Re: New archives layout is not an improvement

> On Apr 20, 2018, at 3:08 PM, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
>
> Here's something funny about the way quoted-printable encodings are
> processed in archives. I attach two screenshots; notice how the From is
> not decoded properly in both cases. Many names with accents are
> displayed just fine, why not these? There are other examples of this
> behavior, including in the subject in a few places.
>
> the messages are
> https://www.postgresql.org/message-id/CY1PR16MB04902D3E0DF77861C206CFB2C6F00%40CY1PR16MB0490.namprd16.prod.outlook.com
> https://www.postgresql.org/message-id/DM3PR17MB076480E0C6E07C3DA9F861F9A2FD0%40DM3PR17MB0764.namprd17.prod.outlook.com <https://www.postgresql.org/message-id/DM3PR17MB076480E0C6E07C3DA9F861F9A2FD0@DM3PR17MB0764.namprd17.prod.outlook.com>
I tried loading these messages locally[1] using to see if I could reproduce.
Unfortunately, it loaded cleanly.

I also tried to see if anything had changed in the archives parsing[2] code
since that message was committed. The one thing I see is that [2] depends
on “tidy” and IIRC there were some issues with libtidy after the boxes the
various apps were on were upgraded. Thus I defer to Magnus.

So perhaps the fix is as simple as the messages need to be reimported.

Jonathan

[1] https://git.postgresql.org/gitweb/?p=pgarchives.git;a=blob;f=loader/load_message.py
<https://git.postgresql.org/gitweb/?p=pgarchives.git;a=blob;f=loader/load_message.py>[2] https://git.postgresql.org/gitweb/?p=pgarchives.git;a=blob;f=loader/lib/parser.py <https://git.postgresql.org/gitweb/?p=pgarchives.git;a=blob;f=loader/lib/parser.py>