|
From: | Dmitry Gutov |
Subject: | bug#24082: 25.1; vc-dir for CVS repositories list all files as if from toplevel directory |
Date: | Sat, 8 Oct 2016 23:06:10 +0300 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:49.0) Gecko/20100101 Thunderbird/49.0 |
On 08.10.2016 18:13, Jérémie Courrèges-Anglas wrote:
Oops, sorry. Yes indeed, I only replaced `files' with `dir'.
OK, good. Any obvious problems with that solution? I don't work with CVS at all, and if you (or anyone else) interact with any particularly big repositories, I'd love to hear how it fares there.
and whether it works well with all CVS repositories you work with.Works fine, there are some differences.
That's too bad. I've tried the patch myself, and I didn't see those, neither with the temporary repository created using Göktuğ's instructions, nor with a random GNU CVS repo I cloned afterwards.
Might there be something special with the repo you're trying it on? Maybe the fact that the working dir and the repository root only have the root directory as the common ancestor? Just a random guess.
So the patched version shows unregistered files and directories, which is good, but has formatting artifacts (`.//' above). Also, pressing `=' on `.//emacs-wip', I get the following in *Messages*: ... Traversing directory /usr/ports/editors/...done Finding changes in /usr/ports/editors/emacs-wip... vc-find-backend-function: Cannot open load file: No such file or directory, vc-nil even though the directory contains a CVS dir, and cvs(1) works fine in it. I don't think it's a big problem though.
Agree, that doesn't sound like too big of a problem.And seeing unregistered files is pretty much working as intended. Alas, I see them in both versions of the code (the one using `dir', and the one using 'cvs update').
Any ideas why you're seeing differently? Do you have a non-default value of vc-cvs-stay-local, maybe?
[Prev in Thread] | Current Thread | [Next in Thread] |