[Privoxy-devel] privoxy debian patches

[Lee]

Hi Roland!

On 10/16/18, Roland Rosenfeld <roland at spinnaker.de> wrote:
> Hi Lee!
>
> On Di, 16 Okt 2018, Lee wrote:
>
>> Can you try something please?
>>   privoxy/doc/source/ldp.dsl.in
>> has a "print" and a "html" section.  Can you see what happens if you
>> change
>>   06_8bit_manual.patch
>> to have the first preserve-sdata bit in the print section be false:
>> (declare-characteristic preserve-sdata?
>>   "UNREGISTERED::James Clark//Characteristic::preserve-sdata?"
>>   #f)
>>
>> and the second preserve-sdata bit in the html section be true:
>> (declare-characteristic preserve-sdata?
>>   "UNREGISTERED::James Clark//Characteristic::preserve-sdata?"
>>   #t)
>
> This has the same result as commenting out both blocks (as in
> 06_8bit_manual.patch), so your change seems the better patch, I'll
> change this here.
>
>> At least for me, that doesn't translate the "&code;" strings. eg:
>> diff --git a/doc/webserver/developer-manual/index.html
>> b/doc/webserver/developer-manual/index.html
>> index 81fd85ad..f32845d3 100644
>> --- a/doc/webserver/developer-manual/index.html
>> +++ b/doc/webserver/developer-manual/index.html
>> @@ -42,7 +42,7 @@
>> HREF="https://www.privoxy.org/user-manual/copyright.html"
>>  TARGET="_top"
>>  >Copyright</A
>>  >
>> -      <A9> 2001-2018 by
>> +      © 2001-2018 by
>>        <A
>
> Same here.  It also uses < and > instead of < and >
> which look very ugly to me when building the docs without changing
> ldp.dsl.in.
>
> But I have to admit, that my developer-manual/index.html always looks
> quite different to the one, that is delivered with 3.0.26 tar ball:
> The tarball uses nice indentation, lowercase HTML tags and ©,
> while the manual, that is created by
> ./configure --with-docbook=/usr/share/sgml/docbook/stylesheet/dsssl/modular
> && env -u LANG LC_ALL=C.UTF-8 make dok
> uses nearly no indentation, uppercase HTML tags and ISO-8859-1 (hex
> A9) as copyright sign.
> After changing the HTML part perserve-sdata to #t the copyright sign
> was changed to ©
>
> Okay, seems that the original doc was generated with "make dok-tidy",
> while I only use "make dok".

The docs in the release tarball - yes, a "make dok-tidy" was done.

> If I change the above line to run "make dok-tidy" after "make dok",
> this results in charset=utf-8 (generated by tidy) as well as an UTF-8
> copyright sign.  I have to append "--preserve-entities yes" to the
> TIDY definition to keep © also after "make dok-tidy".

That line in the makefile is getting kind of long.. I now have
TIDY       = tidy -q -modify -indent -wrap 120 --tidy-mark no
--preserve-entities yes \
 --mute MISSING_ATTRIBUTE --mute TRIM_EMPTY_ELEMENT

>> I also changed p-authors.sgml to be straight 7bit ascii - eg:
>> $ grep '&aacute' *sgml
>> p-authors.sgml: Gábor Lipták
>
> Good idea!

It's getting there..  docbook generated html looks good (at least what
I've checked) but the AUTHORS file is broken now: eg
diff --git a/AUTHORS b/AUTHORS
index 8d21693e..5368aa49 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -21,7 +21,7 @@ Former Privoxy Team Members:
  Karsten Hopp
  Alexander Lazic
  Daniel Leite
- Gábor Lipták
+ Ga'bor Lipta'k
  Adam Lock

>> After rebuilding everything (and using your neat find command :)
>> Lee at i3668 /source/privoxy/privoxy/doc
>> $ find . -name \*.html | xargs file | grep -v ASCII
>
> With or without make dok-tidy?

without

> Anyway, for the next release I'll apply the following changes:

I'd like to get all of the generally applicable debian/patches applied
to privoxy.  That way everybody gets the benefit and you have less
patches to carry forward :)

> - Replace 06_8bit_manual.patch by a new patch, that only changes
>   preserve-sdata to #t in the HTML section.
> - Append --preserve-entities yes to TIDY options in GNUmakefile.in.
> - Change all ISO-8859-1 chars in doc/source/p-authors.sgml to SGML
>   entities

I can make all those changes in git

Regards,
Lee