In an effort to improve things I've been paying more attention to the
error log for
www.sljfaq.org in the same way as I do for
kanji.sljfaq.org, and so I've been doing a bit of work on the programs
on
www.sljfaq.org as follows:
/cgi/kana-romaji.cgi
and the other programs now deals with malformed inputs better. Someone
was typing in percent encoded characters in the URL bar "by hand"
which caused a chain of errors when the gibberish percent encoding
couldn't be decoded. These are now handled better by using Unicode
substitution characters.
/cgi/psc.cgi
Somehow or other there was some old debugging code left in this
program. I'm working on a code linter to remove this kind of thing.
Also, there is a bug where some person sends a bunch of shift-JIS
encoding every day, I'm not sure what the purpose of this is, but the
same person seems to be requesting Shift-JIS again and again, which
has been going on for years.
/books/
* The old link "books.cgi" is now fixed to work.
* Some crazy errors due to malformed inputs (bad links or something)
are now improved, so the user gets a better error message.