summaryrefslogtreecommitdiff
path: root/open_issues/libnetfs_vs_libdiskfs.mdwn
blob: 2fcfbea5cb5b482867eff82ecc47a07482f9f4bf (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
[[!meta copyright="Copyright © 2013 Free Software Foundation, Inc."]]

[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable
id="license" text="Permission is granted to copy, distribute and/or modify this
document under the terms of the GNU Free Documentation License, Version 1.2 or
any later version published by the Free Software Foundation; with no Invariant
Sections, no Front-Cover Texts, and no Back-Cover Texts.  A copy of the license
is included in the section entitled [[GNU Free Documentation
License|/fdl]]."]]"""]]

[[!tag open_issue_hurd]]

[[!toc]]


# Argument Parsing

## IRC, freenode, #hurd, 2013-06-27

    <teythoon> the arg parsing in libdiskfs and libnetfs differ :/
    <teythoon> afaics libdiskfs gets it right, libnetfs does not
    <pinotree> what do you mean?
    <teythoon> wrt to *_std_{runtime,startup}_argp
    <teythoon> see eg netfs.h
    <teythoon> libdiskfs/opts-std-runtime.c:const struct argp
      diskfs_std_runtime_argp =
    <teythoon> libdiskfs/opts-std-runtime.c-{
    <teythoon> libdiskfs/opts-std-runtime.c-  std_runtime_options, parse_opt,
      0, 0, children
    <teythoon> libdiskfs/opts-std-runtime.c-};
    <teythoon> but
    <teythoon> libnetfs/std-runtime-argp.c:const struct argp
      netfs_std_runtime_argp = { 0 };
    <pinotree> well there are no common startup/runtime options provided by
      netfs
    <pinotree> usually netfs-based translators put netfs_std_startup_argp as
      child as their options, so if netfs starts providing options they would
      work
    <teythoon> ah
    <pinotree> if you have a test showing issues, we can certainly look it :)
    <teythoon> ok, m/b I was confused...
    <pinotree> no worries, feel free to ask anytime
    <teythoon> I thought about providing --update as common runtime flag, like
      diskfs does, you think it's the right thing to do?
    <pinotree> what would it do?
    <teythoon> or should it be left for each translator to implement?
    <teythoon> nothing by default I guess
    <pinotree> options provided in libdiskfs are implemented and handled mostly
      in libdiskfs itself
    <pinotree> so imho a new option for libnetfs would be there because its
      behaviour is implemented mostly within libnetfs itself
    <teythoon> libdiskfs calls diskfs_reload_global_state
    <teythoon> libnetfs could do the same, allowing translators to plug in
      anything they wish
    <teythoon> but I'll implement it in procfs for the time being
    <pinotree> ah, its alias is remount
    <teythoon> yes
    <teythoon> I need that working for procfs
    <teythoon> btw, I think I got your mount confusion thing figured out
    <pinotree> but procfs has nothing to update/flush, all the information are
      fetched at every rpc
    <teythoon> yes
    <teythoon> but we still need to ignore the flag
    <teythoon> otherwise the set_options rpc fails
    <teythoon> http://paste.debian.net/12938/
    <teythoon> whee, remounting proc works :)
    <braunr> :)


# IRC, freenode, #hurd, 2013-07-29

    <teythoon> so, what do you folks think about refactoring libdiskfs and
      libnetfs to be more alike?
    <pinotree> what do you mean?
    <teythoon> ah, I mentioned that in the context of my mtab prototype
      1374247519-26589-1-git-send-email-4winter@informatik.uni-hamburg.de
    <teythoon> they are hard to diff against each other b/c they differ in file
      names and identifier names
    <teythoon> while working on the mtab stuff I encountered stuff that was
      implemented in libdiskfs, but never in libnetfs
    <teythoon> mostly support for binding translators to libnetfs nodes
    <braunr> teythoon: sure, but looks a little out of scope
    <teythoon> braunr: I do not mean now, more in general
    <braunr> ok
    <tschwinge> teythoon: I wondered about this, too.  I don't know if it's
      possible to literally merge them (and build the backend-based (libdiskfs)
      vs. volatile-backend one (libnetfs) based on a pre-processor define or
      similar), or just structure the source code (files) in a way such that
      »diff -ru libdiskfs/ libnetfs/« gives meaningful results, figuratively
      spoken.
    <teythoon> tschwinge: my thoughts exactly


# IRC, freenode, #hurd, 2013-08-28

    <teythoon> braunr: do you think another lib*fs would be frowned uppon? I
      like the way procfs is structured and that could be refactored and
      generalized into a library
    <braunr> i think we need more lib*fs libraries
    <braunr> and better integration
    <braunr> that's one of the strengths in linux
    <braunr> it makes writing file systems very easy
    <teythoon> cool :)
    <teythoon> now we only need a snappy name, any suggestions?
    <braunr> i don't know what you like specificlaly in procfs
    <braunr> libpseudofs ?
    <teythoon> well, it's not perfect, but i like the way you just have to
      implement a function for a node and it magically gains the ability to
      being read
    <teythoon> for example
    <braunr> yes i see
    <pinotree> lacks a bit of caching though
    <braunr> no caching for such file systems
    <teythoon> indeed
    <braunr> why would you want caching ?
    <pinotree> you might have files that don't change at all, or rarely do
    <braunr> the premise is that it's meant for files generated on the fly
    <braunr> but are they big ?