1 |
On Sun, 2019-10-20 at 05:21 -0400, Joshua Kinard wrote: |
2 |
> On 10/20/2019 04:32, Michał Górny wrote: |
3 |
> > On Sun, 2019-10-20 at 04:25 -0400, Joshua Kinard wrote: |
4 |
> > > Why is having a max ~24k files in a directory a bad idea? Modern |
5 |
> > > filesystems are more than capable of handling that. |
6 |
> > > |
7 |
> > > - ext4: unlimited files in a directory |
8 |
> > > - xfs: virtually unlimited (hard limit of 2^64-1 total files per volume) |
9 |
> > > - ntfs: 4,294,967,295 |
10 |
> > > |
11 |
> > > And 24k is a bit more than 1/3rd of all distfiles that we currently have. |
12 |
> > |
13 |
> > For the same reason having ~60k files in a directory was a problem. |
14 |
> > There is really no point in changing anything if you change BIG_NUMBER |
15 |
> > to SMALLER_BIG_NUMBER. |
16 |
> |
17 |
> That doesn't answer my question. Why is it a problem? What criteria are |
18 |
> you using to decide that 24k is a "smaller big number"? Is there some issue |
19 |
> highlighted by the mirror admins where having 24k files in a single |
20 |
> directory offers no significant relief versus the current 60k files? |
21 |
|
22 |
IIRC Robin set the goal as: |
23 |
|
24 |
| the number of files in a single directory should not exceed 1000, [1] |
25 |
|
26 |
I don't recall how that number was chosen but it's probably pretty |
27 |
arbitrary. In any case, I can notice the difference between working |
28 |
with a listing of 1k files and 24k files, on the hardware running |
29 |
masterdist. |
30 |
|
31 |
> > > Under which scenario do you wind up with 24k files in a single directory? I |
32 |
> > > consider the tex package an outlier in this case (one package should not be |
33 |
> > > the sole dictator of policy). |
34 |
> > |
35 |
> > Three versions of TeXLive living simultaneously. If one package falls |
36 |
> > completely out of bounds, no problem is solved by the change, so what's |
37 |
> > the point of making it? |
38 |
> |
39 |
> The problem in this case is with texlive, not our current, or future, |
40 |
> distfiles methodology. |
41 |
|
42 |
Is it? Are you suggesting we should ban upstream from using multiple |
43 |
distfiles with similar prefix? What about other potential packages that |
44 |
may suffer from the same problem in the future? Go packages have a good |
45 |
potential, given that majority of them starts with 'github.com'. |
46 |
|
47 |
> Has anyone looked at how other distros deal with texlive? |
48 |
|
49 |
Other distros don't mirror original distfiles. |
50 |
|
51 |
> Has anyone complained or filed a bug to texlive developers |
52 |
> upstream about their excessive amount of distfiles and the burden it places |
53 |
> on distro maintainers? |
54 |
|
55 |
You believe it to be a problem. Don't expect others to bother upstream |
56 |
with your preferences. |
57 |
|
58 |
|
59 |
[1] https://www.gentoo.org/glep/glep-0075.html#algorithm-for-splitting-distfiles |
60 |
|
61 |
> |
62 |
-- |
63 |
Best regards, |
64 |
Michał Górny |