On Wed, Aug 22, 2012 at 5:54 AM, sebb <seb...@gmail.com> wrote: > On 22 August 2012 07:05, Andrea Pescetti <pesce...@apache.org> wrote: >> sebb wrote: >> >>> On 21 August 2012 22:23, Daniel Shahaf wrote: >>>> >>>> Why? Could very well have a "publish all changes" mode (the current >>>> only option) alongside the "cherry picking" (publish only selected file) >>>> mode. >>> >>> This thread started because something was published inadvertently. >> >> >> No, this thread started because people (including me) who did want to >> publish their changes only, and who routinely used the "Diff" command in the >> CMS to examine what would be published, were forced to take the "all or >> nothing" approach (you will find me noting this in the log messages too). > > I was going by the following comments: > >>> >> >> Sorry but IMHO this process failed. Just today evening (Hamburg >>> >> >> time) someone has published again website changes. >>> >> >> >>> >> >> If we rely on a process that is so fragile, then IMHO we shouldn't >>> >> >> do this. Because there will be always somebody: >
I think either solution would work for us; A) Ability to lock publication either of the entire /ooo-site or at the level of a specific subdir. But I suppose from perspective of Infra, everything is a subdir. B) Ability in CMS for user to check off which files are actually published, e.g., which subset. If it can default to select the files that were modified in the current session, that is even better. >> >>> If the purpose of the enhancement is to prevent this happening, then >>> there needs to be some barrier that prevents inadvertent publication. >> >> >> This would be an extra guarantee, but probably an unnecessary complication >> for what we have seen so far. Ability to publish one directory only would >> already help a lot. >> >> Regards, >> Andrea.
