1 |
On Fri, Sep 30, 2022 at 12:49:02PM -0700, Alec Warner wrote: |
2 |
> On Fri, Sep 30, 2022 at 7:53 AM Florian Schmaus <flow@g.o> wrote: |
3 |
> > |
4 |
> > On 30/09/2022 02.36, William Hubbs wrote: |
5 |
> > > On Wed, Sep 28, 2022 at 06:31:39PM +0200, Ulrich Mueller wrote: |
6 |
> > >>>>>>> On Wed, 28 Sep 2022, Florian Schmaus wrote: |
7 |
> > >>> 2.) the number of EGO_SUM entries exceeds 1000 and a Gentoo developer |
8 |
> > >>> maintains the package |
9 |
> > >>> 3.) the number of EGO_SUM entries exceeds 1500 and a proxied |
10 |
> > >>> maintainer maintains the package |
11 |
> > >> |
12 |
> > >> These numbers seem quite large, compared to the mean number of 3.4 |
13 |
> > >> distfiles for packages in the Gentoo repository. (The median and the |
14 |
> > >> 99-percentile are 1 and 22, respectively.) |
15 |
> > |
16 |
> > The numbers may appear large when compared to the whole tree, but I |
17 |
> > think a fair comparison would be within the related programming language |
18 |
> > ecosystem, e.g., Golang or Rust. |
19 |
> > |
20 |
> > For example, analyzing ::gentoo yields the following histogram for |
21 |
> > 2022-01-01: |
22 |
> > https://dev.gentoo.org/~flow/ego_sum_entries_histogram-2020-01-01.png |
23 |
> > |
24 |
> > |
25 |
> > > To stay with your example, restic has a 300k manifest, multiple 30k+ |
26 |
> > > ebuilds and897 distfiles. |
27 |
> > > |
28 |
> > > I'm thinking the limit would have to be much lower. Say, around 256 |
29 |
> > > entries in EGO_SUM_SRC_URI. |
30 |
> > |
31 |
> > A limit of 256 appears to be to low to be of any use. It is slightly |
32 |
> > above the 50th percentile, half of the packages could not use it. |
33 |
> > |
34 |
> > We have to realize that programming language ecosystems that only build |
35 |
> > static binaries tend to produce software projects that have a large |
36 |
> > number of dependencies. For example, app-misc/broot, a tool written in |
37 |
> > Rust, has currently 310 entries in its Manifest. Why should we threat |
38 |
> > one programming language different from another? Will be see voices that |
39 |
> > ask for banning Rust packages in ::gentoo in the future? With the rising |
40 |
> > popularity of Golang and Rust, we will (hopefully) only ever see an |
41 |
> > increase of such packages in ::gentoo. And most existing packages in |
42 |
> > this category will at best keep their dependency count constant, but are |
43 |
> > also likely to accumulate further dependencies over time. |
44 |
> > |
45 |
> > And quite frankly, I don't see a problem with "large" Manifests and/or |
46 |
> > ebuilds. Yes, it means our FTPs are hosting many files, in some cases |
47 |
> > even many small files. And yes, it means that in some cases ebuild |
48 |
> > parsing takes a bit longer. But I spoke with a few developers in the |
49 |
> > past few months and was not presented with any real world issues that |
50 |
> > EGO_SUM caused. If someone wants to fill in here, then now is a good |
51 |
> > time to speak up. But my impression is that the arguments against |
52 |
> > EGO_SUM are mostly of cosmetic nature. Again, please correct me if I am |
53 |
> > wrong. |
54 |
> |
55 |
> I thought the problem was that EGO_SUM ends up in SRC_URI, which ends |
56 |
> up in A. A ends up in the environment, and then exec() fails with |
57 |
> E2BIG because there is an imposed limit on environment variables (and |
58 |
> also command line argument length.) |
59 |
> |
60 |
> Did this get fixed? |
61 |
> |
62 |
> https://bugs.gentoo.org/719202 |
63 |
|
64 |
You are correct this was part of the issue as well. I don't know what |
65 |
the status of this bug is. |
66 |
|
67 |
William |