{"thread":{"id":"1705","subject":"Notes for CVS Converts","startedAt":"2005-08-31T19:08:36Z","lastAt":"2005-09-21T14:04:04Z","messageCount":2,"participants":["Jon Loeliger","Petr Baudis"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"7979","messageId":"1125515315.13577.53.camel@cashmere.sps.mot.com","threadId":"1705","inReplyTo":null,"subject":"Notes for CVS Converts","fromName":"Jon Loeliger","fromEmail":"jdl@freescale.com","sentAt":"2005-08-31T19:08:36Z","receivedAt":"2005-08-31T19:08:36Z","isPatch":false,"sender":{"key":"jdl@jdl.com","avatar":"https://gravatar.com/avatar/75ce9a10b151acd2c28ec4ab2136dba7b2ff1634530bd04b155981a749d08a64?d=mp&s=160"},"body":"Guys,\n\nI was recently bitten by a cg-rm behavior where\nit silently Did The Wrong Thing, IMO.  I think\nthat other CVS Converts might stumble across the\nsame pitfall, so it might warrant a comment in\nthe documentation, perhaps in the \"Notes For CVS\nUsers\" section.\n\nI am in the process of doing some file merging,\nresulting in some files being eliminated from my\nsource tree.  But I wasn't prepared to just out-right\ndelete my original files until some testing had\nbeen completed first.  So I moved my files out of\nthe way so the build would proceed without them.\n\nLater, when I was satisfied the build was good, I\nexecuted a \"cg-rm floof/somefile.h\" and then cg-commit'ed\nthem.  However, I realized, they were not actually\nremoved from the cache.\n\nReading through \"cg-rm\", I figured out that it expected\nto actually rm the files from the filesystem after first\nproving they existed with a \"find\".  That in turn also\ndrove the \"git-cache-update --force-remove\".\n\nHowever, in my case I had already moved/removed the\noriginal files.  The \"find\" didn't find them and the\ncache update didn't happen.  And no output was generated\neither.  But I thought nothing of it; after all, \"rm\"\ndoesn't produce output when it works successfully.\n\nSo two possible suggestions here:\n\nFirst, perhaps cg-rm should issue a warning if the\ngenerated temporary file results in an empty list of\nfiles to be removed (and then maybe with a --silent\noption subsequently?),\n\nSecondly, maybe the documentation needs a section with\n\"Notes for CVS Converts\" section that clearly states\nthat cg-rm _expects_ to actually remove the file from\nthe filesystem, directly opposite the normal CVS behavior.\n\nThanks,\njdl\n"},{"id":"9038","messageId":"20050921140403.GB17981@pasky.or.cz","threadId":"1705","inReplyTo":"1125515315.13577.53.camel@cashmere.sps.mot.com","subject":"Re: Notes for CVS Converts","fromName":"Petr Baudis","fromEmail":"pasky@suse.cz","sentAt":"2005-09-21T14:04:04Z","receivedAt":"2005-09-21T14:04:04Z","isPatch":false,"sender":{"key":"pasky@ucw.cz","avatar":"https://avatars.githubusercontent.com/u/18439?v=4"},"body":"Dear diary, on Wed, Aug 31, 2005 at 09:08:36PM CEST, I got a letter\nwhere Jon Loeliger <jdl@freescale.com> told me that...\n> Guys,\n> \n> I was recently bitten by a cg-rm behavior where\n> it silently Did The Wrong Thing, IMO.  I think\n> that other CVS Converts might stumble across the\n> same pitfall, so it might warrant a comment in\n> the documentation, perhaps in the \"Notes For CVS\n> Users\" section.\n\nThanks for this, I changed the default to not deleting the files.\n\n> However, in my case I had already moved/removed the\n> original files.  The \"find\" didn't find them and the\n> cache update didn't happen.  And no output was generated\n> either.  But I thought nothing of it; after all, \"rm\"\n> doesn't produce output when it works successfully.\n\nThere was a bug like that, but I believe it was already fixed in the\nrelease available by the time you were writing this mail.\n\n-- \n\t\t\t\tPetr \"Pasky\" Baudis\nStuff: http://pasky.or.cz/\nVI has two modes: the one in which it beeps and the one in which\nit doesn't.\n"}]}