Comments on: Code Reviewing: popt http://rusty.ozlabs.org/?p=55 Stealing From Smart People Thu, 19 Aug 2010 18:12:38 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Jeff Johnson http://rusty.ozlabs.org/?p=55&cpage=1#comment-286 Jeff Johnson Mon, 21 Jun 2010 17:27:30 +0000 http://rusty.ozlabs.org/?p=55#comment-286 And the ever so important missing mailing list name spelled out in ASCII to avoid learning Yet Another Editor popt-devel at rpm5.org And the ever so important missing mailing list name spelled out in ASCII
to avoid learning Yet Another Editor
popt-devel at rpm5.org

]]> By: Jeff Johnson http://rusty.ozlabs.org/?p=55&cpage=1#comment-285 Jeff Johnson Mon, 21 Jun 2010 17:07:22 +0000 http://rusty.ozlabs.org/?p=55#comment-285 gak, a typo: "nudging" the Yiddish for "prodding" at least colloquially. gak, a typo: “nudging” the Yiddish for “prodding”
at least colloquially.

]]> By: Jeff Johnson http://rusty.ozlabs.org/?p=55&cpage=1#comment-284 Jeff Johnson Mon, 21 Jun 2010 17:06:08 +0000 http://rusty.ozlabs.org/?p=55#comment-284 I suppose I should leave a pointer to current active development of POPT 2.0 that was largely initiated by this review and private "nuging" from Rusty: Feature requests and patches are most gratefully received at The coding is already underway, perhaps 20% of the way to what *I* want to see in POPT 2.0, will likely result in a release sometime this fall (depending on my time & energy). I suppose I should leave a pointer to current active development
of POPT 2.0 that was largely initiated by this review and private
“nuging” from Rusty:

Feature requests and patches are most gratefully received at

The coding is already underway, perhaps 20% of the way to what *I*
want to see in POPT 2.0, will likely result in a release sometime this fall
(depending on my time & energy).

]]> By: rusty http://rusty.ozlabs.org/?p=55&cpage=1#comment-197 rusty Mon, 29 Mar 2010 03:46:45 +0000 http://rusty.ozlabs.org/?p=55#comment-197 alloca returns a chunk of memory of size "sizeof(*done)"; even though done is yet to be initialized, this declaration is valid C. memset's (little-used) return value is always the same as its first argument, ie. the return of alloca() in this case. The vanilla alternative to "poptDone done = memset(alloca(sizeof(*done)), 0, sizeof(*done));" would use a real variable rather than alloca; which is certainly clearer to those who haven't seen this before. But you can't zero it generically in a one-liner. alloca returns a chunk of memory of size “sizeof(*done)”; even though done is yet to be initialized, this declaration is valid C. memset’s (little-used) return value is always the same as its first argument, ie. the return of alloca() in this case. The vanilla alternative to “poptDone done = memset(alloca(sizeof(*done)), 0, sizeof(*done));” would use a real variable rather than alloca; which is certainly clearer to those who haven’t seen this before. But you can’t zero it generically in a one-liner.

]]> By: cnewbie http://rusty.ozlabs.org/?p=55&cpage=1#comment-191 cnewbie Thu, 11 Mar 2010 21:32:28 +0000 http://rusty.ozlabs.org/?p=55#comment-191 could you explain the trick on how the 'done' struct is zeroing? My first confusion is what I get when you dereference 'done', and then, second, wouldn't alloca() return a pointer to memory NOT the same as the 'done' struct (which would have been allocated automatically on the stack prior to the alloca() call right?)? You wouldn't be zeroing out the right memory? thanks could you explain the trick on how the ‘done’ struct is zeroing? My first confusion is what I get when you dereference ‘done’, and then, second, wouldn’t alloca() return a pointer to memory NOT the same as the ‘done’ struct (which would have been allocated automatically on the stack prior to the alloca() call right?)? You wouldn’t be zeroing out the right memory?

thanks

]]> By: rusty http://rusty.ozlabs.org/?p=55&cpage=1#comment-142 rusty Fri, 29 Jan 2010 05:08:31 +0000 http://rusty.ozlabs.org/?p=55#comment-142 Yes, talloc would be a fair win IMHO. Maybe that will provide the incentive to fork and rewrite? :) Yes, talloc would be a fair win IMHO. Maybe that will provide the incentive to fork and rewrite? :)

]]> By: Brad Hards http://rusty.ozlabs.org/?p=55&cpage=1#comment-140 Brad Hards Fri, 29 Jan 2010 00:16:44 +0000 http://rusty.ozlabs.org/?p=55#comment-140 Also used in openchange, and I find it quite easy to use. Unfortunately it is a bit easy to mis-use. I often find that valgrind tells me I've forgotten to clean up something that is deep in the popt stuff. Also used in openchange, and I find it quite easy to use. Unfortunately it is a bit easy to mis-use. I often find that valgrind tells me I’ve forgotten to clean up something that is deep in the popt stuff.

]]> By: Benno Leslie http://rusty.ozlabs.org/?p=55&cpage=1#comment-139 Benno Leslie Thu, 28 Jan 2010 21:12:43 +0000 http://rusty.ozlabs.org/?p=55#comment-139 Cool code review. Look forward to seeing more! Cool code review. Look forward to seeing more!

]]>