search for: irremedi

On 14/07/15 12:26, Steffen Kaiser wrote: > > You asked about "newer dovecot versions", v2.2 does so. > Fair enough :) So, with v2.2+ the hardlink approach is irremediably gone, at least with LMTP (and without relying to SiS)? -- Danti Gionatan Supporto Tecnico Assyoma S.r.l. - www.assyoma.it email: g.danti at assyoma.it - info at assyoma.it GPG public key ID: FF5F32A8

...s it because of the loss of information in having the v-table stored as a "blob" of bytes ? (which means that Clang should pass more typed information, without changing the exact layout obviously given the ABI constraints) - Or is it something internal to LLVM (the information is somehow irremediably lost) ? I admit that reducing the virtual call overhead is probably not really worth it (in general), however devirtualizing calls also expose more inlining/context opportunities and it's hard (for me) to quantify what such an optimization could bring here. We should also consider the sim...

...s it because of the loss of information in having the v-table stored as a "blob" of bytes ? (which means that Clang should pass more typed information, without changing the exact layout obviously given the ABI constraints) - Or is it something internal to LLVM (the information is somehow irremediably lost) ? I admit that reducing the virtual call overhead is probably not really worth it (in general), however devirtualizing calls also expose more inlining/context opportunities and it's hard (for me) to quantify what such an optimization could bring here. We should also consider the sim...

...v-table >> stored as a "blob" of bytes ? (which means that Clang should pass more >> typed information, without changing the exact layout obviously given >> the ABI constraints) >> >> - Or is it something internal to LLVM (the information is somehow >> irremediably lost) ? >> >> >> I admit that reducing the virtual call overhead is probably not really >> worth it (in general), however devirtualizing calls also expose more >> inlining/context opportunities and it's hard (for me) to quantify what >> such an optimizati...

...nformation in having the v-table > stored as a "blob" of bytes ? (which means that Clang should pass more > typed information, without changing the exact layout obviously given > the ABI constraints) > > - Or is it something internal to LLVM (the information is somehow > irremediably lost) ? > > > I admit that reducing the virtual call overhead is probably not really > worth it (in general), however devirtualizing calls also expose more > inlining/context opportunities and it's hard (for me) to quantify what > such an optimization could bring here. We...

On 14/07/15 08:17, Steffen Kaiser wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On Mon, 13 Jul 2015, Gionatan Danti wrote: > >> On the other hand, private (per-user) sieve file works without >> interfering with hardlinks. In a similar manner, disabling sieve also >> permits dovecot to create multiple hardlinks for a single message. >> >>

...as a "blob" of bytes ? (which means that Clang should pass more > >> typed information, without changing the exact layout obviously given > >> the ABI constraints) > >> > >> - Or is it something internal to LLVM (the information is somehow > >> irremediably lost) ? > >> > >> > >> I admit that reducing the virtual call overhead is probably not really > >> worth it (in general), however devirtualizing calls also expose more > >> inlining/context opportunities and it's hard (for me) to quantify what &g...

...; of bytes ? (which means that Clang should pass more >> >> typed information, without changing the exact layout obviously given >> >> the ABI constraints) >> >> >> >> - Or is it something internal to LLVM (the information is somehow >> >> irremediably lost) ? >> >> >> >> >> >> I admit that reducing the virtual call overhead is probably not really >> >> worth it (in general), however devirtualizing calls also expose more >> >> inlining/context opportunities and it's hard (for me)...

...means that Clang should pass more >>> >> typed information, without changing the exact layout obviously given >>> >> the ABI constraints) >>> >> >>> >> - Or is it something internal to LLVM (the information is somehow >>> >> irremediably lost) ? >>> >> >>> >> >>> >> I admit that reducing the virtual call overhead is probably not really >>> >> worth it (in general), however devirtualizing calls also expose more >>> >> inlining/context opportunities and it...