Hi Peter,
Thank you for taking your time to comment on this long-discussed topic.
I was not around when all the great work was done to implement XSLT
transforms in branch 10 and convert branch 11 to XML, but judging by
this thread
(https://www.postgresql.org/message-id/flat/4ffd72d6-8ab6-37c6-d7d5-dfed8967c0fc%40gmail.com),
much of the testing was automated and could be re-applied to older
branches.
Taking into account that both XSLT transforms and testing scripts can be
reused for the most part, do you think that the benefits of having
consistent source (easy back-porting of doc patches and one-time
translation to all languages) in all the supported branches could
potentially outweigh the inevitable conversion overhead? Are there any
specific features that make branch 10 considerably different from its
predecessors that we are missing?
On 04/30/2018 10:20 PM, Peter Eisentraut wrote:
> On 4/27/18 11:03, Bruce Momjian wrote:
>> On Fri, Apr 27, 2018 at 11:00:36AM -0400, Peter Eisentraut wrote:
>>> On 4/23/18 05:54, Liudmila Mantrova wrote:
>>>> Reading this thread, I got an impression that everyone would benefit
>>>> from converting back branches to XML, but the main concern is lack of
>>>> resources to complete this task. Are there any other issues that affect
>>>> this decision? Looks like Aleksander Lakhin's offer to prepare patches
>>>> was missed somehow as the discussion sidetracked to other issues
>>> That proposal seemed to indicate not only converting the source code to
>>> XML but also the build system to XSL. The latter is out of the
>>> question, I think.
>> Why is that?
> Because there would be a thousand lines of tooling changes to be
> backported and thousands of pages of documentation to be checked
> manually that it doesn't create a mess (times branches times platforms).
>
--
Liudmila Mantrova
Technical writer at Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company