Re: "Python 37/38 conflict, was Re: Trubles compiling lxqt on RPi4" [really: list interface replacement generates bad textual content]

On Tue, May 18, 2021 at 12:08:36AM -0700, Mark Millard wrote:
> On 2021-May-18, at 00:03, Mark Millard <marklmi at yahoo.com> wrote:
> 
> > 
> > On 2021-May-17, at 23:55, Mark Millard <marklmi at yahoo.com> wrote:
> > 
> >> On 2021-May-17, at 23:37, Baptiste Daroussin <bapt at FreeBSD.org> wrote:
> >>> 
> >>> 
> >>> 18 mai 2021 06:50:03 Mark Millard <marklmi@yahoo.com>:
> >>> 
> >>>> On 2021-May-15, at 22:22, Baptiste Daroussin <bapt@FreeBSD.org> wrote:
> >>>> 
> >>>> 
> >>>>> 16 mai 2021 05:29:07 Mark Millard <marklmi@yahoo.com>:
> >>>>> 
> >>>>>> In exploring I've seen various web-page presentation of
> >>>>>> freebsd-bugs messages that seem to be messed up. An
> >>>>>> example is:
> >>>>>> 
> >>>>>> . . .
> >>>>> 
> >>>>> Nice catch, I will fix it
> >>>>> 
> >>>> 
> >>>> The extra text disappeared when I later looked at the
> >>>> pages via the URLs.
> >>>> 
> >>>> But I've run into another oddity. I did a google
> >>>> search for:
> >>>> 
> >>>> "Comparing the OverDrive 1000 (A57) vs. MACCHIATObin"
> >>>> 
> >>>> and then used the 1st link shown, the one from the text:
> >>>> 
> >>>> QUOTE
> >>>> https://lists.freebsd.org › freebsd-arm › 2019-December
> >>>> Comparing the OverDrive 1000 (A57) vs. MACCHIATObin ...
> >>>> Dec 2, 2019 — Comparing the OverDrive 1000 (A57) vs. MACCHIATObin Double Shot (A72) for buildworld and via a CPU/cache/RAM tradeoff-exploring ...
> >>>> END QUOTE
> >>>> 
> >>>> which has the link (it is for one of my old messages):
> >>>> 
> >>>> https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiPhp3OtdLwAhVH6Z4KHaXQC68QFjAAegQIAxAD&url=https%3A%2F%2Flists.freebsd.org%2Fpipermail%2Ffreebsd-arm%2F2019-December%2F020854.html&usg=AOvVaw2VHrNUGc-Vce4CjAmTufxq
> >>>> 
> >>>> Or, extracting/converting the relevant part of that text:
> >>>> 
> >>>> https://lists.freebsd.org/pipermail/freebsd-arm/2019-December/020854.html
> >>>> 
> >>>> This now takes me to the wrong message on the list, one titled:
> >>>> 
> >>>> Re: 64 bit ARM systems with more than four cores Ed Maste
> >>>> 
> >>>> Turns out that the my old message now shows up as 2 later
> >>>> by number (020854 -> 020856):
> >>>> 
> >>>> https://lists.freebsd.org/pipermail/freebsd-arm/2019-December/020856.html
> >>>> 
> >>>> gets to the message with the correct title.
> >>>> 
> >>>> It appears that the numbering has not been preserved
> >>>> in the conversion. This can make google search results
> >>>> that point into the (older) freebsd lists
> >>>> problematical.
> >>>> 
> >>>> 
> >>>> ===
> >>>> Mark Millard
> >>>> marklmi at yahoo.com
> >>>> ( dsl-only.net went
> >>>> away in early 2018-Mar)
> >>> 
> >>> 99.9% of the tume the numbering is preserved, sometimes it is not given the weirdness of the things that happened in the past. i will dig into this specific case to see what I can do. But don t expect much here.
> >>> 
> >> 
> >> FYI:
> >> 
> >> In my "binary search inspired" looking around everything
> >> I looked at that was old but more recent than 2014-Dec-23
> >> had the offset of 2. Everything older than 2014-Dec-10
> >> worked. Those dates are as close together as I got.
> >> 
> >> (Finding things that have useful google search matches in
> >> a time frame is messy so the results are limited. It
> >> would be nice to find a better way to looking up old
> >> references.)
> >> 
> >> Example:
> >> 
> >> "lsof on FreeBSD/arm" goggle search click through to page
> >> 
> >> vs.:
> >> 
> >> Tuesday, 23 December 2014
> >> . . . 
> >> 	• Re: lsof on FreeBSD/arm Larry Rosenman 
> >> 	• [Differential] [Commented On] D1317: Build gperf as a dependency before gcc for arm releases emaste (Ed Maste) 
> >> 	• [Differential] [Updated, 8 lines] D1317: Build gperf as a dependency before gcc for arm releases emaste (Ed Maste) 
> >> 
> >> It gets on to the message for last line (still off by two for the
> >> number, like what I reported already):
> >> 
> >> By contrast:
> >> 
> >> "watchdogd not working anymore (since r273154+)" google search click through
> >> 
> >> vs.:
> >> 
> >> Wednesday, 10 December 2014
> >> . . .
> >> 	• Re: RPi - watchdogd not working anymore (since r273154+) Andreas Schwarz 
> >> 
> >> worked.
> > 
> > A different possible oddity that I just noticed is that
> > the old system used UTC dates/times, such as:
> > 
> > Mark Millard marklmi at yahoo.com 
> > Sat May 1 19:50:18 UTC 2021
> > 
> > (from freebsd-git).
> > 
> > But the new one uses various time zones (the
> > senders time zone?), such as:
> > 
> > From: Mark Millard via freebsd-arm <freebsd-arm_at_freebsd.org> 
> > Date: Sat, 15 May 2021 14:43:45 -0700
> > 
> > (from freebsd-arm).
> > 
> 
> Well . . .
> 
> freebsd-arm's messages have a UTC time after the message, such as:
> 
> Received on Sat May 15 2021 - 21:43:45 UTC
> 
> so both time zones are available. It might be deliberate.

After analysis, freebsd-bugs mailing list archives have been edited manually
multiple times in the last 18 years, making it impossible to regenerates in a
reliable manner archive with compatible URL.

This happenned too often to be able to be easily fixed.

Best regards,
Bapt