Re: RFC: New Response type to aid rlog reponse parsing

[Thorsten Glaser]

Hi patrick,

>Sorry for the delayed reply (project I'm working on is taking
>far too many cycles).

don’t worry, I know all about that.

>Just from reading the diff it looks OK. However, I didn't comprehend
>the sending "partial lines to the client" bit.

There was some comment in cvs_output() about being careful to
only write to the output stream when a full line has been
accumulated.

>> I'm currently sanity.sh-testing it, this will take the night.
>
>Hopefully your tests don't show any regression.

It didn’t. I also manually tested the server by interacting
with 'cvs server' from the commandline, and it looked fine.

>I can patch my CVS server with your diff when I break away from
>my work.

Mind that this diff is based on continued development from
the 1.12.13 release, not the ancient 1.11.x branch, though.

>Thanks for the update!

You’re welcome!

>Happy Thanksgiving!

Erm, not here (Europe), but thanks anyway, and same to you.

>p.s., sending from a different account (not subscribed to CVS list)
>hopefully it'll make it.

It made it, both over the list and separately.

bye,
//mirabilos
-- 
Support mksh as /bin/sh and RoQA dash NOW!
‣ src:bash (358 (380) bugs: 1 RC, 246 (262) I&N, 111 (117) M&W, 0 F&P)
‣ src:dash (110 (129) bugs: 0 RC, 61 (67) I&N, 49 (62) M&W, 0 F&P)
‣ src:mksh (0 bugs: 0 RC, 0 I&N, 0 M&W, 0 F&P)
dash has two RC bugs they just closed because they don’t care about quality…