Обсуждение: Undesirable whitespace treatment in mail archive display
Since we moved to the new website design, the mailing list archives display all messages with leading indentation suppressed. This can be pretty awful, eg in the code fragment here: https://www.postgresql.org/message-id/20180611170812.7ypf6xs6dlteo57j@alap3.anarazel.de Looking closer, I think it may not just be leading indentation that gets lost; seems like *any* whitespace string gets smashed to a single space, cf followup at https://www.postgresql.org/message-id/7382.1528737252%40sss.pgh.pa.us which certainly does not look like it did in my outbox. I assume this is just an ill-considered stylesheet choice somewhere ... can we fix it please? regards, tom lane
> On Jun 11, 2018, at 1:47 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > Since we moved to the new website design, the mailing list archives > display all messages with leading indentation suppressed. This can > be pretty awful, eg in the code fragment here: > > https://www.postgresql.org/message-id/20180611170812.7ypf6xs6dlteo57j@alap3.anarazel.de > > Looking closer, I think it may not just be leading indentation that > gets lost; seems like *any* whitespace string gets smashed to a > single space, cf followup at > > https://www.postgresql.org/message-id/7382.1528737252%40sss.pgh.pa.us > > which certainly does not look like it did in my outbox. > > I assume this is just an ill-considered stylesheet choice somewhere > ... can we fix it please? Hmm…agreed. I’m going to check the old code to see how that was handled. This may have occurred when trying to make the archives readable across different viewports. Jonathan
> On 11 Jun 2018, at 19:52, Jonathan S. Katz <jkatz@postgresql.org> wrote: > >> >> On Jun 11, 2018, at 1:47 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: >> >> Since we moved to the new website design, the mailing list archives >> display all messages with leading indentation suppressed. This can >> be pretty awful, eg in the code fragment here: >> >> https://www.postgresql.org/message-id/20180611170812.7ypf6xs6dlteo57j@alap3.anarazel.de >> >> Looking closer, I think it may not just be leading indentation that >> gets lost; seems like *any* whitespace string gets smashed to a >> single space, cf followup at >> >> https://www.postgresql.org/message-id/7382.1528737252%40sss.pgh.pa.us >> >> which certainly does not look like it did in my outbox. >> >> I assume this is just an ill-considered stylesheet choice somewhere >> ... can we fix it please? > > Hmm…agreed. I’m going to check the old code to see how that was handled. This > may have occurred when trying to make the archives readable across different > viewports. The below diff will retain the whitespace included the post. It doesn’t seem to make it less readable on smaller viewports (although more scrolling), but I only tested quickly so more input is welcome. cheers ./daniel diff --git a/media/css/main.css b/media/css/main.css index 6221d90..46e427d 100644 --- a/media/css/main.css +++ b/media/css/main.css @@ -771,6 +771,7 @@ h3.messages { font-size: 0.8em; line-height: 1.2em; margin: 0 auto; + white-space: pre; }
On Jun 11, 2018, at 1:58 PM, Daniel Gustafsson <daniel@yesql.se> wrote:On 11 Jun 2018, at 19:52, Jonathan S. Katz <jkatz@postgresql.org> wrote:
On Jun 11, 2018, at 1:47 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Since we moved to the new website design, the mailing list archives
display all messages with leading indentation suppressed. This can
be pretty awful, eg in the code fragment here:
https://www.postgresql.org/message-id/20180611170812.7ypf6xs6dlteo57j@alap3.anarazel.de
Looking closer, I think it may not just be leading indentation that
gets lost; seems like *any* whitespace string gets smashed to a
single space, cf followup at
https://www.postgresql.org/message-id/7382.1528737252%40sss.pgh.pa.us
which certainly does not look like it did in my outbox.
I assume this is just an ill-considered stylesheet choice somewhere
... can we fix it please?
Hmm…agreed. I’m going to check the old code to see how that was handled. This
may have occurred when trying to make the archives readable across different
viewports.
The below diff will retain the whitespace included the post. It doesn’t seem
to make it less readable on smaller viewports (although more scrolling), but I
only tested quickly so more input is welcome.
I did some testing locally and it does solve the whitespace problem, but
introduces the following:
1. The line heights are not a bit larger, which could introduce the return of
the “scrolling” problem.
2. We lose the mobile-friendly wrapping that was introduced.
Looking at our mobile traffic numbers, I’d be inclined to solve the immediate
issue of the whitespace problem and make it a bit harder to use the archives on
mobile temporarily while looking for a better solution. In other words, we don’t
have that much mobile traffic in the archives.
I can tweak the line heights the make the scrolling not as bad, if not the same.
