what exactly is going on here?

yeti

On system "destiny":

(yeti@destiny:1)~/wrk/propeller/propgcc-hg/propgcc$ date
Wed Jun 11 08:47:53 CEST 2014
(yeti@destiny:1)~/wrk/propeller/propgcc-hg/propgcc$ hg pull --update
pulling from http://code.google.com/p/propgcc/
searching for changes
no changes found
(yeti@destiny:1)~/wrk/propeller/propgcc-hg/propgcc$ hg branches | awk -vb=$(hg branch) '{ print ($1==b?"--> ":"    ") $0 }'
--> default                     2408:9fa9279ceaba
    p2update                    2251:029074e5e0d2
    release_1_0                 2240:a81b73037fef
    binutils-base                  4:7d2695d7a850

On system "darkstar":

(yeti@darkstar:3)~/wrk/propeller/propgcc-hg/propgcc$ date
Wed Jun 11 08:48:26 CEST 2014
(yeti@darkstar:3)~/wrk/propeller/propgcc-hg/propgcc$ hg pull --update
pulling from http://code.google.com/p/propgcc/
searching for changes
no changes found
(yeti@darkstar:3)~/wrk/propeller/propgcc-hg/propgcc$ hg branches | awk -vb=$(hg branch) '{ print ($1==b?"--> ":"    ") $0 }'
--> default                     2408:9fa9279ceaba
    p2update                    2251:029074e5e0d2
    release_1_0                 1826:a81b73037fef
    binutils-base                  4:7d2695d7a850

Please explain why branch release_1_0 is so much different on both systems.

Additionally, when I try to "hg clone" the repository from scratch again, mercurial tries to get revision 2409 but falls stale after a while. I didn't get a new clone of the repository in more then 5 tries...

Am I confused or is it mercurial/google I should stop to trust?

jazzed

Sometimes googlecode has trouble with "hg clone" it will eventually succeed.

To get a fresh release_1_0 repository, you need to do "hg pull; hg update release_1_0"

yeti

I know how to switch branches.
This is not my question.

Again:

How can 'hg pull' succeed on both systems and yield different revision numbers for release_1_0?

And why does 'hg pull' succeed telling me r2408 is current but trying to 'hg clone' from scratch tells me that r2409 is the current revision number?

That is not suited to build trust in mercurial and/or google-code...

ersmith

yeti wrote: »

How can 'hg pull' succeed on both systems and yield different revision numbers for release_1_0?

And why does 'hg pull' succeed telling me r2408 is current but trying to 'hg clone' from scratch tells me that r2409 is the current revision number?

The Mercurial "revision number" is a local thing, and may be different between different systems. The real identifier for a version is the hash (the hex string that appears after the revision number): as long as that's the same the revisions are the same. I think the Mercurial manual discusses this.

WBA Consulting

It does vary by machine, http://mercurial.selenic.com/wiki/RevisionNumber pitfalls listed on that page:

Pitfalls

Revision numbers referring to changesets are very likely to be different in another copy of a repository. Do not use them to talk about changesets with other people. Use the changeset ID instead.

yeti

Because "gcc -v" contains the release number I always thougt they would be globally unique and meaningfull...

$ /opt/propgcc-1.0.0-2250+/bin/propeller-elf-gcc -v          
(((some lines removed)))
gcc version 4.6.1 (propellergcc_v1_0_0_2250+)
$ /opt/propgcc-1.9.0-2408+/bin/propeller-elf-gcc -v
(((some lines removed)))
gcc version 4.6.1 (propellergcc-alpha_v1_9_0_2408+)

... seems I was too credulous, optimistic or whatever.

Using a changeset hash for telling which of two versions is the younger one definitely is ugly.

Perhaps the changeset timestamp can replace the current use of revision numbers. I need to look closer at mercurial (propgcc is the only contact I have to "hg").