspaper.com`S.+`S20`Ed`Ed`S`Ed`Ed`S`Ed`Ed`S.fvf
==> (www.)?famousnewspaper.com/.+/20dd/dd/dd/

https://www.famousnewspaper.com/opinion/
https://www.famousnewspaper.com/opinion/2022/07/01/why_do_so_called_smart_people_do_dumb_things/

One would think that the first pattern would be the best match the first url and that is so. The second pattern should be the best for the second url. Even though both patterns match it, the second regular expression is longer and that is the ostensible tie-breaker. However, we have a problem since RCFV (rightfully) will always choose the first pattern it finds which matches the whole url. Unfortunately, in this case, when you test each pattern for a complete match by adding "^" at the start and "$" at the end (the protocol prefix is always ignored) you get a spurious match for the shorter pattern since the ".+" matches everything after the first slash up to the final slash. 🙁

Here is the fix:

We have replaced all of the provided file names containing the defective string "`S.+`S" with a better pattern, i.e., "`S[^`S]+`S", which will prevent any unwanted side-effects such as the above from happening. This will happen automatically when you install a new version of RCFV and you allow the installation program to overwrite any matching fvf files.

Again, for new users or those not interested in our predefined fvf files this is not an issue. However, users of older versions will also need to delete the old, imperfect fvf files.

To delete the "bad" files which have been superseded:
(1) Open a command window: type the Windows-key + {R}
(2) Open the CMD (cmd.exe) app
(3) Change the current directory to your filters folder (the one containing your fvf files). For example,
cd "C:Users{your user name}DocumentsRCFVFilters"
(4) Type the following command to delete the "bad" files:
del *`S.+`S*
(5) You can verify that this worked by typing the following command which should return no matches:
dir *`S.+`S*

Sorry for the inconvenience.