| 23 Aug 2025 |
 John Ericson | emily: err so as I interpret this is a way to do something like "if the user specifies --target=..... then inject -L ....." | 00:59:53 |
| John Ericson | not just a static set of flags | 01:00:00 |
 emily | but config files already offer that? they key on the target | 01:00:11 |
| John Ericson | the the conditions are more semantic than syntactic | 01:00:11 |
| John Ericson | can they key off -march and whatnot too? | 01:00:31 |
| emily | have you read the docs for the configuration files? they key on both target and language | 01:00:32 |
| emily | I don't think so, but our wrappers also don't do that | 01:00:42 |
| John Ericson | (other ways of doing target in effect?) | 01:00:43 |
| emily | -march is separate from the target | 01:00:54 |
| John Ericson | yeah I was thinking about how there was that PR for hacking multi arch in CC wrapper for buildingl inux | 01:01:03 |
| emily |
| 01:01:11 |
| emily | Note that options such as --driver-mode=, --target=, -m32 affect the search algorithm. For example, the aforementioned executable called with -m32 argument will instead search for:
| 01:01:13 |
| emily | *
Note that options such as --driver-mode=, --target=, -m32 affect the search algorithm. For example, the aforementioned executable called with -m32 argument will instead search for:
| 01:01:15 |
| emily | https://clang.llvm.org/docs/UsersManual.html#configuration-files | 01:01:18 |
| John Ericson | ok cool | 01:01:25 |
| John Ericson | I guess whoever made the multilib yaml didn't know about this :D | 01:02:29 |
| emily | I think it's specifically for configuring the GCC multilib interface | 01:05:42 |
| emily | but yeah, some overlap | 01:06:02 |
