[Monotone-devel] Re: There's something FISHY with the latest code...

[Bruce Stephens]

Richard Levitte - VMS Whacker <address@hidden> writes:

> It seems like the new code that was merged in last night (swedish
> time) has a weird concept of what the current revision id is *when no
> changes have been made*!
>
> My workspace is currently at ce577c763adf0521c7538255bbbcd4982d21dda3,
> which is visible if you look at _MTN/revision:
>
> -------------------- 8< --------------------
> : ; cat _MTN/revision 
> format_version "1"
>
> new_manifest [933940c810aca3b0edd8651fb5b5848141cac511]
>
> old_revision [ce577c763adf0521c7538255bbbcd4982d21dda3]
> -------------------- 8< --------------------
>
> However, here's the weird thing:
>
> -------------------- 8< --------------------
> : ; mtn automate get_current_revision_id
> d2699bef76a3a8d1aa0b655104d0738d7dbf2b68
> -------------------- 8< --------------------
>
> A little bit of experiment shows what it does:
>
> -------------------- 8< --------------------
> : ; cat _MTN/revision | openssl sha1
> d2699bef76a3a8d1aa0b655104d0738d7dbf2b68
> -------------------- 8< --------------------

This is old behaviour, isn't it?  Revision
38b3ff7851775595777d4804cdf6acf7dcc5e954 (2006-08-23T11:16:04) behaves
similarly:

    brs% cat _MTN/revision
    ce577c763adf0521c7538255bbbcd4982d21dda3

    brs% mtn automate get_current_revision_id
    d2699bef76a3a8d1aa0b655104d0738d7dbf2b68

But you're right that that probably means get_current_revision_id is a
bit wrong: maybe it ought to be get_base_revision_id or something?