Re: split func.sgml to separated individual sgml files - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: split func.sgml to separated individual sgml files
Date
Msg-id ed6705c4-69ba-4a16-b7a9-442dcc1d2ea5@dunslane.net
Whole thread Raw
In response to Re: split func.sgml to separated individual sgml files  (Nazir Bilal Yavuz <byavuz81@gmail.com>)
Responses Re: split func.sgml to separated individual sgml files
Re: split func.sgml to separated individual sgml files
List pgsql-hackers
On 2025-09-12 Fr 10:12 AM, Nazir Bilal Yavuz wrote:
> Hi,
>
> On Tue, 2 Sept 2025 at 17:54, Andrew Dunstan <andrew@dunslane.net> wrote:
>> Ah, you’re right, but then again,  I’d expect ALL_SGML to be used consistently, but it isn't and I didn't check.
>> v3 does that.
>> Note that GENERATED_SGML where'te included in these two targets but I think there's no harm in checking them too.
>>
>> Do we actually care about those? I don't want to add needless cycles anywhere. I note that the meson.build doesn't
appearto have a check target at all, or anything that looks for hard tabs or nbsps.Those checks were added to the
Makefileback in October in commit 5b7da5c261d, but that got missed even though Daniel had mentioned it in the
discussionthread.[1]
 
> I have been working on running these checks under the Meson build
> system. To do this, I converted the checks into a Perl script
> (sgml_syntax_check) and ran it against both the Makefile and Meson.
> Test's name is 'sgml_syntax_check' in the Meson. One difference I
> noticed: I could not find a way in Meson to create a test that does
> not run by default. As a result, this syntax test runs every time you
> run the 'meson test'. This behaviour differs from Autoconf, but I
> think it is acceptable.
>
> Additionally, some of the CI OSes were missing docbook-xml; but it has
> now been installed.
>
> I did not create a new thread for that, I can create one if you think
> that it would be better.
>
> CI run with the attached patch applied:
> https://cirrus-ci.com/build/6610354173640704


Hi Bilal,

This got preempted slightly by Tom's commit 170a8a3f460, but I think 
it's worth doing. I tried to simplify it some. See attached. There 
doesn't seem to me to be any point in using a different set of files for 
the tab tests and the NBSP tests. If we use the same set of files we can 
improve the efficiency easily by opening them only once. Here we just 
look for all the sgml files and all the xsl files and process them all.

WDYT?



cheers


andrew

--
Andrew Dunstan
EDB: https://www.enterprisedb.com

Attachment

pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Re: Fixing MSVC's inability to detect elog(ERROR) does not return
Next
From: Tom Lane
Date:
Subject: Re: split func.sgml to separated individual sgml files