I’ll experiment and once it looks like before push / commit.
Jonathan
On Jun 11, 2018, at 2:09 PM, Jonathan S. Katz <jkatz@postgresql.org> wrote:
On Jun 11, 2018, at 1:58 PM, Daniel Gustafsson <daniel@yesql.se> wrote:On 11 Jun 2018, at 19:52, Jonathan S. Katz <jkatz@postgresql.org> wrote:
On Jun 11, 2018, at 1:47 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Since we moved to the new website design, the mailing list archives
display all messages with leading indentation suppressed. This can
be pretty awful, eg in the code fragment here:
https://www.postgresql.org/message-id/20180611170812.7ypf6xs6dlteo57j@alap3.anarazel.de
Looking closer, I think it may not just be leading indentation that
gets lost; seems like *any* whitespace string gets smashed to a
single space, cf followup at
https://www.postgresql.org/message-id/7382.1528737252%40sss.pgh.pa.us
which certainly does not look like it did in my outbox.
I assume this is just an ill-considered stylesheet choice somewhere
... can we fix it please?
Hmm…agreed. I’m going to check the old code to see how that was handled. This
may have occurred when trying to make the archives readable across different
viewports.
The below diff will retain the whitespace included the post. It doesn’t seem
to make it less readable on smaller viewports (although more scrolling), but I
only tested quickly so more input is welcome.I did some testing locally and it does solve the whitespace problem, butintroduces the following:1. The line heights are not a bit larger, which could introduce the return ofthe “scrolling” problem.2. We lose the mobile-friendly wrapping that was introduced.Looking at our mobile traffic numbers, I’d be inclined to solve the immediateissue of the whitespace problem and make it a bit harder to use the archives onmobile temporarily while looking for a better solution. In other words, we don’thave that much mobile traffic in the archives.I can tweak the line heights the make the scrolling not as bad, if not the same.I’ll experiment and once it looks like before push / commit.
Below is what I’m thinking. This should keep the whitespace and
continue to keep the scroll relatively tight. It also helps to not look as bad on
mobile, though it’s still not great.
If this is acceptable I can get it committed.
Jonathan
diff --git a/django/media/css/main.css b/django/media/css/main.css
index 6221d90..39358d3 100644
--- a/django/media/css/main.css
+++ b/django/media/css/main.css
@@ -768,9 +768,14 @@ h3.messages {
.message-content {
display: inline-block;
font-family: monospace, monospace;
- font-size: 0.8em;
- line-height: 1.2em;
+ font-size: 0.7em;
+ line-height: 1.1em;
margin: 0 auto;
+ white-space: pre;
+}
+
+.message-content p {
+ margin-bottom: 0;
}
On Jun 11, 2018, at 2:22 PM, Jonathan S. Katz <jkatz@postgresql.org> wrote:
On Jun 11, 2018, at 2:09 PM, Jonathan S. Katz <jkatz@postgresql.org> wrote:
On Jun 11, 2018, at 1:58 PM, Daniel Gustafsson <daniel@yesql.se> wrote:On 11 Jun 2018, at 19:52, Jonathan S. Katz <jkatz@postgresql.org> wrote:
On Jun 11, 2018, at 1:47 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Since we moved to the new website design, the mailing list archives
display all messages with leading indentation suppressed. This can
be pretty awful, eg in the code fragment here:
https://www.postgresql.org/message-id/20180611170812.7ypf6xs6dlteo57j@alap3.anarazel.de
Looking closer, I think it may not just be leading indentation that
gets lost; seems like *any* whitespace string gets smashed to a
single space, cf followup at
https://www.postgresql.org/message-id/7382.1528737252%40sss.pgh.pa.us
which certainly does not look like it did in my outbox.
I assume this is just an ill-considered stylesheet choice somewhere
... can we fix it please?
Hmm…agreed. I’m going to check the old code to see how that was handled. This
may have occurred when trying to make the archives readable across different
viewports.
The below diff will retain the whitespace included the post. It doesn’t seem
to make it less readable on smaller viewports (although more scrolling), but I
only tested quickly so more input is welcome.I did some testing locally and it does solve the whitespace problem, butintroduces the following:1. The line heights are not a bit larger, which could introduce the return ofthe “scrolling” problem.2. We lose the mobile-friendly wrapping that was introduced.Looking at our mobile traffic numbers, I’d be inclined to solve the immediateissue of the whitespace problem and make it a bit harder to use the archives onmobile temporarily while looking for a better solution. In other words, we don’thave that much mobile traffic in the archives.I can tweak the line heights the make the scrolling not as bad, if not the same.I’ll experiment and once it looks like before push / commit.Below is what I’m thinking. This should keep the whitespace andcontinue to keep the scroll relatively tight. It also helps to not look as bad onmobile, though it’s still not great.If this is acceptable I can get it committed.
