Am Donnerstag, 20. September 2012 um 10:14:56, schrieb T. Modes <[email protected]> > Hi Kornel, > > > Since nobody answered, I committed the change in the default branch. To my > > surprise Thomas Modes > > also already committed to branch gui_overhaul. > > > > I committed the changeset to the default branch (and forgot to leave a > note) - not to gui_overhaul branch. The gui_overhaul branch was close > some time ago, all changes to it are gone into the default branch. > > > Now, there is a problem. After updating to branch gui_overhaul, the changes > > (from Thomas _and_ me) are lost. > > Sure I don't understand how hg acts here. (I suspect the use of "hg push > > -f" causes this.) > > > > Normally, there should no need to use the -f switch. It's only needed > when you create a new branch. In all other cases the warning is > justified. Then you should check your local mercurial repo and often > it is necessary to do a merge. > > > So, sorry to causing troubles. Could someone point me to a reasonable > > handling as how to > > correctly resolve such conflicts? > > You had an old and unpushed changeset to the gui_overhaul branch in > your local repo. This created a new head because the development in > the gui_overhaul branch continued. That caused the warning you got. > > For a check of the outgoing changesets you could use "hg outgoing". > For me a GUI helps to keep track of the changes. Also the web > interface can be helpful to keep the overview (e.g. the graph): > http://hugin.hg.sourceforge.net/hgweb/hugin/hugin/graph > > Thomas > > PS: The backed out your change to the default branch. The variable > maxvalue is from type long. There is no need to explicitly cast it to > long. The changes to the format string is sufficient.
Yes Thomas, I knew. But my experience is that cast does not hurt, and, in case
of change maxvalue-type, stays safe. But I'm fine with your view.
Kornel
signature.asc
Description: This is a digitally signed message part.
