Quoth Stuart Morrow <morrow.stuart@gmail.com>:
> proof: don't confuse ""(1)
>
> diff -r d072b459519d sys/src/cmd/proof/htroff.c
> --- a/sys/src/cmd/proof/htroff.c Thu Nov 26 11:18:41 2020 -0800
> +++ b /sys/src/cmd/proof/htroff.c Thu Nov 26 22:40:31 2020 +0000
> @@ -550,7 +550,7 @@
> continue;
> }
>
> - fprint(2, "illegal; try q, 17, +2, -1, p, m.7, /2, x1, y-.5 or return\n");
> + fprint(2, "illegal -- try q, 17, +2, -1, p, m.7, /2, x1, y-.5 or return\n");
> }
> return 0;
> }
>
Will look at the page patch later, but for this one,
I'll commit with one change: '--' to ':' should avoid
confusing "" and looks better to my eyes.
Thanks for pointing it out.
Probably the reason for the semicolon was it's actually a sentence that way. So I followed that line of thought with my '--'. Anyway, I was in proof (quite a few days ago now, sorry for delay) making it consistent with what I did to page (don't know why I didn't think to do that earlier, when the whole point of the page patch was consistency) and now I think we should just delete proof, at least provided nobody claims to use it. The whole thing is bugs. Easy to fix, but still. It doesn't do anything that page doesn't do better. Here's what would need to change: man (delete -p or alias it to -P) man(1) troff(1) rio(1), which inexplicably references proof(1). Executables stick around after the source is gone (I still have ape/mv|cp), so I would want the deletion of cmd/proof to be on a time-delay, so that that one last bug fix ('illegal;') can mk install for everyone. Stuart
Quoth Stuart Morrow <morrow.stuart@gmail.com>:
> Probably the reason for the semicolon was it's actually a sentence
> that way. So I followed that line of thought with my '--'.
>
> Anyway, I was in proof (quite a few days ago now, sorry for delay)
> making it consistent with what I did to page (don't know why I didn't
> think to do that earlier, when the whole point of the page patch was
> consistency) and now I think we should just delete proof, at least
> provided nobody claims to use it. The whole thing is bugs. Easy to
> fix, but still. It doesn't do anything that page doesn't do better.
>
> Here's what would need to change:
>
> man (delete -p or alias it to -P)
> man(1)
> troff(1)
> rio(1), which inexplicably references proof(1).
>
> Executables stick around after the source is gone (I still have
> ape/mv|cp), so I would want the deletion of cmd/proof to be on a
> time-delay, so that that one last bug fix ('illegal;') can mk install
> for everyone.
>
> Stuart
>
I'm ok with this, as long as nobody else objects.
On December 5, 2020 7:19:52 PM EST, ori@eigenstate.org wrote:
>Quoth Stuart Morrow <morrow.stuart@gmail.com>:
>> Probably the reason for the semicolon was it's actually a sentence
>> that way. So I followed that line of thought with my '--'.
>>
>> Anyway, I was in proof (quite a few days ago now, sorry for delay)
>> making it consistent with what I did to page (don't know why I didn't
>> think to do that earlier, when the whole point of the page patch was
>> consistency) and now I think we should just delete proof, at least
>> provided nobody claims to use it. The whole thing is bugs. Easy to
>> fix, but still. It doesn't do anything that page doesn't do better.
>>
>> Here's what would need to change:
>>
>> man (delete -p or alias it to -P)
>> man(1)
>> troff(1)
>> rio(1), which inexplicably references proof(1).
>>
>> Executables stick around after the source is gone (I still have
>> ape/mv|cp), so I would want the deletion of cmd/proof to be on a
>> time-delay, so that that one last bug fix ('illegal;') can mk install
>> for everyone.
>>
>> Stuart
>>
>
>I'm ok with this, as long as nobody else objects.
i don't use proof at all.
sl