Bug 724880 – packages.gentoo.org atom feeds are corrupted

• Status: RESOLVED FIXED
• Product: Websites

[Description Enne Eziarc 2020-05-24 01:11:11 UTC]

Feeds contain invalid XML in package descriptions:

 $ curl -sSL https://packages.gentoo.org/packages/added.atom | xmllint -
-:422: parser error : xmlParseEntityRef: no name
            <title>media-video/vidcutter (FFmpeg-based simple video cutter & joi
                                                                            ^
-:431: parser error : xmlParseEntityRef: no name
            <title>media-video/vidcutter (FFmpeg-based simple video cutter & joi
                                                                            ^
-:3050: parser error : StartTag: invalid element name
        <title>sci-libs/m4rie (Fast dense matrix arithmetic over GF(2^e) for 2 <
                                                                               ^
-:3050: parser error : StartTag: invalid element name
   <title>sci-libs/m4rie (Fast dense matrix arithmetic over GF(2^e) for 2 <= e <
                                                                               ^
-:4742: parser error : xmlParseEntityRef: no name
            <title>dev-haskell/basement (Foundation scrap box of array & string)
                                                                        ^
-:5111: parser error : xmlParseEntityRef: no name
            <title>dev-libs/boxfort (Convenient & cross-platform sandboxing C li
                                                 ^
-:6137: parser error : xmlParseEntityRef: no name
      <title>dev-python/pyproject2setuppy (Cheap setup.py hack to install flit &
                                                                               ^

According to my feed reader, the last time they were working was 2020-05-02 16:00

[Comment 1 Luca Santarelli 2020-05-25 22:12:39 UTC]

Can confirm it's a broken/missinge escaping of entities.

[Comment 2 Enne Eziarc 2020-05-27 07:43:39 UTC]

This has possibly been broken for a long time (or just never worked), except prior to commit bd59c15958c3[1] it was infrequent enough to go ignored.

[1] https://gitweb.gentoo.org/sites/soko.git/commit/?id=bd59c15958c319d0922bc52b20b2783be6740ae6

[Comment 3 Alec Warner (RETIRED) 2020-05-27 21:04:38 UTC]

We rewrote the app earlier this year and that rewrite launched with busted feeds. I plan on fixing them soon.

-A

[Comment 4 Luca Santarelli 2020-05-27 21:12:20 UTC]

(In reply to Alec Warner from comment #3)
> We rewrote the app earlier this year and that rewrite launched with busted
> feeds. I plan on fixing them soon.
> 
> -A

https://golang.org/pkg/html/#EscapeString

Called on every string that gets written in a XML (title, content)

[Comment 5 Luca Santarelli 2020-05-27 21:14:30 UTC]

Plus, albeit unrelated to the current bug (sorry), please add a README/BUILDING/some docs to help people help you.

[Comment 6 Alec Warner (RETIRED) 2020-05-27 23:53:01 UTC]

(In reply to Luca Santarelli from comment #4)
> (In reply to Alec Warner from comment #3)
> > We rewrote the app earlier this year and that rewrite launched with busted
> > feeds. I plan on fixing them soon.
> > 
> > -A
> 
> https://golang.org/pkg/html/#EscapeString
> 
> Called on every string that gets written in a XML (title, content)

At this point we have migrated to gorilla feeds; the existing feeds were  using text/template which was never appropriate for atom feeds anyways.

-A