This is only an example: net-libs/gnutls:0 (net-libs/gnutls-3.5.15:0/30::gentoo, ebuild scheduled for merge) conflicts with net-libs/gnutls:=[abi_x86_32(-),abi_x86_64(-)] required by (app-emulation/wine-vanilla-3.2:3.2/3.2::gentoo, installed) >=net-libs/gnutls-2.12.23-r6[abi_x86_32(-),abi_x86_64(-)] required by (net-print/cups-2.1.4:0/0::gentoo, installed) net-libs/gnutls:0/30=[abi_x86_32(-),abi_x86_64(-)] required by (app-emulation/wine-vanilla-3.2:3.2/3.2::gentoo, installed) Most new users will have no clue what underlying error is, this isn't the only example of portage being overly symbolic in its error displaying. Use Flag constraints and circular dependencies are also a problem. Portage should be a little bit more descriptive. Great way to chase away new users.
Let me add that there is absolutely no online documentation detailing all of portages possible output errors.
(In reply to Zakhar from comment #0) > > Portage should be a little bit more descriptive. Normally emerge produces some more output that you didn't post, maybe it was suppressed by the -q/--quiet option? If we're going to constructively criticize the output, then we need to see all of it so that we can properly evaluate it for deficiencies.
+ The interface (emerge) is inconsistent and awful. Moreover, the documentation is not accurate. # Undefined symbols. For instance, the user does know the meaning of square brackets, parentheses and dash: dev-python/setuptools[python_targets_pypy(-)?...] These information add no value for a user. Overview: [1]. -------------------------------------------------------- -> [1] https://forums.gentoo.org/viewtopic-p-8205918.html#8205834 # Unsuitable information. -> [2] https://forums.gentoo.org/viewtopic-p-8194844.html#8194844 -> [3] https://forums.gentoo.org/viewtopic-p-8205560.html#8205560 1. Here, the user partially apprehend the issue (look at (-pypy)). -> [4] https://forums.gentoo.org/viewtopic-p-8198380.html#8198380 # Symbols having different meanings (e.g.: U): 1. Upgrade a package (new version available). [ebuild U ] dev-libs/libassuan-2.5.1::gentoo [2.4.3-r1::gentoo] USE="-static-libs" 552 KiB [ebuild r ] ... 2. Request to search packages for which USE flags changed since installation. user $ emerge -pvUu @world # Undefined behaviors (divergent, implied or underlied) following requests. 1. For example, --tree option should not be used certain circumstances: although mentioned in the man page (see attachments: rust and vim). 2. Another example, there are things superficially similar but really a little bit different. -> [5] https://forums.gentoo.org/viewtopic-p-8190798.html#8190798 3. Here, emerge should warns about possible breaking(s). -> [6] https://forums.gentoo.org/viewtopic-p-8205558.html#8205558 CONCLUSION: I suppose the mentioned things cause unpleasantness!!
Created attachment 526600 [details] "emerge -tpv vim"
Created attachment 526602 [details] "emerge -tpv rust"
I misinterpreted the --tree option therefore you may ignore my comment about that.
(In reply to Kevin Vigouroux from comment #3) > 3. Here, emerge should warns about possible breaking(s). > > -> [6] https://forums.gentoo.org/viewtopic-p-8205558.html#8205558 > > CONCLUSION: I suppose the mentioned things cause unpleasantness!! I've replied in the above forum thread here, explaining how the slot operator is intended to behave: https://forums.gentoo.org/viewtopic-p-8206050.html#8206050 I'll follow that thread to find out what's going on. If there turns out to be a portage problem there then we should handle it with a separate bug report.
