asciilifeform: i suspect it's at least partially meat-powered shop : 'Please be advised that a sales (Account manager) will reach out to you at their earliest convenience .'
asciilifeform: latest update -- they ate asciilifeform's 17bux, but did not disgorge a form for choosing the # yet
asciilifeform: 'One of our representatives will review your order and enable your account shortly'
asciilifeform: aite i'ma get the slightly cheaper-per one, we can switch laters
asciilifeform: mircea_popescu: how much traffic do you expect on the 800..? there's variant 14.49/0.07min and 17.99/0.066min where latter includes 223/mo
asciilifeform: ( aside from rampant brokenness, in the years since gentoo author took 30 silvers from microshit and fucked it all ) would routinely get into circular wedges and fail in other 'interesting' ways
asciilifeform: and it dun even break the ancient format.
asciilifeform: file names are covered by the gpg seal of given patch, tho, so it isn't as if people can get away with blindly renaming items in a patch. so if taking all of mircea_popescu's algo but the hashed-names part, you have a usable algo.
asciilifeform: ( could no longer identify moved files automatically )
asciilifeform: it is my understanding tho that it is impossible to apply the algo in the thread if one were to make file name part of file's hash.
asciilifeform: phf: i think errything you need, is in there.
asciilifeform: ( in thread, mircea_popescu detailed how to make it go )
asciilifeform: i actually posted an experimental vtron that hashed the names
asciilifeform: mircea_popescu: if he's stuck to the old gnudiff format, which per my test appears to be the case, names aint hashed at all ( they're covered by the gpg sig of given patch, but that's it )
asciilifeform: mircea_popescu: i'm mapping out the size of the hole, rather than dancing on mircea_popescu's toes for phun; have patience, if i've found this earlier, yer toes would not hurt nao.
asciilifeform: phf then you dun need bin diff/create.
asciilifeform: possibly, i'ma have to rewalk the pertinent l0gz when i get a chance
asciilifeform: but nobody liked this, seems like errybody is addicted to lines.
asciilifeform: phf: recall, this is why asciilifeform earlier wanted a non-linecentric diff representation. then would not need special cases for clean ascii & bins.
asciilifeform: ( creations & diffs for bins, specifically )
asciilifeform: phf: creations & diffs do not presently have a standard format; if mircea_popescu says that we need it, will have to make one.
asciilifeform: also while mircea_popescu is awake : if you have the dest # for the 1-800 thing, plox to gpggram, i'ma bake it ( theoretically oughta be live in day or 2 from revv-up )☟︎
asciilifeform: ( if only deletions, than the file renameage thing will suffice )
asciilifeform: mircea_popescu: do you have bins in eulora tree ?
asciilifeform: phf: do ya think you can post the binaries-eating version soon-ish ? i'd like to regrind FG, for example.
asciilifeform: ( i suggested several ways of doing this )
asciilifeform: phf: if binariola handler is added carefully, then yes, should not break existing patches
asciilifeform: diana_coman evidently did not test this ( not really criticism of diana_coman , iirc we did decide that republican proggies ought not to include raw binariolade )
asciilifeform: then diana_coman & mircea_popescu say 'wtf your waiting for' so thought 'so it's done...'
asciilifeform: and i do recall that mircea_popescu specced this feature. but seems like it aint there yet.
asciilifeform: mircea_popescu: sadly enuff, i was right in re the current phf-vdiff . observe : a) http://p.bvulpes.com/pastes/5ZYgF/?raw=true << example of deleting a text file. in fact uses same deletion representation as ordinary gnudiff, where entire payload is quoted. b) http://p.bvulpes.com/pastes/vl2Ca/?raw=true << attempt to represent deletion of 1MB from FG. result : same as in oldschool gnudiff.☟︎
