| 1 |
On 04/01/2018 06:54 AM, Michał Górny wrote: |
| 2 |
> W dniu czw, 29.03.2018 o godzinie 15∶34 -0700, użytkownik Zac Medico |
| 3 |
> napisał: |
| 4 |
>> For binary packages, honor the INSTALL_MASK configuration that |
| 5 |
>> exists at install time, since it might differ from the build time |
| 6 |
>> setting. |
| 7 |
>> |
| 8 |
>> Fixes: 3416876c0ee7 ("{,PKG_}INSTALL_MASK: python implementation") |
| 9 |
>> Bug: https://bugs.gentoo.org/651952 |
| 10 |
>> --- |
| 11 |
>> bin/misc-functions.sh | 23 +++++++++++++++++++++++ |
| 12 |
>> bin/phase-functions.sh | 10 +--------- |
| 13 |
>> pym/portage/dbapi/vartree.py | 5 +++++ |
| 14 |
>> 3 files changed, 29 insertions(+), 9 deletions(-) |
| 15 |
>> |
| 16 |
>> diff --git a/bin/misc-functions.sh b/bin/misc-functions.sh |
| 17 |
>> index 26f589915..a6330ee93 100755 |
| 18 |
>> --- a/bin/misc-functions.sh |
| 19 |
>> +++ b/bin/misc-functions.sh |
| 20 |
>> @@ -323,6 +323,29 @@ postinst_qa_check() { |
| 21 |
>> done < <(printf "%s\0" "${qa_checks[@]}" | LC_ALL=C sort -u -z) |
| 22 |
>> } |
| 23 |
>> |
| 24 |
>> +preinst_mask() { |
| 25 |
>> + # Remove man pages, info pages, docs if requested. This is |
| 26 |
>> + # implemented in bash in order to respect INSTALL_MASK settings |
| 27 |
>> + # from bashrc. |
| 28 |
>> + local f x |
| 29 |
>> + for f in man info doc; do |
| 30 |
>> + if has no${f} ${FEATURES}; then |
| 31 |
>> + INSTALL_MASK+=" /usr/share/${f}" |
| 32 |
>> + fi |
| 33 |
>> + done |
| 34 |
>> + |
| 35 |
>> + # Store modified variables in build-info. |
| 36 |
>> + cd "${PORTAGE_BUILDDIR}"/build-info || die |
| 37 |
>> + set -f |
| 38 |
>> + |
| 39 |
>> + IFS=$' \t\n\r' |
| 40 |
>> + for f in INSTALL_MASK; do |
| 41 |
> |
| 42 |
> This loop along with the whole indirection is entirely pointless, given |
| 43 |
> that you're processing exactly one variable. |
| 44 |
|
| 45 |
I did this for consistency with the related loop in dyn_install, since |
| 46 |
we might save values of other variables from the binary package |
| 47 |
environment. In fact, I think we should record DOC_SYMLINKS_DIR here, |
| 48 |
since it affects CONTENTS, much like INSTALL_MASK. |
| 49 |
|
| 50 |
>> + x=$(echo -n ${!f}) |
| 51 |
>> + [[ -n ${x} ]] && echo "${x}" > "${f}" |
| 52 |
> |
| 53 |
> There's probably no point in this [[ -n ... ]], as that: |
| 54 |
> |
| 55 |
> a. requires you to special-handle missing INSTALL_MASK file, while it's |
| 56 |
> easier to just ensure that it's there (and I think you requested |
| 57 |
> the same thing from me before you rewritten my commit into breakage), |
| 58 |
> |
| 59 |
> b. makes it impossible to distinguish packages from before INSTALL_MASK |
| 60 |
> storing was added from those where it is empty. |
| 61 |
|
| 62 |
Maybe this only matters within the context of bug 364633 [1], and for |
| 63 |
that I think we need to introduce a separate CONTENTS.INSTALL_MASK file |
| 64 |
so that we can easily toggle collision-protect behavior to use |
| 65 |
CONTENTS.INSTALL_MASK when desired. |
| 66 |
|
| 67 |
[1] https://bugs.gentoo.org/364633 |
| 68 |
-- |
| 69 |
Thanks, |
| 70 |
Zac |
Archives