What is wrong. AS names lose or garble accented letters, in two places:
The AS list built into the program has no letter outside ASCII. AS1251 reads "FUNDAO DE AMPARO PESQUISA DO ESTADO SO PAULO"; the source, https://github.com/ipverse/asn-info (as.csv), has "FUNDAÇÃO DE AMPARO À PESQUISA DO ESTADO SÃO PAULO". A name search for "São Paulo" finds nothing.
WHOIS answers from LACNIC are in Latin-1, and are stored as received. The log shows "Brasrede Telecomunica��es LTDA", and the pages show the same replacement characters.
Also, make asupdate, which rebuilds the built-in list, now stops with "Unexpected CSV header": as.csv has a fourth column, country, after asn, handle and description.
What acceptable looks like.
cmd/asinfo-gen reads today's as.csv header and keeps names exactly as the source spells them. Keeping the country column is out of scope; the generator reads past it.
pkg/asinfo/asdata.json.gz regenerated with make asupdate, accents intact.
internal/whois stores answers as UTF-8: an answer that is already valid UTF-8 is kept, and one that is not is decoded as Latin-1 (golang.org/x/text/encoding/charmap).
Definition of done.
Tests that fail on the current code: the generator's parsing of a four-column header with an accented name, and a Latin-1 WHOIS answer giving a correct UTF-8 organisation name.
The regenerated list, and one reviewed PR into next.
Model: opus-5-5
**What is wrong.** AS names lose or garble accented letters, in two places:
- The AS list built into the program has no letter outside ASCII. AS1251 reads "FUNDAO DE AMPARO PESQUISA DO ESTADO SO PAULO"; the source, https://github.com/ipverse/asn-info (`as.csv`), has "FUNDAÇÃO DE AMPARO À PESQUISA DO ESTADO SÃO PAULO". A name search for "São Paulo" finds nothing.
- WHOIS answers from LACNIC are in Latin-1, and are stored as received. The log shows "Brasrede Telecomunica��es LTDA", and the pages show the same replacement characters.
Also, `make asupdate`, which rebuilds the built-in list, now stops with "Unexpected CSV header": `as.csv` has a fourth column, `country`, after `asn`, `handle` and `description`.
**What acceptable looks like.**
- `cmd/asinfo-gen` reads today's `as.csv` header and keeps names exactly as the source spells them. Keeping the country column is out of scope; the generator reads past it.
- `pkg/asinfo/asdata.json.gz` regenerated with `make asupdate`, accents intact.
- `internal/whois` stores answers as UTF-8: an answer that is already valid UTF-8 is kept, and one that is not is decoded as Latin-1 (`golang.org/x/text/encoding/charmap`).
**Definition of done.**
- Tests that fail on the current code: the generator's parsing of a four-column header with an accented name, and a Latin-1 WHOIS answer giving a correct UTF-8 organisation name.
- The regenerated list, and one reviewed PR into `next`.
Model: opus-5-5
clawbot
self-assigned this 2026-10-26 21:33:15 +01:00
Built in #172: the generator reads today's as.csv, the built-in list is regenerated with its accents, and a WHOIS answer that is not valid UTF-8 is decoded as Latin-1.
One correction to this issue: LACNIC answers in UTF-8. For a Brazilian AS its answer already holds the replacement characters, so those WHOIS names stay garbled after this change. The PR body has the details.
Model: opus-5-5
Built in https://git.eeqj.de/sneak/routewatch/pulls/172: the generator reads today's `as.csv`, the built-in list is regenerated with its accents, and a WHOIS answer that is not valid UTF-8 is decoded as Latin-1.
One correction to this issue: LACNIC answers in UTF-8. For a Brazilian AS its answer already holds the replacement characters, so those WHOIS names stay garbled after this change. The PR body has the details.
Model: opus-5-5
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
What is wrong. AS names lose or garble accented letters, in two places:
as.csv), has "FUNDAÇÃO DE AMPARO À PESQUISA DO ESTADO SÃO PAULO". A name search for "São Paulo" finds nothing.Also,
make asupdate, which rebuilds the built-in list, now stops with "Unexpected CSV header":as.csvhas a fourth column,country, afterasn,handleanddescription.What acceptable looks like.
cmd/asinfo-genreads today'sas.csvheader and keeps names exactly as the source spells them. Keeping the country column is out of scope; the generator reads past it.pkg/asinfo/asdata.json.gzregenerated withmake asupdate, accents intact.internal/whoisstores answers as UTF-8: an answer that is already valid UTF-8 is kept, and one that is not is decoded as Latin-1 (golang.org/x/text/encoding/charmap).Definition of done.
next.Model: opus-5-5
Built in #172: the generator reads today's
as.csv, the built-in list is regenerated with its accents, and a WHOIS answer that is not valid UTF-8 is decoded as Latin-1.One correction to this issue: LACNIC answers in UTF-8. For a Brazilian AS its answer already holds the replacement characters, so those WHOIS names stay garbled after this change. The PR body has the details.
Model: opus-5-5