Add a version parsing and comparison API to librpmio #1221
Conversation
No functional change, but we'll need this in the next step.
Adding a new header just for this seems a bit much but we'll be adding stuff there shortly. No functional changes as such, this is prerequisite for supporting version comparison in expressions.
No functional changes, just lose extra baggage that was needed to append initialize and append rpm.vercmp() from librpm side when the rest is in librpmio. We might need to bring it back some day, but given the history here it doesn't seem all that likely.
Epoch promotion was a thing around and before the turn of the millenium, we really shouldn't be carrying that cruft around anymore. Remove all traces that we can without too much guilt about breaking ABI/API and not bumping soname which we dont want to do for this stupid old thing: all the symbols are left in place, they just don't work anymore. Nobody should notice as nobody can have been using this stuff in the last 15+ years.
|
Rebase to clear fixups. |
|
Removing the nopromote API in Python but keeping it in C seems inconsistent. Not sure if I really care. |
|
Overall this looks good. I am wondering what all the malloc and free costs us as we might do a few version comparisons during a transaction. Probably not enough to actually worry. Converting the EVR of the rpmds object just for comparison seems kinda weird. I am fine with this as a first step but we should consider moving the rpmds to use versions from the start. Which asks the question how this relates to rpmsid... |
|
This doesn't actually add any mallocs that were not there before for the busy case of rpmds comparisons: previously the strings were strdup()'ed in rpmdsCompareEVR() before passing to parseEVR(), now this is happens in rpmevr.c instead. The ver handle is all alloced as a single blob so there's no added cost (except for some error checking). As for rpmds and string pool, all good questions that occurred to me too, but I don't have immediate answers. It'd seem attractive to run the rpmds versions through this just for enforced error checking, but then those are in the strpool, and the pool has its own non-trivial cost, and the private pool trick we use elsewhere seems way out of proportion for storing a couple of tiny strings. My excuse for ripping epoch promotion from Python but leaving the ruins in C is basically just that in C, removing would require soname bump, but Python has nothing of the sort. I can certainly put it back there if you prefer. BTW, last fixup added support for python (e,v,r) tuples as a way of creating version objects, I initially pushed the python bindigs a little bit too hastily :) |
|
Another fixup to avoid multiple reallocations in rpmverEVR(). |
It's more than a little hysterical that rpm hasn't had a meaningful public API for parsing and comparing version strings. The API may seem kinda overkill for some things but then this does give us a place to check for invalid version strings, which is a thing we never had before (although the current code doesn't do much in the way of checking) Fixes: rpm-software-management#561
…software-management#897) The low-level rpmvercmp() was always the wrong thing to export to Lua, but its what people have used because it's there and its used for comparing full versions. So rather than add a separate API for the higher level comparison, lets just change rpm.vercmp() do the right thing instead. Fixes: rpm-software-management#897
Use the newly added version converter function for parsing labelCompare() arguments, gaining automatic access to all formats that we support in rpm.ver() constructor. Currently this means (E,V,R) tuples which labelCompare() always used, plus plain old strings. In future, might be something more.
|
Rebased once more with some tweaks to gain support strings in Python labelCompare() essentially for free things like |
|
Just to make sure, this behaviour change of Currently we're doing the parsing "by hand" in lua: https://build.opensuse.org/package/view_file/Base:System/rpm/rpmsort?expand=1 I'm really looking forward to drop that for good. |
|
It should be compatible for legit values of E, V, R. I'm quite positive you can come up with strange behavior differences if you feed it versions containing ':' or '-' though. |
This adds a (beginnings of) a version parsing and comparison API to librpmio.
Details in the commit messages, but the (obvious) motivation is to finally have a meaningful API for handling rpm versions in and out of rpm. For maximum coverage, this needs to be in librpmio which allows usage from Lua, macro engine and expression parser.
--nopromote support is ripped out to the extent possible without requiring soname bump to avoid dragging that ancient remnant into the new API.
Fixes #561 and #897 and paves the way for #1217.