Thread: Link missing in the 14 release notes
Hi, "Add predefined roles pg_read_all_data and pg_write_all_data (Stephen Frost) These non-login roles can be used to give read or write permission to all tables, views, and sequences." pg_read_all_data links to "Predefined roles", pg_write_all_data does not. Is that on purpose? At least it looks strange. Regards Daniel
On Wed, Jun 30, 2021, at 11:47 AM, Daniel Westermann (DWE) wrote:
"Add predefined roles pg_read_all_data and pg_write_all_data (Stephen Frost)These non-login roles can be used to give read or write permission to all tables, views, and sequences."pg_read_all_data links to "Predefined roles", pg_write_all_data does not. Is that on purpose? At least it looks strange.
It seems so. Once you click on the link, you will notice that pg_write_all_data
is there too. Role names are similar to make the reader suspect that both
descriptions will be on the same page.
Greetings, * Euler Taveira (euler@eulerto.com) wrote: > On Wed, Jun 30, 2021, at 11:47 AM, Daniel Westermann (DWE) wrote: > > "Add predefined roles pg_read_all_data and pg_write_all_data (Stephen Frost) > > > > These non-login roles can be used to give read or write permission to all tables, views, and sequences." > > > > pg_read_all_data links to "Predefined roles", pg_write_all_data does not. Is that on purpose? At least it looks strange. > It seems so. Once you click on the link, you will notice that pg_write_all_data > is there too. Role names are similar to make the reader suspect that both > descriptions will be on the same page. I tend to agree that it'd make sense to have them both as links. I've CC'd Bruce to make sure he sees this discussion. Thanks, Stephen
Attachment
On Thu, Jul 1, 2021 at 05:46:16PM -0400, Stephen Frost wrote: > Greetings, > > * Euler Taveira (euler@eulerto.com) wrote: > > On Wed, Jun 30, 2021, at 11:47 AM, Daniel Westermann (DWE) wrote: > > > "Add predefined roles pg_read_all_data and pg_write_all_data (Stephen Frost) > > > > > > These non-login roles can be used to give read or write permission to all tables, views, and sequences." > > > > > > pg_read_all_data links to "Predefined roles", pg_write_all_data does not. Is that on purpose? At least it looks strange. > > It seems so. Once you click on the link, you will notice that pg_write_all_data > > is there too. Role names are similar to make the reader suspect that both > > descriptions will be on the same page. > > I tend to agree that it'd make sense to have them both as links. > > I've CC'd Bruce to make sure he sees this discussion. I normally link to only the _first_ mention of something, and since they are both in the same section, I didn't add a link for the second one. Adding a second link might suggest that there is more information available, even though the coarseness of our links means that section is the same. -- Bruce Momjian <bruce@momjian.us> https://momjian.us EDB https://enterprisedb.com If only the physical world exists, free will is an illusion.
Greetings,
On Thu, Jul 1, 2021 at 21:35 Bruce Momjian <bruce@momjian.us> wrote:
On Thu, Jul 1, 2021 at 05:46:16PM -0400, Stephen Frost wrote:
> Greetings,
>
> * Euler Taveira (euler@eulerto.com) wrote:
> > On Wed, Jun 30, 2021, at 11:47 AM, Daniel Westermann (DWE) wrote:
> > > "Add predefined roles pg_read_all_data and pg_write_all_data (Stephen Frost)
> > >
> > > These non-login roles can be used to give read or write permission to all tables, views, and sequences."
> > >
> > > pg_read_all_data links to "Predefined roles", pg_write_all_data does not. Is that on purpose? At least it looks strange.
> > It seems so. Once you click on the link, you will notice that pg_write_all_data
> > is there too. Role names are similar to make the reader suspect that both
> > descriptions will be on the same page.
>
> I tend to agree that it'd make sense to have them both as links.
>
> I've CC'd Bruce to make sure he sees this discussion.
I normally link to only the _first_ mention of something, and since they
are both in the same section, I didn't add a link for the second one.
Adding a second link might suggest that there is more information
available, even though the coarseness of our links means that section is
the same.
Perhaps the wording would be better as something like:
New <a href=“…”>predefined roles</a> for granting access to read/write all tables have been added, called pg_read_all_data and pg_write_all_data ..?
Or something along those lines?
Just a thought.
Perhaps another idea would be to make one link which includes both names.
Thanks,
Stephen
On Thu, Jul 1, 2021 at 09:47:47PM -0400, Stephen Frost wrote: > On Thu, Jul 1, 2021 at 21:35 Bruce Momjian <bruce@momjian.us> wrote: > I normally link to only the _first_ mention of something, and since they > are both in the same section, I didn't add a link for the second one. > Adding a second link might suggest that there is more information > available, even though the coarseness of our links means that section is > the same. > > Perhaps the wording would be better as something like: > > New <a href=“…”>predefined roles</a> for granting access to read/write all > tables have been added, called pg_read_all_data and pg_write_all_data ..? > > Or something along those lines? I have avoided doing that since the link is about the roles, not about the term "predefined roles". > Just a thought. > > Perhaps another idea would be to make one link which includes both names. Uh, that is going to look odd, I am afraid. -- Bruce Momjian <bruce@momjian.us> https://momjian.us EDB https://enterprisedb.com If only the physical world exists, free will is an illusion.