Committed and pushed. Changes should be visible in that thread and as caches
clear.
I did some more research into the CSS directive Daniel posted and found that
“pre-wrap” can give us the same effect that we had on mobile before (though
long lines don’t look great) so I think we have as close to best-of-all-worlds for
now.
Please let me know if this adversely affects scrolling and I can look into adjusting.
Thanks,
Jonathan
> On 11 Jun 2018, at 20:40, Jonathan S. Katz <jkatz@postgresql.org> wrote: > I did some more research into the CSS directive Daniel posted and found that > “pre-wrap” can give us the same effect that we had on mobile before (though > long lines don’t look great) so I think we have as close to best-of-all-worlds for > now. For reading indented code, sideways scrolling is preferrable to wrapping or whitespace squashing so +1 on the current approach. If we wan’t to improve further I think pulling down the font-size for mobile devices in landscape mode is the way forward, but let’s await actual complaints before hacking on it further since desktop will most likely be the main viewport for the lists. cheers ./daniel
"Jonathan S. Katz" <jkatz@postgresql.org> writes: > Committed and pushed. Changes should be visible in that thread and as caches > clear. Hm ... there's now a surprisingly large amount of vertical space between paragraphs (ie, double newlines in the original text). Weird that this affects that. regards, tom lane
> On Jun 11, 2018, at 3:06 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > "Jonathan S. Katz" <jkatz@postgresql.org> writes: >> Committed and pushed. Changes should be visible in that thread and as caches >> clear. > > Hm ... there's now a surprisingly large amount of vertical space between > paragraphs (ie, double newlines in the original text). Weird that this > affects that. Yeah, that’s an effect of the “pre” text. I tried to reduce the impact by lowering the font size and line height, but it still starkly stands out. I’ll do some research to see if there is another CSS property to help minimize that gap. Jonathan
On Jun 11, 2018, at 3:11 PM, Jonathan S. Katz <jkatz@postgresql.org> wrote:On Jun 11, 2018, at 3:06 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
"Jonathan S. Katz" <jkatz@postgresql.org> writes:Committed and pushed. Changes should be visible in that thread and as caches
clear.
Hm ... there's now a surprisingly large amount of vertical space between
paragraphs (ie, double newlines in the original text). Weird that this
affects that.
Yeah, that’s an effect of the “pre” text. I tried to reduce the impact by
lowering the font size and line height, but it still starkly stands out.
I’ll do some research to see if there is another CSS property to help
minimize that gap.
Based upon some research this did not sound like a problem people were
trying to solve, but I tested, committed and pushed a change that should help
reduce the gap. Will be available as caches clear, but I did apply it to originally
discussed thread[1].
Thanks,
Jonathan
"Jonathan S. Katz" <jkatz@postgresql.org> writes: >> On Jun 11, 2018, at 3:11 PM, Jonathan S. Katz <jkatz@postgresql.org> wrote: >>> On Jun 11, 2018, at 3:06 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: >>>> Hm ... there's now a surprisingly large amount of vertical space between >>>> paragraphs (ie, double newlines in the original text). Weird that this >>>> affects that. > Based upon some research this did not sound like a problem people were > trying to solve, but I tested, committed and pushed a change that should help > reduce the gap. Will be available as caches clear, but I did apply it to originally > discussed thread[1]. Thanks, it looks more proportionate now. The text is all rather tightly set now, but I guess we're doing that intentionally to conserve screen space. regards, tom lane
> On Jun 11, 2018, at 4:23 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > "Jonathan S. Katz" <jkatz@postgresql.org> writes: >>> On Jun 11, 2018, at 3:11 PM, Jonathan S. Katz <jkatz@postgresql.org> wrote: >>>> On Jun 11, 2018, at 3:06 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote: >>>>> Hm ... there's now a surprisingly large amount of vertical space between >>>>> paragraphs (ie, double newlines in the original text). Weird that this >>>>> affects that. > >> Based upon some research this did not sound like a problem people were >> trying to solve, but I tested, committed and pushed a change that should help >> reduce the gap. Will be available as caches clear, but I did apply it to originally >> discussed thread[1]. > > Thanks, it looks more proportionate now. The text is all rather tightly > set now, but I guess we're doing that intentionally to conserve screen > space. Pushed a fixed to relax the spacing that should still conserve screen space. Available after next cache clear, etc. Jonathan