According to [EMAIL PROTECTED]:
> > Upon discovering on Wednesday that there was a chunk of documentation
> > missing from the external_parsers entry in attrs.html, I decided to
> > do a "diff -bu" of it against an older copy (dated Aug 26). I found
> > numerous errors were introduced as a result of the move to the new
> > cf_generate.pl script. Here's my summary of the problems:
> >
> > - string list now simply identified as string, which doesn't destinguish
> > attributes that allow more than one string
>
> Restored.
Great, thanks. I found and fixed a couple ones that you missed.
> > - quoted string list simply identified as "quoted"
>
> All of them were quoted string list and are restored.
Excellent.
> > - hrefs to htdig.html now set target='_top' (should be just for
> > htsearch.html, which is the only one that opens a new frameset)
> > - same for links to htnotify.html, htfuzzy.html, htmerge.html
>
> Fixed cf_generate.pl to handle this case.
Looks good now.
> > - cf_byprog.html links to htsearch.html without setting target frame
>
> Fixed cf_generate.pl now set target='_top' for htsearch.html only.
This one is still a problem. I'm talking about cf_byprog.html here, not
attrs.html.
> > - no commas or ands in "used by" lists
>
> Added comma.
>
> > - default values of "0" are listed as "No default"
>
> Fixed cf_generate.pl (replaced !$value by $value =~ /^\s*$/)
Great on both counts.
> > - <, > and & in default not changed to <, > and &
>
> I did not find this problem. For instance keywords_meta_tag_names
> description contains < and > that are < and > as expected. It would
> help me if you could point to one specific error.
Not the description, but the default field. This is important to get
right, because the field does double duty. It's used by the programs
to intialise the attributes (so you can't bury &foo; entities in this
field), but also used in the generated documents, where you don't want
bare ampersands or angle brackets. Specific problems are the default
values for valid_punctuation (contains &) and noindex_start (contains <).
A few substitutions in the Perl script should do it.
All the other changes look good. The description for search_results_header
wound up in the spot for allow_in_form, but I fixed that.
> > - descriptions add BIG strings to programs, which aren't even used by them
> > (should maybe be done as comments instead?)
>
> It could be done. It adds roughly 60k to the program. If 60k is considered
> unbearable, I can change default.cc so that descriptions are in comments.
> My opinion is that 60k is nothing and that these strings could be usefull if
> someone wants to implement some kind of online help. Let me know what you
> prefer.
I didn't bother to do the math before. 60k isn't much - I wouldn't worry
about it for now, unless others take exception to it.
> > several inconsistencies that were there, and it should be a big help
> > in preventing further inconsistencies from creeping in.
>
> Thanks for saying this. I deeply apologize for the trouble. I was
> careless in implementing this program and should have spent time verifying
> the inconsistencies myself.
That's why the "many eyes" philosophy to open source is a good thing.
No apologies necessary. Thanks for all the work you did on this.
--
Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]>
Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil
Dept. Physiology, U. of Manitoba Phone: (204)789-3766
Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED] containing the single word "unsubscribe" in
the SUBJECT of the message.
