diff options
author | Thierry Vignaud <tv@mageia.org> | 2012-03-12 19:31:03 +0000 |
---|---|---|
committer | Thierry Vignaud <tv@mageia.org> | 2012-03-12 19:31:03 +0000 |
commit | 95e031578306cf1ef7656404d3e2b49db80497f4 (patch) | |
tree | a96c4252a6cc3b13ea2f5f911af4318390df1f9a /advertising | |
parent | 7dd226f80f477a768b808a4efe761f78c345c247 (diff) | |
download | drakx-95e031578306cf1ef7656404d3e2b49db80497f4.tar drakx-95e031578306cf1ef7656404d3e2b49db80497f4.tar.gz drakx-95e031578306cf1ef7656404d3e2b49db80497f4.tar.bz2 drakx-95e031578306cf1ef7656404d3e2b49db80497f4.tar.xz drakx-95e031578306cf1ef7656404d3e2b49db80497f4.zip |
(cmp_kernel_versions) move it from any.pm to common.pm; thus preventing lazy loading of syscall.ph through
any->detect_devices>modules>list_modules>MDK::Common>MDK::Common::System>syscall.ph
This can end in crashing mgaaplet when performing live update (mga#3042)
because of lazy loading of syscall (by which time perl may as well have been
updated including syscall.ph which leave eg perl/5.12.x for perl/5.14.x)
This also saves quite some memory for:
- net_applet (5Mb aka 11%),
- mgapplet (7Mb aka 14%)
(as well as maybe for drakxnet but that's more transient):
21736 root 20 0 240m 44m 10m S 0 1.1 0:00.63 net_applet
21501 root 20 0 236m 39m 11m S 0 1.0 0:01.09 net_applet
23116 root 20 0 297m 50m 12m S 0 1.3 0:00.58 mgaapplet
23259 root 20 0 290m 43m 12m S 0 1.1 0:00.40 mgaapplet
Diffstat (limited to 'advertising')
0 files changed, 0 insertions, 0 deletions