Mohammed Al Sahaf

Got it. Try: ``` xcaddy build --with github.com/vrongmeal/caddygit/module/git --with github.com/vrongmeal/caddygit=./caddygit ``` This will underscore-import `github.com/vrongmeal/caddygit/module/git` and `github.com/vrongmeal/caddygit` (the latter of which is harmless), but replaces `github.com/vrongmeal/caddygit` with the local directory.

> I build caddy like this on my local Mac. Server is running Ubuntu. > > ``` > GOOS=linux GOARCH=amd64 xcaddy build --with github.com/caddy-dns/cloudflare --with github.com/vrongmeal/caddygit/module/git=./caddygit > ``` > >...

> I'd like to give implementing this a shot, where could I find the path selector in this case? You'll be working in this range: https://github.com/caddyserver/caddy/blob/821a08a6e39ed0e7c43b0271ccf126c194eb6339/caddyconfig/caddyfile/parse.go#L369-L410

The issue is with on Windows side. Perhaps some further configuration parameters need to be set. I found this, but I cannot validate if this is truly the issue or...

> Thanks for the signature improvements! With stable 2.6.0, results are better, and a little weird, though it's well possible this isn't the way a release blob is supposed to...

I'm happy this PR exists. However, it introduces a breaking change: `*os.File` -> `Pty`. The func `Start` and `StartWithAttrs` return `*os.File` currently, but the PR changes this to type `Pty`.

Kinda but not exactly. It's expected if it were part of the selected columns, because SQL engines end up naming the column with that. It's unexpected when it's one of...

> This has been fixed > > https://play.sqlc.dev/p/8b270e64462166a51326f9c26712b83cc39b5e69c67617b6d14185604ab33e07 Is the field name becoming `Lower` expected behavior? https://play.sqlc.dev/p/3d035e555d980f57e8a93672f9a1aec4a25fbcce02edff8946a528f3fc989b92

The `git archive` command only packs tracked and committed files. So unignoring the `vendor` directory won't help, neither will running `go mod vendor` to create the directory. It can pack...

> then, if we do this, we might need to implement the archiving ourselves instead of relying on `git-archive`... I think you could still piggy back on `git archive` by...