Greetings,
On Thu, Jul 1, 2021 at 21:57 Bruce Momjian <bruce@momjian.us> wrote:
On Thu, Jul 1, 2021 at 09:47:47PM -0400, Stephen Frost wrote:
> On Thu, Jul 1, 2021 at 21:35 Bruce Momjian <bruce@momjian.us> wrote:
> I normally link to only the _first_ mention of something, and since they
> are both in the same section, I didn't add a link for the second one.
> Adding a second link might suggest that there is more information
> available, even though the coarseness of our links means that section is
> the same.
>
> Perhaps the wording would be better as something like:
>
> New <a href=“…”>predefined roles</a> for granting access to read/write all
> tables have been added, called pg_read_all_data and pg_write_all_data ..?
>
> Or something along those lines?
I have avoided doing that since the link is about the roles, not about
the term "predefined roles".
I tend to agree that it’s about the roles, but it isn’t about just one of them …
> Just a thought.
>
> Perhaps another idea would be to make one link which includes both names.
Uh, that is going to look odd, I am afraid.
This thread started with the point that the current entry looks odd. I’m not sure this would look more odd than what is there now.
Thanks,
Stephen
On Thu, Jul 1, 2021 at 10:01:46PM -0400, Stephen Frost wrote:
> > Just a thought.
> >
> > Perhaps another idea would be to make one link which includes both names.
>
> Uh, that is going to look odd, I am afraid.
>
> This thread started with the point that the current entry looks odd. I’m not
> sure this would look more odd than what is there now.
If you look at the current doc version:
https://momjian.us/pgsql_docs/release-14.html
Notice that for these two items:
Remove factorial operators ! and !! (Mark Dilger)
The factorial() function is still supported. Also remove function numeric_fac().
Disallow factorial() of negative numbers (Peter Eisentraut)
Previously such cases returned 1.
Only the first mention of "factorial()" uses a link; the second item
doesn't have any link.
--
Bruce Momjian <bruce@momjian.us> https://momjian.us
EDB https://enterprisedb.com
If only the physical world exists, free will is an illusion.
>On Thu, Jul 1, 2021 at 10:01:46PM -0400, Stephen Frost wrote: >> > Just a thought. >> > >> > Perhaps another idea would be to make one link which includes both names. >> >> Uh, that is going to look odd, I am afraid. >> >> This thread started with the point that the current entry looks odd. I’m not >> sure this would look more odd than what is there now. >If you look at the current doc version: > > https://momjian.us/pgsql_docs/release-14.html > >Notice that for these two items: > > Remove factorial operators ! and !! (Mark Dilger) > The factorial() function is still supported. Also remove function numeric_fac(). > > Disallow factorial() of negative numbers (Peter Eisentraut) > Previously such cases returned 1. > >Only the first mention of "factorial()" uses a link; the second item >doesn't have any link. There are other places as well, like this on: "The postgres_fdw supports these type of scans if async_capable is set." async_capable is not a link. Or this one: "This speeds normalize() and IS NORMALIZED." I still think it should be more consistent and all should be links, even if they point to the same section. Regards Daniel
On Thu, Jul 1, 2021, at 11:01 PM, Stephen Frost wrote:
I have avoided doing that since the link is about the roles, not aboutthe term "predefined roles".I tend to agree that it’s about the roles, but it isn’t about just one of them …
I think it is clear that "predefined roles" link contains the new roles, no?
On Fri, Jul 2, 2021 at 05:15:23AM +0000, Daniel Westermann (DWE) wrote: > >On Thu, Jul 1, 2021 at 10:01:46PM -0400, Stephen Frost wrote: > >Only the first mention of "factorial()" uses a link; the second item > >doesn't have any link. > > There are other places as well, like this on: > "The postgres_fdw supports these type of scans if async_capable is set." > > async_capable is not a link. Or this one: > "This speeds normalize() and IS NORMALIZED." > > I still think it should be more consistent and all should be links, even if they point to the same section. I am using standard typographic style, where the first mention of something is often italics, but later mentions might not be. I am not in favor of adding links to everthing that can have a link _unless_ there is a unique URI for each item. -- Bruce Momjian <bruce@momjian.us> https://momjian.us EDB https://enterprisedb.com If only the physical world exists, free will is an illusion.
Greetings, * Bruce Momjian (bruce@momjian.us) wrote: > On Fri, Jul 2, 2021 at 05:15:23AM +0000, Daniel Westermann (DWE) wrote: > > >On Thu, Jul 1, 2021 at 10:01:46PM -0400, Stephen Frost wrote: > > >Only the first mention of "factorial()" uses a link; the second item > > >doesn't have any link. > > > > There are other places as well, like this on: > > "The postgres_fdw supports these type of scans if async_capable is set." > > > > async_capable is not a link. Or this one: > > "This speeds normalize() and IS NORMALIZED." > > > > I still think it should be more consistent and all should be links, even if they point to the same section. > > I am using standard typographic style, where the first mention of > something is often italics, but later mentions might not be. I am not > in favor of adding links to everthing that can have a link _unless_ > there is a unique URI for each item. While I understand that's the approach you've been following, I don't agree with it. Providing links gives users a way to find more information and that applies to each equally. I don't believe users of the website and release notes would be confused, upset, or otherwise misunderstand when two items which are covered in the same place have a repeated link for each of the items. Thanks, Stephen