dprintf() polyfill for Illumos #1911
Conversation
sjmulder
changed the title
| int len, nwritten; | ||
|
|
||
| va_start(ap, format); | ||
| len = vasprintf(&s, format, ap); |
There was a problem hiding this comment.
Although posix doesn't require it, the general expectation is for dprintf to not do any memory allocation so that it can be used in signal handler context. Not doing allocation would also remove allocation related failure point.
There was a problem hiding this comment.
I agree it would be better but I couldn't find a better alternative with the available primitives. One could use a static buffer but then you're length limited.
| @@ -121,6 +121,10 @@ | |||
| #define alloca(size) __builtin_alloca(size) | |||
| #endif | |||
|
|
|||
| #ifdef __sun | |||
There was a problem hiding this comment.
If Illumos is maintained, why don't they add dprintf? It's more economical for them to add it instead of N projects to carry their own implementation.
There was a problem hiding this comment.
Would be better if they did indeed. But that does leave legacy systems, and even there nnn can be very useful.
|
@N-R-K please merge when ready. |
sjmulder
changed the title
|
Turns out there aren't that many dprintf() calls so perhaps this alternative is better: #1912 |
This the improved version submitted to upstream: jarun/nnn#1911
|
Closed with #1912. |
I realise it's a bit of a long shot, submitting a PR to fix a legacy platform, but I thought I might at least try upstreaming this patch we've been using for a while.
Edit: this is also for Illumos, which is actively maintained.