@tolysz I left T.toLower instead of lowering the char in the filter both because of readability and because it seems that Text.toLower handle more cases than map Char.toLower
doc says
O(n) Convert a string to lower case, using simple case conversion. The result string may be longer than the input string. For instance, "İ" (Latin capital letter I with dot above, U+0130) maps to the sequence "i" (Latin small letter i, U+0069) followed by " ̇" (combining dot above, U+0307).

Still not sure it is the way to go, and anyway, build fails so more is to be done anyway.
Leaving as-is

Since you're only dealing with ASCII letters (A-Z), there shouldn't be any difference in behavior. I don't have an opinion on which form looks nicer TBH.

Can I ask you to ping me when you think this is ready to be merged? I'm taking from your comment that you still are going to make a few changes.

Sure, but we already have only [a-zA-Z_] at the time of T.toLower application and those do not benefit from full unicode normalization.

@tolysz that's why I then switched in my second PR to T.map sanitize' . T.toLower .

Anyway, in this third version, I did what you wanted - code is now:

    sanitize = T.map sanitize'
    sanitize' c
        | 'A' <= c && c <= 'Z' = C.toLower c
        | 'a' <= c && c <= 'z' = c
        | '0' <= c && c <= '9' = c
        | otherwise = '_'

@snoyberg it can me berged (build is passing except on 7.4 because iproute-1.4.0 dependency failed to install)

Beautiful :)

Thanks!

Should this same change of lowercasing the app name be applied to the scaffolding, specifically this part" https://github.com/yesodweb/yesod-scaffold/blob/df45cb46ff4c7c1dd9b41ad30ea847f6fef9d4c8/config/settings.yml#L25

Yeah, that makes sense. I'll make that change, there's something slightly (though not significantly) tricky that needs to happen there.

Hi,
I do not know :) but I think it has only isues with generated username/password.
But I checked by changing dbname to rAndOmcAse and it works
edit: just lowercase the user name is enough:
( http://www.postgresql.org/message-id/201004191852.o3JIqiMK012314@wwwmaster.postgresql.org )

Scratch that, @tolysz's right I think, mixed case seems to be just fine.